John Doerr

1. Snapshot and testable edge thesis

Doerr's most defensible investment system is not “spot the future.” It is a conversion loop: use dense operator, founder, and university networks to see enabling shifts; select ambitious and adaptive teams inside those shifts; obtain partnership and governance support; then convert the thesis into recruiting and measurable execution. The scope is KPCB's early company-building and internet eras, not every later vehicle or public-policy activity. [researcher inference · moderate] profile thesis · talk frameworks

The public denominator is 49 searched routes: 46 eligible company candidates plus three passes; 14 are sufficiently reconstructed, 32 remain unresolved, and zero have audited economic reconstructions. Career-wide hit rate, return, IRR, DPI, TVPI, alpha, and personal economics are not computable. [researcher inference · strong] count rules · economics boundary

2. Core philosophy

1. People before plans. Doerr repeatedly says teams are the scarce input and describes plans as revisable. The support is direct doctrine and selected behavior, not a validated founder-prediction model. [investor-stated · strong] Edge, team and plan passages
2. Mission before map, evidence before company. A consequential problem can define where to look; customer need, technical proof, price-performance, team capacity, and financing duration must decide whether to invest. [researcher inference · strong] writings synthesis, sourcing and picking · cleantech forensics
3. Execution must be visible. Objectives make direction legible; key results make progress inspectable; transparent review should trigger learning rather than mechanical control. [investor-stated · strong] Wired excerpt, OKR superpowers
4. Service requires work and attribution. Recruiting, coaching, introductions, and board intervention can help, but founders, operators, partners, and vehicles keep their own credit. [researcher inference · strong] NVCA speech, “Assist entrepreneurs” · people map
5. Capital must fit the clock. Internet software and physical climate ventures retire different risks at different speeds. Price, terms, ownership, reserves, exits, fees, carry, LP rights, and return doctrine have insufficient public record across the career. [researcher inference · strong] Bloom lab · vehicle map

This synthesis is deliberately narrower than Doerr's slogans. “Ideas are easy; execution is everything” becomes investable only after the analyst defines evidence, owner, deadline, capital, decision consequence, and counterfactual. [researcher inference · strong] writings decision playbook

3. Sourcing

The historical sourcing engine was networked rather than purely inbound. Intel and semiconductor relationships, Stanford-area technical networks, founders, executives, KPCB partners, and trusted intermediaries produced both opportunities and diligence. The 1996 profile's 2,000-plan funnel describes partnership reach; it cannot be assigned to Doerr personally. [contemporaneous record · strong] Institutional Investor, network and funnel passages

A reusable sourcing receipt should name the problem, platform change, original source, trusted referrer, partner sponsor, expert challenge, conflicts, and rejected alternatives. This is a researcher-designed control, not a recovered KPCB form. [researcher inference · strong] networked-diligence framework

4. Picking

Doerr's 2018 five-screen summary is the cleanest direct checklist: technical excellence, ability to build a team, a large unmet need, reasonable financing, and speed. It is a retrospective public summary, not a decision-date scorecard or proof of predictive efficacy. [investor-stated · strong] CNN transcript, final Zakaria/Doerr block

The evidence supports five harder translations:

• Product before label: identify the technical or customer superiority that works now. Google had a better search product; Amazon had observable customer focus; a “platform” remains a hypothesis until the wedge earns expansion. [retrospective witness · moderate] Google lab · Amazon lab
• Ambition with adaptability: test whether the team can recruit, revise, and operate under bad news—not merely articulate a huge market. [researcher inference · moderate] Edge, team construction · Fisker lab
• Fatal-risk sequence: identify technical, manufacturing, distribution, regulatory, capital, and adoption risks; specify which evidence retires each one. [researcher inference · strong] QuantumScape lab · Bloom lab
• Reasonable financing means full duration: an initial round is not “reasonable” if commercial proof requires repeated large financings that the vehicle cannot support. [researcher inference · strong] cleantech portfolio forensics
• Mission does not waive economics: “clean discount,” cost parity, and customer willingness are company gates even when the externality is urgent. [investor-stated · moderate] 2026 letter, PDF p. 3

Price, security, ownership target, dilution, reserves, and exit criteria have insufficient public record as a general Doerr doctrine. Do not infer them from celebrated outcomes or later valuations. [researcher inference · strong] uncertainty ledger

5. Deal and portfolio mechanics

The public KPCB process around 1996 required two sponsoring partners, allowed theoretical vetoes, generally sought to lead, and expected a board seat. This is decision-right evidence for that period, not proof that the rules applied unchanged to every fund or later deal. [contemporaneous record · strong] Institutional Investor, partnership-process passages

Use a three-ledger discipline: operating outcome, security/vehicle outcome, and investor cash flow. Never derive return from company valuation, paper mark, share count alone, or a hypothetical alternative. [researcher inference · strong] economics boundary

6. Vehicle and incentive fit

Doerr's public record spans numbered KPCB/KP funds and associates, side/nominee entities, Digital Growth Funds I–III, family and trust vehicles, Foris-related positions, Doerr Capital, Breakthrough Energy Ventures, a family foundation, and philanthropy. They are not one pool. [contemporaneous record · strong] vehicle table

Vehicle fit is a substantive part of the thesis: the fund's duration, reserve capacity, mandate, governance, and LP obligations must match the technology clock. The cleantech record makes that a failure control, not a footnote. [researcher inference · strong] Bloom and Fisker labs

7. Team attribution and work with founders

Doerr's “servant VC” language is strongest when converted into named work: recruit missing leaders, connect technical and market experts, coach without taking the founder's decision, and install an agreed operating cadence. It is weakest when “we” compresses firm capital, partner work, founder execution, and later outcome. [researcher inference · strong] NVCA speech, service passages · people map

A founder-support contract should name the founder-owned objective, investor deliverable, owner, deadline, evidence, decision right, conflict, and stop date. This is a researcher-designed governance control, not a universal KPCB policy. [researcher inference · strong] modern playbook

8. Views on the craft

• Venture is company building, not transaction collecting. Doerr tells new investors to differentiate, network, gain operating experience, and assist entrepreneurs as builders. [investor-stated · strong] NVCA speech, career advice and “Assist entrepreneurs”
• Teams adapt; plans do not. The investor should underwrite who can recruit, learn, and change course rather than worship the original deck. [investor-stated · strong] Edge, team and plan passages
• Platforms create openings, not winners. Workstations, the web, mobile/cloud, climate infrastructure, and AI can expand the search space while leaving company selection unresolved. [researcher inference · strong] Senate AI statement, p. 1 · forecast scorecard
• Capital is an intervention tool with failure modes. The 2002 reset says too much money can damage standards; cleantech shows too little duration can also kill learning. [researcher inference · strong] Charlie Rose 2002 · cleantech reconciliation
• Measurement must preserve judgment. Doerr says OKRs cannot replace leadership, judgment, or culture and may be revised after reassessment. [investor-stated · strong] Wired excerpt, “not a silver bullet” and endangered-result passages
• Public ecosystems affect private outcomes. Talent, research, infrastructure, regulation, acquisitions, and IPOs shape company possibility; investor incentives must be disclosed when policy can affect holdings. [researcher inference · strong] Senate AI statement, pp. 4–5 · Senate climate hearing

9. Signature bets and decision process

The decision-date curriculum lives in investments.md. Each full lab freezes known, unknown, believed, disputed, alternatives, context, risks, terms, vehicle, and roles before the outcome; each asks “Would you invest then?”, kill criteria, terms/ownership, and the next evidence check. [researcher inference · strong]

1. Google — superior search and enormous ambition, but no mature monetization and a founder-imposed two-firm structure. [retrospective witness · moderate]
2. Amazon — customer obsession and fast learning inside an uncertain retail/internet market, with filing-visible KPCB vehicles. [retrospective witness · moderate]
3. Netscape — enabling-browser thesis and rapid market formation, with incompatible check/ownership recollections. [contemporaneous record · moderate]
4. DoorDash — crowded delivery market reframed as local logistics infrastructure; round size has reported $35/$40/$45 million variants. [retrospective witness · moderate]
5. Bloom Energy — an ordinary/mixed physical-venture case that tests time, capital intensity, controls, and board persistence. [contemporaneous record · strong]
6. QuantumScape — technical promise separated from multi-layer, manufacturing, financing, and commercialization risk. [contemporaneous record · strong]
7. Fisker versus Tesla — the conflict-constrained choice and the strongest admitted selection failure. [retrospective witness · moderate]
8. Sun Microsystems and Intuit — compact early cases that constrain hero attribution and expose missing terms/economics. [retrospective witness · strong]

The case set is pedagogically broad but not statistically representative. It cannot produce a career hit rate or investment rule frequency. [researcher inference · strong] denominator

10. Mental models and maxims

The model lane froze exactly nine evidence-supported rules. Titles and links below mirror the physical field guide; no additional slogan is promoted to a model. [researcher inference · strong] model reconciliation

1. Let mission define the search space; let evidence select the company
2. Underwrite adaptive teams, then coach without taking false credit
3. Ride enabling platform waves, not product labels
4. Turn the thesis into measurable milestones and transparent review
5. Revise the plan without erasing the audit trail
6. Retire fatal risks in sequence and match capital to the technology clock
7. Use networked diligence to turn uncertainty into intervention levers
8. Underwrite the ecosystem and disclose the investor's policy incentive
9. Keep people, vehicles, and outcomes in separate ledgers

Operator apprenticeship remains career advice; the 13-year technology-wave cadence remains an unvalidated narrative; startup/speedup/co-venture taxonomy has insufficient decision evidence; and standalone clean-discount or capital-abundance slogans are folded into the relevant risk and company-selection models. [researcher inference · strong] candidate-model disposition

11. Evolution and contradictions

The stable core is team, ambition, enabling change, networked help, and execution discipline. The real updates are capital restraint after the crash, capital-duration awareness after cleantech, and revision transparency after the climate scorecard. The principal unresolved conflict is whether these updates changed actual IC, reserve, or shutdown decisions; that has insufficient public record. [researcher inference · strong] cross-case forensics

12. Failures and limits

Failure is not one bucket. The six full adverse postmortems distinguish selection, technical, financing, governance, exit, and attribution errors. [researcher inference · strong] postmortem set

Bloom is the most useful ordinary/mixed limit: early technical importance and long governance exposure coexist with capital intensity, losses, control weaknesses, and no audited KPCB return. It prevents the curriculum from becoming only spectacular wins and terminal failures. [researcher inference · strong] Bloom lab

The method's structural limits are survivor-selected narratives, a 14-of-46 sufficient public reconstruction, zero audited economics, missing passes, uncertain partner roles, changing vehicles, and the absence of general price/terms/reserve/exit doctrine. [researcher inference · strong] uncertainty ledger

13. Historical operating environment

No decision-date base-rate dataset covers all these eras on compatible definitions. The historical record is therefore contextual, not a quantitative adjustment to returns. [researcher inference · strong] evidence boundary

14. Modern VC translation (as of 2026-08-23)

The following playbook is a researcher-designed translation for decisions made as of 2026-08-23. Its premises must be revalidated for each company; it is not represented as KPCB policy or as a measured causal study. [researcher inference · strong] writings playbook

Current IC minimum. Freeze known/unknown/believed/disputed; list alternatives; identify five fatal risks; specify source, sponsor, diligence, approval, board and follow-on owners; set price/security/ownership/reserve or write insufficient public record; define kill criteria; name the next evidence check; and show outcome only after the decision. [researcher inference · strong] decision labs

15. Comparison with completed peers

Only the three requested completed local peers are compared: Vinod Khosla, Michael Moritz, and Bob Kagle. Their adversarial reviews mark the packages as bounded curricula, not compatible personal-return studies. [researcher inference · strong] Khosla review · Moritz review · Kagle review

The comparative result is not a ranking. Doerr is most distinctive in pairing networked company building with a portable execution system; Khosla in explicit fatal-risk experimentation and vehicle breadth; Moritz in wedge-level observation and institutional renewal; Kagle in equal-partner service and behavioral customer evidence. Across all four, team, access, vehicle, era, and outlier controls are stronger than public evidence for repeated personal alpha. [researcher inference · strong] Doerr denominator

16. The strongest case against greatness

The steelman skeptical case is that Doerr's visible record is KPCB's elite access plus founder execution and public-market timing, narrated through Google. KPCB already had brand, partner networks, capital, and company-building infrastructure; Google was a joint, founder-conditioned round; Amazon, Netscape, Sun, and Intuit used fund vehicles and multiple contributors; later climate stories are selected and economically unaudited. [researcher inference · strong] Google announcement · Pareto analysis

The denominator intensifies the objection: 14 sufficiently reconstructed candidates out of 46 eligible company leads, 32 unresolved, three passes, no complete opportunity set, and zero audited economic reconstructions. Google likely dominates the public economic ranking, yet an exact outlier share or ex-Google performance cannot be computed. [researcher inference · strong] denominator

The operating doctrine is also hindsight-exposed. OKRs are inherited from Grove and publicized after Google; founder/team stories are winner-selected; a 2002 capital-discipline update sits beside a later large cleantech allocation; “servant VC” language coexists with board and financing power; and climate target revisions weaken accountability without a versioned ledger. [researcher inference · strong] contradiction scorecard · adverse boundaries

The favorable case survives only at a narrower level: Doerr repeatedly helped institutionalize technical-market learning, recruiting, goal discipline, and partnership succession, and he publicly admitted meaningful errors. The theses diverge on hidden evidence. Dated funnels, passes, IC votes, board labor, founder assessments, vehicle cash flows, reserve decisions, and an ordinary-case matched cohort would support or falsify personal repeatability. [researcher inference · strong] retrieval priorities

17. Analyst study guide, glossary, and go deeper

Recommended study path

1. Read the chronological profile and write the edge, alternative explanation, predictions, and falsifiers in your own words.
2. Read the 1997 Edge interview, then the 2002 reset, before reading outcome-rich retrospectives. [researcher inference · strong]
3. Replay Google, Bloom, and Fisker/Tesla without revealing the outcomes.
4. Use the writings chronology and talk contradiction scorecard to distinguish contemporaneous doctrine from retrospective repair.
5. Apply all nine mental models, then attack them with the six postmortems and uncertainty ledger.
6. Compare the same eight dimensions across Khosla, Moritz, and Kagle; do not rank returns from incompatible evidence.

This sequence is a researcher-designed curriculum, not Doerr's prescribed reading order. [researcher inference · strong] package evidence boundary

Investor-specific glossary

Reusable checklists

Sourcing and picking — define mission and why venture; record source/referrer/sponsor; separate platform from product; test technical or customer superiority; reference adaptability; name missing hires; list alternatives; disclose conflicts. [researcher inference · strong] models 1–3

Diligence and IC — freeze known/unknown/believed/disputed; map five fatal risks; specify the cheapest discriminating evidence, threshold, owner, time, and spend; record partner vote/dissent; set price, security, ownership, reserves, and vehicle—or insufficient public record; write kill criteria and next check. [researcher inference · strong] decision labs

Governance and follow-on — name founder-owned objective and investor deliverable; version milestones; separate advice from authority; re-underwrite every follow-on; disclose related-party and policy incentives; preserve an abandonment path; log board/recruiting work and exit decision. [researcher inference · strong] models 4–9

Attribution and economics — keep source, sponsor, diligence, approval, board, recruiting, follow-on, exit, narrator, and vehicle roles separate; never convert share count or valuation into return; distinguish company outcome, security outcome, vehicle cash flow, LP return, carry, and personal wealth. [researcher inference · strong] economics boundary

Ranked sources and next retrievals

1. Institutional Investor, 1996 — best contemporaneous biography, funnel, network, and partnership-process account. [contemporaneous record · strong]
2. Edge, 1997 — earliest full direct philosophy on teams, platforms, Amazon, Netscape, Intel, and coaching. [investor-stated · strong]
3. Google 1999 announcement and 2004 S-1 — joint financing, governance, vehicle, and ownership controls. [contemporaneous record · strong]
4. Charlie Rose, 2002 — strongest candid boom/bust update and rejected strategic alternatives. [investor-stated · strong]
5. Wired Measure What Matters excerpt — best operating doctrine plus Google retrospective, with amount conflict retained. [investor-stated · strong]
6. Speed & Scale official sample — clearest climate framework, collaborator record, and Fisker admission; access is partial. [investor-stated · strong]
7. SEC/company case ledger — strongest independent vehicle, board, adverse, and outcome controls. [contemporaneous record · strong]
8. Climate One, 2026 — current off-track admission, deadline/baseline revision, and collaborator attribution. [contemporaneous record · strong]

Highest-value next evidence is original IC/pass/follow-on memoranda; a complete KPCB opportunity and loss ledger; partnership agreements, LPAs, side letters, fees/carry and cash flows; board packets and founder assessments; full Google/Netscape/Amazon cap-table histories; a complete cleantech vehicle schedule; and versioned OKR/climate scorecards. Until recovered, price, terms, reserves, exits, carry, and returns have insufficient public record except where a case filing discloses a bounded fact. [researcher inference · strong] ranked retrieval priorities

Pareto 80/20 — what drove the record

The public record supports a power-law interpretation, not an audited 80/20 calculation. No KPCB cash-flow ledger, LP statement set, or partner-level carry schedule was recovered, so the rankings below use three different measures and never turn valuations, nominee holdings, or reported paper values into returns. [researcher inference · strong] 1996 KPCB selection process and ownership target economic boundary

Ranked view 1 — likely economic importance, evidence-bounded

Dominant-outlier exclusion test. Remove Google entirely. Amazon and Netscape still explain the historical record through two different Internet mechanisms—transaction infrastructure and browser/platform formation—and Sun supplies an earlier open-systems precedent. What disappears is any defensible ordering of economic magnitude: without realized cash flows, the remaining cases cannot be ranked by MOIC, DPI, or net IRR. [researcher inference · strong] Amazon case Netscape case Sun compact case

Ranked view 2 — historical importance

This ranking measures category and method influence, not money. The official current portfolio is self-selected and no career-wide accepted/rejected denominator exists. [researcher inference · strong] Kleiner Perkins profile

Ranked view 3 — learning value

Evidence boundary and denominator

Count rules and mechanical reconciliation

The bounded public denominator is 49 searched decision routes: 46 eligible company candidates plus three explicit passes/misses. The 46 consist of all 25 cards on Doerr's live Kleiner Perkins profile, seven additional direct company associations, twelve KPCB-only Internet-strategy leads whose individual attribution remains unresolved, and two additional adverse cleantech candidates. [researcher inference · strong] official 25-card profile 1996 company/funnel record 1999 board biography

Sufficiently reconstructed means the record supports a dated decision/thesis or documented association, vehicle/terms evidence or an explicit statement that it is absent, role/governance boundaries, outcome, and unresolved fields. It does not mean the economics are complete. Excite and @Home count as two original candidates because each was an investment before their merger; one Excite@Home postmortem covers the combined outcome without collapsing the two original decisions. [researcher inference · strong] 1999 merger record

The scoped reconstruction fraction is therefore 14/46 (30.4%). It describes only this named public candidate universe, not career coverage, opportunity coverage, or capital-weighted coverage. Career-wide coverage and the investment hit rate are not computable because no complete accepted/rejected ledger, fund schedule, or audited cash-flow series is public. [researcher inference · strong]

Deduplication rules are literal: iControl is one identity; Healtheon/WebMD is one evolving company; AOL/America Online and SportsLine spelling variants each count once; @Home and Excite remain separate original investments; Netscape and acquirer AOL remain separate; Homeshop.com in 1999 is not silently normalized to Homestore.com in 2004; and the 1996 OnLive! Technologies is not assumed to be the later game-streaming company. [researcher inference · strong] Google 1999 biography Google 2004 biography

An NFX Signal route is retained only as a discovery warning: its page showed 121 “Investments on Record” while the table header showed 75, so neither number is imported into the denominator. [contemporaneous record · provisional] Signal profile

Economics boundary

No case has a complete same-vehicle cash ledger. Reported valuations, financing totals, IPO holdings, beneficial-ownership tables, and acquisition consideration are not returns. This chapter therefore reports exact dated checks and securities when available, identifies whether economics are reported or computable, and calculates no IRR, MOIC, TVPI, DPI, cleantech hit rate, or personal profit. [researcher inference · strong]

Historical and decision-date context

Vehicle economics and incentives

Across all vehicles, management fees, carry percentages, hurdle or clawback terms, LP rights, reserve policies, opportunity allocation, side-letter terms, and Doerr's realized proceeds remain insufficient public record. The 2016 chair transition and 2018 growth split establish organizational boundaries, not those economics. [researcher inference · strong] chair statement Axios split

People, capital, and influence map

Comprehensive company ledger

The ledger includes every eligible candidate once. Sufficient means sufficient for decision-case analysis under the gate above, never full economics. Profile-card claims use the live Kleiner Perkins catalog; the twelve firm-only Internet leads use the 1996 profile and remain individually unresolved. [researcher inference · strong] official profile 1996 company list

Explicit passes and misses — excluded from the 46 investments

Ex-ante cases and decision labs

Each lab freezes the record before disclosing the outcome. The prompts are analyst-created teaching judgments, not claims about the historical committee. [researcher inference · strong]

Case 1 — Google

Freeze — May/June 1999

Reader prompts before the reveal

1. Would you invest then? Conditionally, after blind relevance and retention tests confirm a material edge and the round preserves meaningful ownership. [researcher inference · moderate]
2. What would kill the deal? No durable relevance advantage, weak repeat use, serving costs outrunning revenue, or inability to complete the leadership team. [researcher inference · strong]
3. What ownership/terms fit the risk? Preferred equity, board/information and pro-rata rights, a reserve plan, and a founder-compatible executive-search covenant; these are teaching terms, not asserted historical terms. [researcher inference · moderate]
4. What evidence unlocks the next check? Named-engine benchmarks, query cohorts, cost/query, signed licensing revenue, and dated leadership hires. [researcher inference · strong]

Reveal

The 2004 S-1 lists Doerr as director since May 1999 and 23,893,800 shares [single-source] through KPCB Holdings; he disclaimed beneficial ownership except his pecuniary interest. It also documents preferred-holder director rights and later founder-control architecture. Google S-1, management/principal holders/capital stock [documented behavior · strong] No KPCB cash-distribution ledger exists, so the round size, nominee shares, or later valuation cannot produce a realized return. [researcher inference · strong]

Case 2 — Amazon

Freeze — June 1996

Reader prompts before the reveal

1. Would you invest then? Yes, if repeat-order and unit-contribution data support the approximately documented ownership and the burn has a financeable milestone path. [researcher inference · strong]
2. What would kill the deal? Repeat decay, structurally negative contribution after fulfillment/service, working-capital stress, or an incumbent price war. [researcher inference · strong]
3. What ownership/terms fit the risk? The documented lead Series A preferred with liquidation, conversion, protective, information, and board rights. [researcher inference · strong]
4. What evidence unlocks the next check? Monthly cohorts, frequency, gross/contribution margin, fulfillment accuracy/cycle time, CAC payback, supplier terms, and downside runway. [researcher inference · strong]

Reveal

The final prospectus shows 3,401,376 KPCB common-equivalent shares, 16.4% pre-offering and 14.6% post-offering [single-source], and continued Doerr board service. Amazon S-1/A, principal holders [documented behavior · strong] Paper-gain and later billion-dollar lore lack sale dates, follow-ons, distributions, fees, and carry, so no return is computed. [researcher inference · strong]

Case 3 — Netscape

Freeze — June presentation / September 1994 investment

Reader prompts before the reveal

1. Would you invest then? Conditionally, with evidence of distribution, server willingness-to-pay, and a credible team-building plan. [researcher inference · moderate]
2. What would kill the deal? A free bundled browser, unclean IP, loss of technical lead, weak server economics, or missing enterprise leadership. [researcher inference · strong]
3. What ownership/terms fit the risk? Lead preferred, meaningful ownership, board/information/pro-rata rights, and reserves for a platform fight; do not assume the disputed $5m/25% terms. [researcher inference · strong]
4. What evidence unlocks the next check? Adoption/retention, server pipeline, OEM/ISP distribution, release cadence, executive accepts, and written IP provenance. [researcher inference · strong]

Reveal

Netscape went public in 1995 and agreed to be acquired by AOL in 1998. KP case [retrospective witness · strong] Fortune's roughly $400 million KPCB outcome [single-source] is a reported sponsor figure, not an audited cash result. Fortune archive, pp. 2–3 [retrospective witness · provisional] The June/September date difference may be presentation, approval, signing, or closing; it remains unresolved. [researcher inference · strong]

Case 4 — DoorDash

Freeze — March 2015 Series B

Reader prompts before the reveal

1. Would you invest then? Only if several mature cities show contribution profit before shared overhead and a credible path to density without permanent subsidy. [researcher inference · strong]
2. What would kill the deal? Cohorts whose order economics worsen with age, weak merchant retention, courier scarcity, adverse classification economics, or no route-density advantage. [researcher inference · strong]
3. What ownership/terms fit the risk? Preferred with board/information/pro-rata rights, labor/legal reporting, city-level milestone releases, and reserves tied to mature-market proof; these are teaching terms. [researcher inference · moderate]
4. What evidence unlocks the next check? City cohorts, contribution/order, courier utilization, retention, refund/support burden, restaurant churn, and classification downside. [researcher inference · strong]

Reveal

DoorDash's S-1 records Doerr's board service, Series B designee status, and DGF holdings; by March 2021 the Doerr beneficial-ownership row mixed DGF nominee shares, DGF II Associates, VVT, and other controlled entities. S-1 2021 proxy, footnote 7 [documented behavior · strong] A June 2020 DGF II Series H purchase of 164,595 shares at $45.9062 cost $7,555,931 [single-source]; no full vehicle ledger makes the outcome computable. 2021 proxy, related transactions [documented behavior · strong]

Case 5 — Bloom Energy (ordinary/mixed case)

Freeze — 2002 first venture financing

Reader prompts before the reveal

1. Would you invest then? Only as a milestone-tranched technical option with independently replicated durability and a financeable demonstration. [researcher inference · moderate]
2. What would kill the deal? No credible delivered-cost path, field degradation, warranty tail, customer-finance dependence that hides economics, or capital needs beyond a pre-agreed financing chain. [researcher inference · strong]
3. What ownership/terms fit the risk? Tranched preferred, board/information/audit rights, pro-rata rather than automatic follow-on, technical and commercial covenants, and explicit reserve/stop rules; not asserted historical terms. [researcher inference · moderate]
4. What evidence unlocks the next check? Independent field hours/degradation, yield and material cost, service/warranty cohorts, unsubsidized delivered cost, contracted backlog quality, and project-finance availability. [researcher inference · strong]

Reveal

Bloom reached its IPO after sixteen years. Its filing reported revenue of $208.5 million in 2016 and $376.0 million in 2017, material-cost reduction of more than 75%, roughly $1.5 billion of gross private capital raised, and about $2.3 billion accumulated deficit. 2018 S-1, business/financial statements/capitalization [contemporaneous record · strong] Later accounting corrections and litigation/settlement records make this a mixed governance and duration case, not a clean win; none establishes entry economics or a KPCB cash return. Bloom 2019 10-K [researcher inference · strong]

Case 6 — QuantumScape

Freeze — December 2010

Reader prompts before the reveal

1. Would you invest then? Only as a small, staged technical option with explicit replication gates, because the market is enormous but feasibility dominates. [researcher inference · moderate]
2. What would kill the deal? Dendrites, weak cycle life, inability to scale layers/yield, automotive qualification delay, or factory economics requiring implausible capital. [researcher inference · strong]
3. What ownership/terms fit the risk? Tranched preferred with technical milestones, independent validation, board/information/pro-rata rights, and no valuation-based return shortcut; these are teaching terms. [researcher inference · moderate]
4. What evidence unlocks the next check? Repeated multilayer cells, yield distributions, full-cycle performance, safety abuse tests, OEM validation, and a bottoms-up factory cost/capex plan. [researcher inference · strong]

Reveal

Volkswagen's relationship began in 2012 and the company later entered public markets through a business combination with issuer forecasts that were not decision-date evidence. 2020 S-1 investor presentation [documented behavior · strong] A securities settlement of $47.5 million carried no admission; it is an adverse disclosure record, not proof that the 2010 science thesis was fraudulent. settlement administration [contemporaneous record · moderate] No same-vehicle cash ledger supports a return. [researcher inference · strong]

Case 7 — Fisker versus Tesla

Freeze — 2007–08 choice

Reader prompts before the reveal

1. Would you invest then? Only as a staged option, not a conviction-sized category winner; design and policy tailwinds did not retire integration risk. [researcher inference · moderate]
2. What would kill the deal? Failure of an independently validated production-intent vehicle, stable battery supply, pilot gross-margin path, or survival without undrawn government funding. [researcher inference · strong]
3. What ownership/terms fit the risk? Milestone-tranched preferred, actual-sponsor board representation, supplier/DOE information rights, pro-rata without automatic follow-on, and predeclared reserve limits; historical specifics remain unknown. [researcher inference · provisional]
4. What evidence unlocks the next check? Battery validation/recall data, enforceable supplier quality remedies, pilot unit economics, delivery without covenant relief, and a downside liquidity plan. [researcher inference · strong]

Reveal

Fisker stopped production in August 2012, laid off most staff in April 2013, and entered bankruptcy in November 2013 after roughly $1.1 billion of aggregate capital; DOE had committed $528.7 million and advanced $192 million. The bankruptcy asset sale was approximately $149.2 million, but neither that amount nor aggregate capital is a KPCB return. Orgone chronology and merits disclaimer bankruptcy opinion, opening facts [contemporaneous record · strong] Doerr later said Tesla out-executed the selected company and that KPCB's choice was wrong. Ritholtz [retrospective witness · strong]

Compact case — Sun Microsystems

Freeze — November 1982

The sponsor record describes an open Unix workstation, a Stanford/Berkeley team, and $8 million of sales in the first two quarters; it reports a $1.7 million KPCB investment [single-source]. The public record lacks the decision memo, security, vehicle, ownership, reserves, and vote. KP Sun case [retrospective witness · moderate] Doerr's directorship from 1982 is independently confirmed. Sun proxy [documented behavior · strong]

1. Would you invest then? Conditionally, on technical differentiation, customer proof, and open-systems adoption. [researcher inference · moderate]
2. What would kill the deal? Proprietary competitors, weak manufacturing, or no developer/network ecosystem. [researcher inference · moderate]
3. What ownership/terms fit the risk? Lead preferred and board/information/pro-rata rights; actual terms are unknown. [researcher inference · provisional]
4. What evidence unlocks the next check? Customer orders, gross margin, manufacturing yield, software compatibility, and the hiring plan. [researcher inference · strong]

Reveal

Sun became a major public workstation/server company and was acquired by Oracle. The opened sources provide no KPCB distribution ledger, so company success is not a computable sponsor return. Kleiner Perkins Sun case [researcher inference · strong]

Compact case — Intuit

Freeze — August 1990

A 1996 profile describes Intuit as profitable or near-profitable and close to an IPO; the product thesis was mass-market personal-finance software. Institutional Investor, stage discussion [retrospective witness · moderate] The proxy confirms Doerr's board service from August 1990, while McMurtry's oral history warns that company financing and board work were multi-party. Intuit proxy McMurtry oral history, 03-00:26 and 03-00:47:49 [retrospective witness · moderate] Check, security, vehicle, ownership, diligence, and vote are absent.

1. Would you invest then? Yes if retention, low support burden, and distribution economics justify a late private entry. [researcher inference · moderate]
2. What would kill the deal? Microsoft bundling, weak upgrade behavior, support cost, or a price that already assumes the IPO. [researcher inference · strong]
3. What ownership/terms fit the risk? Governance and downside rights appropriate to a near-IPO preferred; actual terms are unknown. [researcher inference · provisional]
4. What evidence unlocks the next check? Product cohorts, retail/OEM distribution, support cost, profitability quality, competitive response, and IPO downside. [researcher inference · strong]

Reveal

Intuit later survived Microsoft's competitive pressure; the DOJ challenged and Microsoft abandoned its proposed acquisition. DOJ case record [documented behavior · strong] No vehicle cash flows make the KPCB or Doerr result computable. [researcher inference · strong]

Portfolio forensics

Selection funnel, power law, and absence of a cash denominator

The 1996 partnership funnel—roughly 2,000 plans, 200 serious reviews, and 20–25 financings annually—shows extreme selection, while the reconstructed signature cases show the separate power-law problem: a few Internet platforms dominate the historical narrative. Institutional Investor, selection and Monday meeting sections [contemporaneous record · strong] Neither fact supplies a return distribution. The official 25-card page omits many failures and fund-only stakes, and none of the 46 candidates has audited entry-to-distribution economics. [researcher inference · strong]

Excluding Google does not erase the historical pattern—Amazon, Netscape, and Sun still span transactions, browsers, and open systems—but it destroys any defensible claim to an economic ranking. Excluding the top three Internet narratives leaves a record whose most instructive evidence is duration and failure, not return. [researcher inference · strong] Pareto exclusion test

Cleantech denominator reconciliation

The [self-reported] snapshots 9, roughly 70, and roughly 100, the direct capital/value claims $100m, $200m, roughly $1b, roughly $3b, and 80×, the [single-source] Fortune 54/$630m, and the [disputed] 70/100 and $3b/$3.2b variants are preserved rather than merged. Charlie Rose TED Fortune Ferriss Ritholtz Speed excerpt Forbes MIT's sector finding that more than 90% of its post-2007 cleantech companies failed to return capital explains the model stress but cannot substitute for KPCB results. MIT, pp. 4–5 and 17–18 [researcher inference · strong]

Cross-case forensic findings

1. Company and sponsor outcomes diverge. Excite@Home common equity became worthless after KPCB distributed shares to LPs; Amyris old equity was canceled, and the filing says DIP claims received successor equity but contradicts itself about Foris prepetition claims; Zynga sold without a public trail showing whether the 2013 KPCB/DGF stake remained. [researcher inference · strong] Excite@Home record Amyris 8-K, Items 1.02 and 3.02 Zynga acquisition
2. Nominee rows exaggerate apparent personal concentration. Google, DoorDash, Bloom, QuantumScape, Coursera, Amyris, and Zynga records decompose multiple funds, trusts, and affiliated entities; shared voting power is not personal cash ownership. [researcher inference · strong] Coursera proxy example
3. Persistence is observationally ambiguous. Bloom's long duration and Amyris/Fisker follow-ons can represent rational risk retirement or escalation of commitment. The absent milestone, reserve, and stop ledgers prevent choosing. Bloom S-1 Amyris 2023 proxy Fisker chronology [researcher inference · strong]
4. Governance attribution is role-specific. Doerr, Ray Lane, Bing Gordon, Tom Perkins, the Managing LLCs, DGF managers, Foris lenders, founders, and company boards made different decisions. “John Doerr deal” is a retrieval label, not a safe causal statement. [researcher inference · strong] Perkins oral history, role allocation
5. The public record is strongest on ownership snapshots, weakest on exits. Filings disclose entry securities in Amazon and later holdings elsewhere, but almost never distributions, secondary sales, write-offs, carry, or LP net cash flows. Amazon S-1/A Coursera proxy [researcher inference · strong]

Failure, miss, and anti-portfolio postmortems

Postmortem 1 — Fisker Automotive versus Tesla

Postmortem 2 — Amyris

Postmortem 3 — Excite@Home

Postmortem 4 — Zynga

Postmortem 5 — Weitek (pass/miss)

Postmortem 6 — MiaSolé

Uncertainty and retrieval ledger

Material conflicts retained

Legal and adverse-governance boundary

Ranked retrieval priorities

1. Fund and LP cash flows. KPCB VII–XIV and DGF I–III audited statements, capital calls, distributions, write-offs, fees/carry, side letters, and deal-level attribution would convert reported outcomes into economics. [researcher inference · strong]
2. Executed signature-deal records. Google closing cap table/stock agreement; Netscape term sheet/purchase documents; DoorDash Series B allocation; Amazon distribution ledger; Sun/Intuit security and vehicle records. [researcher inference · strong]
3. Cleantech program schedule. Company-by-company fund/side/family vehicle, check, follow-on, reserve, write-off, realization, and valuation-method schedule behind the 54/70/100 and $1b/$3b–$3.2b claims. [researcher inference · strong]
4. Fisker and Amyris governance. IC memos, board minutes, milestone and conflict decisions, round-by-round capital, and recovery waterfall; actual Tesla alternative terms/conflict policy. [researcher inference · strong]
5. Excite/@Home and Zynga exits. Entry basis, LP distribution records, entity-level sales/distributions, and merger/acquisition holdings continuity. [researcher inference · strong]
6. Ordinary and unresolved portfolio. Source/sponsor/terms/board/exit records for the 32 unresolved candidates, especially Go, Silicon Compilers, Coursera, Compaq, Slack, Twitter, Uber, and the twelve KPCB-only Internet leads. [researcher inference · strong]
7. Identity and archive gaps. Homeshop/Homestore legal identity, 1996 OnLive entity, official Netscape filing, Doerr oral history/papers, Rice 2000 VHS, Pao trial transcript/exhibits, and regulator/entity searches. [researcher inference · strong]

Search and saturation receipt

The integrated set uses more than 50 linked works, including more than 25 Tier 1 issuer filings, SEC records, court opinions, official announcements, government records, and institutional oral histories. Sponsor pages and retrospective interviews are used for stated thesis and reconstruction, not silently upgraded to audited economics. [researcher inference · strong]

The criticism/lawsuit battery combined Doerr/KPCB with failure, underperformance, cleantech, Fisker, MiaSolé, Amyris, Zynga, Excite@Home, Weitek, lawsuit, fraud, conflict, SEC, DOJ, governance, fund returns, and LP. It surfaced Pao, Alphabet, Sandys, Orgone, and Roth; no opened result established a Doerr-specific SEC or DOJ enforcement action. That is a bounded search result, not proof none exists. [researcher inference · moderate]

After incorporating MiaSolé, the filing families, and the legal dispositions, final targeted searches for MiaSolé ordinary-loss evidence, Weitek/Perkins attribution, and an Excite@Home-specific process change produced no new scoped material fact beyond the sources already captured. Practical web saturation was reached subject to the retrieval priorities above. [researcher inference · strong]

Coverage and evidence base

The written-corpus denominator is 11 priority works: all 11 ★★★ rows under “Books and signed writings” in sources.md. All 11 have source notes and all 11 receive substantive analysis below. Nine artifacts were accessible in full as published; two books were accessible only in part. “Full” for the WIRED and Fast Company items means the complete authorized excerpt, not the complete underlying book.

Direct voice establishes what Doerr said, not audited history or sole credit. His Google excerpt reports $11.8 million for about 12%, the publisher description reports $12.5 million, and Google's contemporaneous announcement discloses only an aggregate $25 million round led by Sequoia and KPCB. [retrospective witness · provisional] The firm split, vehicle, and personal-versus-partnership capital remain unresolved. WIRED, opening investment passage Penguin Random House, Book Description Google, June 7, 1999 announcement

The same caution applies to cleantech. Doerr's written account reports roughly $1 billion of “green venture investments” becoming worth $3 billion but gives no fund schedule, cash-flow basis, valuation date, or realized/unrealized split. [investor-stated · provisional] Separate direct interviews describe the cohort as roughly 100 companies and roughly 70 ventures; those counts conflict and cannot support a hit rate. Fast Company, “We pursued mobile apps” and “Kleiner Perkins got beaten up” Tim Ferriss transcript, cleantech discussion Masters in Business transcript, cleantech discussion

Source-by-source analysis

1. A New Role as KPCB Chair (2016)

• Authorship, context, and access. The complete signed firm announcement records Doerr's March 2016 transition to KPCB chair after advice he attributes to Bill Campbell. It is primary evidence for intended role design, not proof that the design persisted or worked. Kleiner Perkins, opening and “I will remain”
• Central argument and reasoning. Doerr frames the transition as acceleration: remain active in daily firm life and investment committees while using the chair role to invest, help teams scale, scan themes, recruit investors, and coach successors. Because venture changes continuously, the institution must change while remaining organized around entrepreneur service. [investor-stated · strong] Kleiner Perkins, “The more I thought,” “The venture industry,” and five objectives
• Evidence, examples, and assumptions. The evidence is an announced responsibility map and named leadership group, not outcome data. It assumes a senior investor can be both active “player” and succession “coach” without obscuring authority. [researcher inference · moderate] Kleiner Perkins, “Meanwhile” and “This last goal”
• Rule and operational implication. Keep sourcing and portfolio support while deliberately transferring pattern recognition and relationships; audit succession through committee seats, votes, deal ownership, carry, and time allocation rather than title. The first clause is Doerr's rule; the audit is an analytical translation. [investor-stated · strong] [researcher inference · moderate] Chair announcement, five objectives
• Decisions, biography, and attribution. Doerr says he would invest for both KPCB funds and his family foundation and remain on investment committees. That identifies distinct capital pools but not opportunity-allocation or conflict rules; they cannot be merged. [contemporaneous record · strong] Kleiner Perkins, “I will remain”
• Evolution and contradiction. The five items are called OKRs but contain no measurable key results, baselines, dates, owners, or grading cadence—a material tension with the doctrine published in 2018. [documented behavior · strong] Retaining daily and committee influence while empowering successors also leaves decision-right transfer untestable. [researcher inference · moderate] Kleiner Perkins, five objectives WIRED, “OKR Superpowers”

2. Measure What Matters (2018, partial)

• Authorship, context, and access. Penguin Random House identifies a 320-page U.S. Portfolio edition by John Doerr, with a foreword by Larry Page, published April 24, 2018. Only publisher metadata/description plus Google Books contents and selected pages were legitimately available; the complete text, notes, and continuous chapters were not. Penguin Random House, Product Details Google Books, Contents
• Central argument and reasoning. The accessible description defines objectives as desired achievements and key results as time-bounded measurable action. Grove teaches Doerr a repeatable system; transparent priorities align people; measurement creates focus and accountability; the system travels across organizations. [investor-stated · moderate] This is publisher-mediated description, so unavailable nuance is not assumed. Penguin Random House, Book Description
• Evidence, examples, and assumptions. The contents name Google, Intel, Remind, Nuna, MyFitnessPal, the Gates Foundation, Chrome, Zume, Lumeris, and ONE. That establishes case-study architecture, not causal validity. [contemporaneous record · moderate] No denominator, control group, failed-adoption inventory, or attribution method is accessible. [researcher inference · strong] Google Books, Contents
• Rule and operational implication. After investing, turn strategy into transparent milestones and use the record to focus teams and identify losing propositions. [investor-stated · moderate] Automatically linking scores to reserves, CEO changes, or shutdowns is unsupported because accessible material exposes no kill protocol. [researcher inference · strong] Penguin Random House, Book Description
• Decisions, biography, and attribution. The description makes Grove-to-Doerr transmission central and says Doerr shared the system with more than fifty companies. [retrospective witness · moderate] Its $12.5 million Google claim conflicts with $11.8 million/12% in the excerpt and with the contemporary $25 million aggregate round; none proves a personal check or sole decision. [retrospective witness · provisional] Penguin Random House, Book Description Google announcement
• Evolution and contradiction. Publication marks a shift from applying an inherited practice to codifying it publicly. [researcher inference · moderate] The partial record cannot establish whether KPCB's process changed or whether the full book adequately tests gaming, implementation cost, and failed use. [researcher inference · strong] Penguin Random House, Product Details and Book Description

3. When John Doerr Brought a “Gift” to Google's Founders (2018)

• Authorship, context, and access. WIRED published this complete authorized edited excerpt from Measure What Matters on release day. It is direct retrospection about 1999, not a contemporaneous IC memo or the complete book chapter. WIRED, opening and adaptation notice
• Central argument and reasoning. Ambition becomes executable when a team names a few objectives and attaches specific, time-bound, measurable, verifiable results. Focus, alignment, accountability, and stretch follow; review turns misses into corrective action. [investor-stated · strong] Doerr's own teaching objective included three observable results and a three-month trial, so adoption was an experiment. WIRED, “OKR Superpowers,” definitions, and “My objective that day”
• Evidence, examples, and assumptions. Intel's Operation Crush, Doerr's 14-month Sun role, Google's quarterly reviews, and a 2008 Code Yellow campaign demonstrate use, not isolated effect. [investor-stated · moderate] The excerpt itself acknowledges narrowed attention, unethical behavior, excess risk, weak cooperation, and reduced motivation as goal-setting hazards. [investor-stated · strong] WIRED, illustration, Sun, Code Yellow, and “In 2009” paragraphs
• Rule and operational implication. Diagnose the bottleneck, obtain leadership buy-in, set a small visible scorecard, review it, and revise endangered results after judgment rather than worshiping the metric. [investor-stated · strong] Process supplements—not replaces—product judgment, leadership, and culture. WIRED, “Superpower #3” and method boundary
• Decisions, biography, and attribution. Doerr remembers superior beta search, Page and Brin's huge ambition, and conviction despite limited management experience and no mature monetization plan. [retrospective witness · moderate] The first session included the founders, Marissa Mayer, Susan Wojcicki, Salar Kamangar, and roughly thirty colleagues; adoption and execution were collective. WIRED, pitch and first-session passages
• Evolution and contradiction. This is the clearest Grove-to-portfolio-company lineage, but also an attribution trap: Google's survivor outcome cannot isolate OKRs, and the $11.8 million figure conflicts with publisher copy while the contemporary record identifies a joint round and two board appointees. [researcher inference · strong] WIRED, opening and later-product passages Google announcement

4. John Doerr NVCA lifetime achievement award acceptance speech (2019)

• Authorship, context, and access. The complete signed text of the February 27, 2019 award speech was published by Kleiner Perkins on March 11. An award venue favors gratitude and success selection; its capital figures are unaudited retrospection. Kleiner Perkins, full speech
• Central argument and reasoning. Doerr rejects a lone-hero origin: Byers and Caufield supplied apprenticeship, Grove execution discipline, Campbell coaching, partner generations the institution, and founders the building. He derives a sequence: specialize, network, gain operating experience, incubate with founders, and serve as a builder rather than trade deals. [investor-stated · strong] Kleiner Perkins, mentor, “To those of you,” and “Assist entrepreneurs” passages
• Evidence, examples, and assumptions. Arrival in Silicon Valley, internship rejections, Intel, continued contact with Byers, Kleiner entry, plan reviews, and named partner generations illustrate the model. [retrospective witness · moderate] The speech assumes operating empathy and specialization improve service without comparing decisions made with and without them. [researcher inference · moderate] Kleiner Perkins, apprenticeship passages
• Rule and operational implication. Choose a field in which you can differentiate, build relationships before transactions, and help founders recruit and execute as a “servant VC.” [investor-stated · strong] An IC must still map where service ends and fiduciary control begins. [researcher inference · moderate] Kleiner Perkins, “To those of you” and “Assist entrepreneurs”
• Decisions, biography, and attribution. “Together we started Silicon Compilers, and @Home” compresses founders, partners, vehicles, and board roles. [retrospective witness · provisional] The stated 29 Kleiner funds, more than $11 billion raised, and alumni nearly $2 billion in ten funds cannot be converted into AUM, returns, or Doerr performance. Kleiner Perkins, “Since 1980” and incubation passages
• Evolution and contradiction. Team/service framing corrects personal-brand history. Yet narrow specialization sits uneasily with Doerr's breadth across semiconductors, internet, consumer, education, health, climate, and AI; the text never says whether the rule applies to a period, partner, or firm. [researcher inference · moderate] Kleiner Perkins, specialization advice and career inventory

5. DoorDash: delivering the future of local (2020)

• Authorship, context, and access. Doerr's complete seven-paragraph firm essay appeared December 9, 2020 at a celebratory company milestone. It is primary evidence for his remembered thesis and interested evidence for deal history. Kleiner Perkins, full essay
• Central argument and reasoning. Local merchants faced a convenience gap; restaurants and consumers were poorly served despite competition; the four-person founding team combined technology, customer commitment, and execution; therefore DoorDash could become last-mile infrastructure beyond meal delivery. [investor-stated · moderate] Kleiner Perkins, “Small businesses,” “When we met,” and “Just five years ago”
• Evidence, examples, and assumptions. Doerr names Tony Xu, Andy Fang, Stanley Tang, and Evan Moore and reports scale across merchants, dashers, customers, and countries. [retrospective witness · moderate] [contemporaneous record · provisional] The piece assumes expansion and mission continuity equal durable value but supplies no unit economics, cohorts, stakeholder distribution, or competitor test. [researcher inference · strong] Kleiner Perkins, “When we met” through “Today”
• Rule and operational implication. In a crowded market, prefer a team solving a persistent multi-sided problem with operating technology and a platform path; require each adjacency to improve density or customer utility. [researcher inference · moderate] Kleiner Perkins, “Just five years ago”
• Decisions, biography, and attribution. Doerr says “we” led the 2015 Series B; DoorDash's proxy records his board service from March 2015 and separates KPCB holdings, a family trust, and other controlled entities. [documented behavior · strong] The supported wording is KPCB led / Doerr served on the board, not personal check or sole sponsorship. DoorDash, 2021 proxy
• Evolution and contradiction. The essay applies the earlier people/execution rhetoric to a company but omits sourcing, price, ownership, reserves, board interventions, labor risk, and exit logic. [researcher inference · strong] Serving merchants can conflict with fees or bargaining power; the article does not test that tension. [researcher inference · moderate] Kleiner Perkins, full essay

6. John Doerr on addressing climate change (2021)

• Authorship, context, and access. Fast Company published this complete authorized Speed & Scale excerpt on November 8, 2021. It exposes the catalyst story, portfolio retrospection, and execution thesis, but not the complete book or endnotes. Fast Company, full excerpt
• Central argument and reasoning. A time-constrained emergency needs more than solutions: deploy what exists, invent what is missing, translate both into measurable OKRs, combine capital with policy, and coordinate institutions. [investor-stated · strong] Applying a company-management tool to global coordination is a proposed transfer, not proof OKRs reduce emissions. [researcher inference · provisional] Fast Company, “Fortunately,” “While a good many solutions,” and “Now I'm proposing”
• Evidence, examples, and assumptions. Doerr contrasts quick mobile wins with slow failure-heavy climate companies and reports $1 billion becoming worth $3 billion. [investor-stated · provisional] No vehicle, valuation date, write-off schedule, follow-on history, or counterfactual shows whether persistence created value; “surviving investments” is selection. [researcher inference · strong] Fast Company, mobile and cleantech passages
• Rule and operational implication. Test customer demand, but let an urgent systemic problem justify uncertain-market search; pair nearer-term opportunities with longer-duration experiments. [investor-stated · moderate] No sizing, reserve, ownership, stop, or exit rule appears, so urgency cannot justify unlimited patience. [researcher inference · strong] Fast Company, customer-demand and urgency passages
• Decisions, biography, and attribution. Doerr dates his climate priority to his daughter's reaction after a 2006 screening of An Inconvenient Truth, followed by a partnership priority and Al Gore's recruitment. [investor-stated · moderate] It is retrospective causation; “my partners and I” and Kleiner capital preclude sole credit. Fast Company, opening and partnership passages
• Evolution and contradiction. The text moves the execution doctrine from companies to climate systems and acknowledges failure more candidly than celebratory essays. Yet $1 billion/$3 billion cannot reconcile the roughly 70 versus 100 company counts or prove a fund return. [researcher inference · strong] Tim Ferriss transcript Masters in Business transcript

7. Speed & Scale (2021, partial)

• Authorship, context, and access. Penguin Random House dates the 448-page U.S. book to November 9, 2021. The official UK sample exposed front matter and printed pp. xi–xxxii and 1–12; most chapters and endnotes were unavailable. Credits name Ryan Panchadsaram, a six-person team, six consulting organizations, and many interviewees, so the plan is Doerr-authored but collaborative. Publisher record Official sample, PDF pp. 7–8
• Central argument and reasoning. Six sector solutions and four accelerants translate net zero by 2050 and halfway by 2030 into cost, adoption, policy, innovation, and capital milestones. [investor-stated · strong] [researcher inference · provisional] Doerr permits several “right” plans and acknowledges uncertainty, making revision part of the method. Official sample, printed pp. xvii–xxix
• Evidence, examples, and assumptions. The Fisker/Tesla choice is adverse evidence: Kleiner saw Fisker's designer, luxury-to-mass sequence, outsourced manufacturing, and A123 relationship as risk reducers; conflict prevented backing both; Doerr later calls the choice wrong. [retrospective witness · moderate] Terms, partner votes, dissent, reserves, and the IC memo are absent. Official sample, printed pp. 1–2
• Rule and operational implication. “Go for the gigatons,” then require price-performance parity. [investor-stated · strong] Separately underwrite reachable share, margin, capital intensity, financing path, and durability; societal importance is not venture fitness. [researcher inference · moderate] Official sample, printed pp. xxx–xxxi and 7–9
• Decisions, biography, and attribution. The sample traces a 2006 family challenge, Kleiner's allocation, 2009 advocacy, a 2010 workshop, policy work, and Grove's influence. [investor-stated · moderate] The move from about 10% to nearly half of the portfolio names no fund, date, capital basis, or authority. [investor-stated · provisional] Official sample, printed pp. xi–xv and xxii–xxiv
• Evolution and contradiction. The book integrates management system, climate commitment, failure, and policy into one plan. The 2026 plan materially changes baselines and targets, confirming revisability while preventing simple progress comparison. [documented behavior · strong] 2026 plan

8. Speed & Scale Action Plan (2021)

• Authorship, context, and access. The full official two-page poster is the contemporaneous baseline: ten objectives and 55 KRs. It is unsigned and cannot allocate authorship among Doerr, Panchadsaram, the research team, advisers, or institutions. 2021 action plan, pp. 1–2
• Central argument and reasoning. Six emissions solutions and four accelerants are decomposed into quantities, deadlines, price thresholds, policy, and capital flows. [contemporaneous record · strong] [researcher inference · provisional] The implied measurement-to-ownership link lacks an owner map or review protocol. [researcher inference · moderate]
• Evidence, examples, and assumptions. Transport, grid/heating, agriculture, nature, industry, and removal destinations plus cost thresholds are plan assumptions—not observed facts or validated forecasts. [contemporaneous record · strong] [researcher inference · provisional] 2021 plan, objectives 1–6 and KRs 1.1, 2.2–2.3, 9.2–9.5
• Rule and operational implication. Translate themes into cost, scale, deadline, and capital milestones; distinguish government support, R&D, venture, project finance, and philanthropy. [researcher inference · moderate] Never treat a gigaton target or $50 billion ecosystem flow as company-level ownership or return evidence. [researcher inference · strong] 2021 plan, KRs 10.1–10.5
• Decisions, biography, and attribution. This records what the climate program measured, not a Kleiner or personal portfolio decision. [contemporaneous record · strong] No company, board action, vehicle, return, or exit appears. 2021 plan, p. 2
• Evolution and contradiction. The 2026 update changes 55 KRs to 52, revises baselines, and removes or reframes deadlines including “no new coal or gas plants after 2021.” [documented behavior · strong] Without a changelog, removal can mean failure, better modeling, or scope change. 2021 KR 2.4 2026 plan

9. Prepared statement for the Senate AI Insight Forum (2023)

• Authorship, context, and access. Doerr submitted this complete five-page statement on October 24, 2023. It is primary evidence for policy positions; forecasts, venture statistics, and firm impact are interested and unaudited. Senate statement, pp. 1–5
• Central argument and reasoning. AI is the latest roughly 13-year technology “tsunami.” The U.S. should manage factuality, bias, privacy, and disinformation without surrendering leadership; leadership depends on education, RD&D, talent, government funding, shared compute/data, clusters, and viable IPO/acquisition paths. [investor-stated · strong] [investor-stated · moderate] Senate statement, pp. 1–5
• Evidence, examples, and assumptions. Lung-cancer detection, stroke triage, retinopathy screening, grid resilience, and tutoring illustrate possible value. [investor-stated · moderate] Selected dates make the cadence memorable but not predictive, while recommendations develop innovation capacity more than governance. [researcher inference · strong] Senate statement, pp. 2–4
• Rule and operational implication. Underwrite talent, compute, research, commercialization, regulation, and liquidity—not just applications. [researcher inference · moderate] Independently validate adoption, cost, capability, and durable value before using a narrative cadence as an investment clock. [researcher inference · provisional] Senate statement, pp. 1–5
• Decisions, biography, and attribution. Doerr reports more than forty years with hits and misses and says Kleiner backed 1,000+ companies creating 1.6 million jobs and $4.5 trillion market cap. [investor-stated · provisional] No deduplicated list, vehicle scope, method, or attribution rule exists; named AI examples are not proven investments. Senate statement, p. 1
• Evolution and contradiction. His role expands from company coaching to market-design advocacy. [documented behavior · strong] Acquisition-friendly policy aligns with venture liquidity and omits the concentration/buyer-power countercase; acknowledged AI risks receive less detail than the investment agenda. [researcher inference · strong] Senate statement, “IPOs, Mergers, and Growth,” p. 5

10. Let’s Build, Friends, Build (2026)

• Authorship, context, and access. This complete three-page April 2026 letter carries Doerr's handwritten signature. The text is attributable to him, but its research team and numbered footnote references are absent. Signed letter, pp. 1–3
• Central argument and reasoning. Rising electricity demand, China's clean-manufacturing scale, and falling costs converge; repetition and scale can create a clean discount; therefore builders and policymakers must construct profitable alternatives fast enough to displace fossil fuels. [investor-stated · strong] [investor-stated · moderate] Signed letter, numbered sections 1–3
• Evidence, examples, and assumptions. AI data centers, EVs, Global South development, Chinese exports, fusion, geothermal, storage, and low-carbon steel illustrate the thesis. [investor-stated · moderate] Intermittency, transmission, permitting, minerals, industrial policy, and distributional costs are not modeled. [researcher inference · strong] Signed letter, numbered sections 1–3
• Rule and operational implication. Source where large demand, strategic dependence, and steep cost curves meet; prefer technologies able to reach a clean discount, then test deployment and policy feasibility. [researcher inference · moderate] No company, valuation, ownership, reserve, governance, or exit rule follows. Signed letter, numbered sections 1–3
• Decisions, biography, and attribution. The signature and explicit 2021-plan recast show Doerr revising a public program. [documented behavior · moderate] The technology list is advocacy, not evidence of investments by Doerr, a family vehicle, or Kleiner. Signed letter, closing and signature
• Evolution and contradiction. Emphasis shifts from urgency and targets toward build quantities, price, energy security, and geopolitics. [researcher inference · moderate] China is proof of scale and strategic rival; rising emissions despite falling costs are counterevidence that economics alone suffices. [researcher inference · strong]

11. Speed & Scale: 2026 Action Plan

• Authorship, context, and access. The complete official two-page April 2026 scorecard retains ten objectives, reduces 55 KRs to 52, and adds five statuses. It is unsigned; the companion letter links Doerr to the recast but does not allocate research or ratings. 2026 plan, p. 2 Signed letter
• Central argument and reasoning. Physical counts for vehicles, generation, storage, transmission, and facilities plus unit-price gates and color status should make drift visible and enable correction. [contemporaneous record · strong] [researcher inference · provisional] No formulas, source dates, thresholds, or owners make ratings reproducible. 2026 plan, p. 2
• Evidence, examples, and assumptions. Baselines/destinations materially change: transport 9→2 Gt, grid/heating 30→1, food 8→4, nature 13→3, industry 13→4, and annual removal 14 Gt. [contemporaneous record · strong] [researcher inference · insufficient public record] Unit-price thresholds are market-readiness signals, not company value capture. [researcher inference · moderate] 2026 plan, objectives 1.0–6.0 and KRs 9.1–9.9
• Rule and operational implication. Convert a thesis into physical, price, policy, and capital gates; expose off-track status; demand a predeclared action for misses. The first two are in the plan; the action protocol is missing. [documented behavior · moderate] [researcher inference · strong] 2026 plan, p. 2
• Decisions, biography, and attribution. KR 10.3 targets $50 billion annually in early-stage climate venture while KR 10.4 targets $5 trillion in deployment capital. [contemporaneous record · strong] That is ecosystem architecture, not a Doerr/Kleiner commitment or return claim. 2026 plan, KRs 10.1–10.5
• Evolution and contradiction. The scorecard operationalizes revision but breaks comparability through changed categories, baselines, destinations, and deadlines; several price KRs lack dates, and status colors are unaudited. [researcher inference · strong] It demonstrates willingness to update more strongly than accountable achievement. 2021 plan 2026 plan

Cross-source synthesis

Execution makes ambition legible

Across the Google excerpt, the partial Measure What Matters record, both climate plans, and the 2026 letter, Doerr's most stable belief is that large ambition becomes useful only when decomposed into observable results. [investor-stated · strong] At Google the unit is a quarterly objective; in climate it becomes gigatons, physical deployment, unit prices, policy, and capital flows. WIRED, definitions 2021 plan 2026 plan

The implied causal chain is purpose → priorities → observable milestones → transparent review → corrective action → execution, with leadership, judgment, and culture as prior conditions. [researcher inference · strong] That is more defensible than claiming OKRs caused Google or can coordinate climate alone. WIRED, method boundary Fast Company, plan-is-not-guarantee passage

People, service, and institutional compounding

Doerr consistently assigns building to teams: Google's adoption group, all four DoorDash founders, Kleiner generations, and the Speed & Scale collaboration. [investor-stated · strong] His own intellectual biography is social—Grove for goals, Byers/Caufield for apprenticeship, Campbell for coaching. WIRED DoorDash essay NVCA speech Speed & Scale credits

The operational implication is to evaluate edge as an institutional production function—domain knowledge, network, founders, partners, board service, vehicle, and follow-on support—not personal anecdotes. [researcher inference · strong] DoorDash supports KPCB led / Doerr board member; Google supports a joint Sequoia/KPCB round and two board appointees. DoorDash proxy Google announcement

“Servant VC” is a norm, not evidence of equal power. [researcher inference · moderate] Boards, financing, reserves, executive changes, and exits can put fund and founder interests in tension; the writings provide no conflict protocol. NVCA speech, “Assist entrepreneurs” Chair announcement

Sourcing and picking

The recurring selection pattern is: pursue a consequential problem, identify an execution-oriented team with product insight, and require a route to customer value or price-performance parity. [researcher inference · moderate] Google, DoorDash, and climate's gigaton/clean-discount tests illustrate it. WIRED, pitch passages DoorDash essay Speed & Scale sample, pp. xxx–xxxi

Each element has a boundary. Need is not automatically a venture market; a founder is not a complete team; price parity does not prove margin, capital efficiency, financing, or defensibility. [researcher inference · strong] Fisker warns against mistaking an attractive plan and apparent risk reducers for execution. Speed & Scale sample, pp. 1–2

Portfolio construction, governance, and exits

Mobile-versus-climate and the multi-solution plans support holding different time horizons and treating policy, infrastructure, and market formation as risk. [investor-stated · moderate] They do not reveal concentration, ownership targets, reserves, loss budgets, or stop rules. Fast Company, mobile/cleantech passages 2021 plan

The narrow transferable lesson is to match fund duration and follow-on capacity to development/deployment cycles and precommit continuation/stop evidence. [researcher inference · moderate] The $1b→$3b value, 10%→nearly-half allocation, and 70/100 counts are too incomplete for an efficacy conclusion. Speed & Scale sample, pp. xv and xxii Tim Ferriss Masters in Business

Coaching and transparent review are Doerr's distinctive governance contribution. [investor-stated · strong] The missing half is authority: who sets or changes goals, how scores connect to reserves or executive decisions, and how a player coach yields control. [researcher inference · strong] WIRED Chair announcement

Climate and AI expand underwriting to public R&D, immigration, compute/data, regulation, project finance, philanthropy, grids, industrial policy, IPOs, and acquisitions. [investor-stated · strong] That is useful and interested: venture investors benefit from abundant capital and liquidity, while the Senate statement barely tests concentration or buyer power. [researcher inference · strong] Senate statement, pp. 4–5 2026 plan, KRs 7 and 10

Unsupported conclusions

The corpus does not disclose a pricing doctrine, ownership targets, fund construction, reserve algorithm, liquidation discipline, realized-return series, or complete anti-portfolio. [researcher inference · strong] It does not prove OKRs caused Google, the climate cohort produced a fund return, Doerr alone originated team decisions, or a 13-year cadence predicts timing. The writings are strongest on ambition, execution, coaching, and ecosystems; weakest on venture economics and attribution. Publisher description Senate statement, p. 1

Intellectual chronology, evolution, and contradictions

Three tensions must remain visible:

1. Measurement versus unmeasurable public goals. Specific/verifiable KRs contrast with broad 2016 chair “OKRs” and 2026 items without dates or formulas. [documented behavior · strong] Chair announcement 2026 plan
2. Customer demand versus mission-led entry. Doerr says test whether customers want a technology but also permits urgency to define uncertain-market search. [investor-stated · moderate] The best reconciliation is mission chooses the search space; price-performance and deployment test adoption. [researcher inference · moderate] Fast Company
3. Service/team rhetoric versus concentrated authority and credit. The writings celebrate collaborators while Doerr retains committee roles and narrates successes. [researcher inference · moderate] Map sourcing, approval, vehicle capital, board work, and outcome rather than choosing one narrative. Chair announcement Google announcement

Decision playbook and modern VC translation

1. Make a consequential thesis measurable

• Historical rule. Choose an important objective, attach specific/time-bound/verifiable results, expose them, and review them often enough to act. [investor-stated · strong] WIRED
• Diagnostics. What must be true by the next financing? Which result measures outcome? Who owns it? What action follows green, yellow, or red? What evidence justifies changing the KR rather than the thesis?
• Boundary/failure. Metrics can narrow attention, invite gaming, and conceal weak judgment; goals without owners or consequences become theater. [investor-stated · strong] [researcher inference · strong] WIRED, goal-setting hazards
• Modern VC translation (as of 2026-08-23). What transfers is the visible hypothesis-and-milestone loop. What fails is treating standardized SaaS or AI benchmarks as universal truth. Changed assumptions include cheaper iteration, volatile model cost/capability, and abundant telemetry. Applicability signals are short feedback cycles, controllable drivers, and predeclared board action. Likely misuse is turning stretch goals into forecasts. [researcher inference · strong] Grounded current example: the 2026 climate scorecard exposes status but lacks reproducible formulas and action rules, showing the method's value and incompleteness. [documented behavior · moderate] 2026 plan

2. Let mission choose the search space; let evidence choose the company

• Historical rule. Search consequential problems, then test team, product advantage, demand, price-performance, and deployability. [researcher inference · moderate] WIRED DoorDash essay Speed & Scale sample
• Diagnostics. Is pain large? What is better now? Who pays? Can mission value become a discount? What complementary infrastructure is required? Who wins if this thesis is wrong?
• Boundary/failure. Gigaton importance, social need, or conviction does not prove capture, margin, capital efficiency, financing, or defensibility. Fisker is the counterexample. [retrospective witness · moderate] Speed Scale sample, printed pp. 1–2
• Modern VC translation (as of 2026-08-23). What transfers is mission scale plus customer economics. What fails is using economy-wide TAM or AI/climate urgency as diligence. Changed assumptions include AI electricity demand, industrial policy, grid/permitting bottlenecks, and different hardware/software clocks. Signals are signed customers, falling delivered cost, repeatable deployment, and a financing chain. Likely misuse is calling every company in a favored theme inevitable. [researcher inference · strong] Grounded current example: Doerr's 2026 letter names the system but no winning company. [investor-stated · strong] Signed letter

3. Underwrite the team; coach without taking false credit

• Historical rule. Favor teams with insight, customer commitment, ambition, and execution; help with priorities, recruiting, and scaling while treating operators as builders. [investor-stated · moderate] NVCA speech DoorDash essay
• Diagnostics. Who owns product, engineering, go-to-market, and operations? How does the team disagree and learn? What exact help can this board member supply? Who sourced, approved, negotiated, recruited, and governed?
• Boundary/failure. Charisma hides gaps; “we” hides work; post-outcome essays can turn board presence into authorship. DoorDash's proxy supports a board role and vehicle distinctions, not sole credit. [researcher inference · strong] DoorDash proxy
• Modern VC translation (as of 2026-08-23). What transfers is team-level diligence and a written board-support contract. What fails is assuming one star partner or generic platform fills a capability gap. Changed assumptions include distributed teams and specialized platform functions. Signals are role clarity, evidence-based conflict, hiring throughput, and founder pull. Likely misuse is post hoc credit inflation. [researcher inference · strong] Analyst-created hypothetical: an AI infrastructure memo should name who evaluated model economics, security, distribution, and regulation and preserve that work allocation. [researcher inference · moderate]

4. Use the board scorecard for learning, not automatic control

• Historical rule. Expose goals, review misses, and revise endangered results after judgment; coaching and culture remain indispensable. [investor-stated · strong] WIRED, “Superpower #3” and method boundary
• Diagnostics. Is a miss a bad thesis, execution, metric, or changed environment? What intervention is proportionate? Which right belongs to management, board, IC, or vehicle?
• Boundary/failure. Boards can weaponize metrics or revise endlessly to avoid failure; player-coach succession can obscure authority. [researcher inference · strong] WIRED, goal-setting hazards Chair announcement
• Modern VC translation (as of 2026-08-23). What transfers is separating indicator, interpretation, and action. What fails is dashboard governance without rights. Changed assumptions include real-time data and greater Goodhart risk. Signals are stable definitions, escalation rules, and owners. Likely misuse is treating a metric miss as founder failure. [researcher inference · strong] Analyst-created hypothetical: pair AI revenue retention with gross-margin-per-inference and customer outcome, then define which combination changes pricing, hiring, or follow-on. [researcher inference · moderate]

5. Match technology duration to fund and financing duration

• Historical rule. Climate can require more capital, time, courage, and persistence than conventional venture tolerates; do not finance long physical deployment with a fast-feedback vehicle design. [investor-stated · moderate] Speed & Scale sample, p. xv
• Diagnostics. What technical, commercial, regulatory, and project-finance milestones consume time/cash? Which vehicle owns each stage? What reserves, syndicate capacity, and kill criteria exist?
• Boundary/failure. Persistence can be skill or delayed loss recognition; survivor value proves neither. The $1b→$3b and 70/100 counts cannot distinguish them. [researcher inference · strong] Fast Company, cleantech passage Tim Ferriss transcript, cleantech discussion
• Modern VC translation (as of 2026-08-23). What transfers is duration/capital-stack matching. What fails is collapsing deployment, project finance, philanthropy, and venture. Changed assumptions include larger infrastructure needs and subsidies. Signals are milestone-financing fit, bankable offtake, repeatable construction, and later capital. Likely misuse is waiving stops due to urgency. [researcher inference · strong] Grounded current example: the 2026 plan separates $50 billion early-stage venture from $5 trillion deployment capital. [contemporaneous record · strong] 2026 plan, KRs 10.3–10.4

6. Revise the plan, but preserve the audit trail

• Historical rule. Revise when evidence changes; misses trigger action, not concealment. [investor-stated · strong] WIRED Signed letter
• Diagnostics. Which target was achieved, missed, redefined, merged, or dropped? What changed? Is the old baseline calculable? Who approved revision and what decision followed?
• Boundary/failure. Revision without a bridge erases failure; status without formulas is branding rather than accountability. [researcher inference · strong] 2021 plan 2026 plan
• Modern VC translation (as of 2026-08-23). What transfers is versioned underwriting: retain the original memo and reconcile old/new metrics. What fails is rewriting after every round. Faster markets increase both revision need and audit duty. Signals are a changelog, unchanged historical actuals, and decision-linked revisions. Likely misuse is celebrating flexibility without recording misses. [researcher inference · strong] Grounded current example: the 2021/2026 posters differ without a public itemized bridge. [documented behavior · strong] 2021 plan 2026 plan

7. Underwrite the innovation ecosystem and disclose the investor's incentive

• Historical rule. Waves need education, research, talent, compute/data, public funding, commercialization, clusters, and exits. [investor-stated · strong] Senate statement, pp. 4–5
• Diagnostics. Which complement is scarce? Who funds it? Which rule enables trust or blocks entry? What if acquisitions or IPOs close? Does policy improve innovation broadly or mainly portfolio liquidity?
• Boundary/failure. A narrative cycle is not a timing signal, and pro-M&A policy may favor exits while worsening concentration. [researcher inference · strong] Senate statement, “IPOs, Mergers, and Growth,” p. 5
• Modern VC translation (as of 2026-08-23). What transfers is adding compute, energy, talent, regulation, and exit sensitivity to the memo. What fails is assuming every wave means a funding cycle. Changed assumptions include AI dependence on concentrated compute, power, data rights, and policy. Signals are lower input cost, broader access, durable customer value, and multiple exits. Likely misuse is lobbying for investor-friendly rules as pure public interest. [researcher inference · strong] Grounded current example: Doerr's statement joins AI to public compute, immigration, R&D, clusters, IPOs, and M&A while offering thinner governance detail. [investor-stated · strong] Senate statement

8. Keep people, vehicles, and outcomes in separate ledgers

• Historical rule. Doerr names mentors, partners, founders, collaborators, funds, and a family foundation; the historical unit is not automatically John Doerr. [contemporaneous record · strong] NVCA speech Chair announcement
• Diagnostics. Who sourced, sponsored, approved, negotiated, governed, recruited, financed follow-ons, and narrated? Which vehicle supplied capital? Is ownership nominal, beneficial, proportional, or personal? Is outcome company value, fund value, or cash?
• Boundary/failure. Brand storytelling converts partnership capital and founder execution into personal performance. Google, DoorDash, and cleantech conflicts show why ledgers matter. [researcher inference · strong] Google announcement DoorDash proxy
• Modern VC translation (as of 2026-08-23). What transfers is an attribution ledger beside every case/statistic. What fails is inferring personal returns from a board seat, firm investment, beneficial-ownership row, or market cap. Changed assumptions include multi-stage firms and special-purpose/personal vehicles. Signals are named roles, legal entities, dates, and cash-flow definitions. Likely misuse is ranking by company fame. [researcher inference · strong] Analyst-created hypothetical: write “Fund X invested; Partner A sourced; Partner B diligenced; Partner C joined the board; founders executed; proceeds unknown” until evidence permits more. [researcher inference · strong]

Best writings and unresolved record

Ranked top five

1. When John Doerr Brought a “Gift” to Google's Founders. Clearest mechanism, boundaries, governance cadence, intellectual lineage, and remembered case; it also exposes the Google amount and survivor-attribution problems. [researcher inference · strong] WIRED
2. Speed & Scale official sample. Despite partial access, it joins macro ambition, operating architecture, price-performance, duration, a failure-rich Fisker/Tesla choice, and collaborative authorship. [researcher inference · strong] Official sample
3. Speed & Scale: 2026 Action Plan. Strongest evidence of revision: physical quantities and statuses make the theory more testable, while missing methodology exposes the accountability limit. [researcher inference · strong] 2026 plan
4. Senate AI statement. Fullest ecosystem model—education, talent, compute/data, RD&D, regions, regulation, IPOs, and M&A—and enough incentive visibility for skeptical review. [researcher inference · strong] Senate statement
5. NVCA acceptance speech. Best compact apprenticeship biography and strongest corrective to lone-investor mythology, with differentiation, operating, network, builder, and service rules. [researcher inference · moderate] NVCA speech

Unresolved record

• Partial books. Complete Measure What Matters chapters/notes/negative cases and Speed & Scale beyond printed p. 12 plus pp. 345–396 endnotes remain unavailable. Authorized library or purchased review could change the analysis. Measure What Matters record Speed & Scale record
• Google economics and credit. $11.8m/12%, $12.5m, and $25m aggregate remain distinct; no term sheet, cap table, cash-flow ledger, or vehicle schedule resolves them. Moritz, Sequoia, KPCB, Doerr, founders, advisers, and team remain separate. WIRED Publisher Google
• Cleantech denominator and return. The $1b→$3b assertion cannot be reconciled to funds, companies, dates, realized/unrealized values, write-offs, or follow-ons; direct accounts conflict at about 70 versus 100 ventures. [researcher inference · insufficient public record] Fast Company Tim Ferriss Masters in Business
• Climate audit trail. No public data dictionary, source workbook, KR owner map, status formula, or itemized 2021→2026 changelog was recovered. 2021 plan 2026 plan
• OKR efficacy and failure denominator. Adoption and belief are documented, not causal effect across 50+ organizations. Rejection, abandonment, gaming, implementation cost, and board decisions triggered by KRs remain missing. Publisher WIRED
• Firm and vehicle mechanics. KPCB funds, family foundation, and committees lack opportunity allocation, votes, carry, decision-right, and succession records. DoorDash and Google do not fully allocate sourcing, diligence, board work, reserves, or proceeds. Chair announcement DoorDash proxy Google
• AI interests and safeguards. The Senate statement lacks an investment-interest schedule, named holdings behind examples, and safeguards Doerr would accept if they reduced capability, venture formation, or acquisitions. The 13-year cadence lacks a falsification rule. [researcher inference · strong] Senate statement, complete pp. 1–5

Coverage and evidence base

The spoken-corpus denominator is 15 mapped ★★★ source rows representing 21 unique appearances. Fourteen rows each represent one appearance; the Charlie Rose archive row represents seven unique air dates. Its live index exposes 12 cards, but five full-program cards duplicate five segment cards, so the archive contributes seven appearances—not 12. Charlie Rose guest index, cards 1–12

TED supplied official cue-level transcripts; CNN, Tim Ferriss, Ritholtz, What Matters, Climate One, Edge, Stanford, the White House, GovInfo, and Charlie Rose supplied official or publisher-hosted text. Climate One and Ritholtz warn of transcript defects, while the White House file route proved fragile; those boundaries remain visible in the notes. No audio transcription was necessary because complete publisher or institutional text was available.

Direct voice establishes what Doerr or a named collaborator said, not that the claim was correct, caused an outcome, represented an entire partnership, or generated a fund return. [researcher inference · strong] The corpus is vulnerable to success selection, retrospective compression, collective work collapsed into one public figure, and unaudited economics. Ferriss, Google group-decision and cleantech turns Climate One, Panchadsaram speaker turns

Dated source-by-source guide

1. The Coach — Edge, 1997-01-06

The complete interview is available through Edge's canonical transcript and legacy eight-page Digerati route. The canonical heading dates it January 6, 1997; the earlier 2000 inventory date was wrong. Edge, canonical heading and Doerr turns

Doerr argues that capital and ideas are abundant relative to team construction. He divides opportunities into startups, moving “speedups,” and partner-created “co-ventures,” then uses Netscape recruiting to argue that leadership teams convert platform shifts into companies. [investor-stated · strong] Edge, venture-role and Netscape turns; legacy p. 4

Amazon appears in a contemporaneous transaction-business taxonomy, while Netscape is framed as participation in a growing intranet market rather than a binary browser war. [contemporaneous record · moderate] The interview says my partners and I but does not allocate sourcing, approval, recruiting, board work, or follow-on capital. Edge, Internet-business and Netscape turns; legacy pp. 4–6

2. John Doerr — Charlie Rose, 1999-05-31

The full official transcript was accessible; Doerr’s terminal cue is 10:04; the full-program card duplicates the segment card. Charlie Rose, John Doerr segment

Doerr calls the internet an early networked economy and argues that teams, leadership, character, and education are scarcer than capital or technology. [contemporaneous record · strong] He also predicts always-on broadband reaching more than half of households within two or three years. [contemporaneous record · moderate] Preserve the dated forecast alongside the durable team claim. Charlie Rose, 1:05–5:51

He says he joined Intel in 1974, a date that later conflicts with a 1975 TED scene. [contemporaneous record · provisional] Charlie Rose, 6:32–8:18

3. John Doerr after the dot-com bust — Charlie Rose, 2002-03-18

The full official transcript was accessible; Doerr’s terminal cue is 27:43; the full-program card is a duplicate. Charlie Rose, John Doerr segment

Doerr assigns responsibility for the bubble to excessive venture capital and weak entrepreneurs, acknowledges that Kleiner kept investing over the top, and contrasts that with decisions not to raise a billion-dollar fund or open European offices. He argues that smaller funds and less capital per startup were then necessary for returns. [contemporaneous record · strong] Charlie Rose, 4:28–7:54 and 15:33–16:39

After September 11 he broadens the entrepreneurial agenda to security, urbanization, water, transportation, health, education, and environmental problems. [contemporaneous record · moderate] This predates the later Mary-centered climate origin story, though it does not establish the later investment program. Charlie Rose, 16:39–22:12 and 25:14–27:43

4. Transformation of the Internet — Charlie Rose, 2004-11-16

The full official transcript was accessible; Doerr’s terminal cue is 25:23; its full-program card is a duplicate. Charlie Rose, Transformation of the Internet

Doerr links innovation to public R&D, education, immigration, and enabling policy, while describing wireless services, clean distributed energy, and network security as active investment areas. [contemporaneous record · strong] Charlie Rose, 3:53–6:56 and 16:15–25:23

He favors technical founders near Stanford and Berkeley and predicts China will have more internet users than the United States within five years. [contemporaneous record · moderate] The source shows a sourcing preference and dated forecast but no accepted/rejected funnel or outcome scorecard. Charlie Rose, 17:01–21:27

5. How To Be a Venture Capitalist — Stanford, 2005-02-02

Stanford exposes the complete 0:19 clip and a five-paragraph printable transcript. It is complete as published but cannot stand in for a full lecture. Stanford eCorner, 0:00–0:19 and Read Transcript

Doerr says aspiring venture capitalists should first become strong entrepreneurs or technology-company operators because the company is where the “real action” occurs. [investor-stated · strong] The clip neither defines qualifying operating responsibility nor reconciles the rule with Doerr's own Intel-employee-to-venture-associate path. [researcher inference · strong] Stanford, complete clip and transcript

6. The World Is Flat — Charlie Rose, 2006-01-06

The full official transcript was accessible; Doerr’s terminal cue is 22:30; the full-program card duplicates this segment. Charlie Rose, The World Is Flat

Doerr translates Thomas Friedman's globalization thesis into an innovation agenda spanning education, immigration, R&D, broadband, health care, and energy. [contemporaneous record · strong] Charlie Rose, 0:41–15:01

He calls for 100,000 additional engineers and scientists within four years, U.S. energy independence within five, and 100-megabit connectivity to every home, school, and health center within ten. [contemporaneous record · strong] These are falsifiable targets; the appearance supplies no later scorecard. Charlie Rose, 15:01–20:40

7. Thomas Friedman interview, Doerr as guest host — Charlie Rose, 2006-05-22

The full official transcript was accessible; Doerr’s terminal cue is 51:29. Doerr is the guest host and questioner; Friedman's answers are not Doerr doctrine. Charlie Rose, Thomas Friedman

In his own disclosure, Doerr says Kleiner had quietly invested in clean technologies for five years, had nine energy ventures, and had earmarked at least $100 million from an $800 million fund. [contemporaneous record · strong] Charlie Rose, 28:57–30:52

Five years of activity reaches back to roughly 2001, complicating simplified claims that clean investing began in 2006. [researcher inference · strong] A narrower formal zero-emissions strategy might explain the difference, but no opened record reconciles it. Charlie Rose, 28:57–30:52

8. Alternative Energy Panel — Charlie Rose, 2006-12-29

The full official transcript was accessible; Doerr’s terminal cue is 27:16; its full-program card is a duplicate. Charlie Rose, Alternative Energy Panel

Doerr says energy transition requires policy as well as innovation, favors cap-and-trade over a politically infeasible carbon tax, and argues that backed companies must offer disruptive cost and emissions performance. [contemporaneous record · strong] Charlie Rose, 1:59 and 13:33–17:05

He says companies should work with oil at both $60 and $40 per barrel. [investor-stated · strong] This is a rare pre-outcome investment test; the public record does not show how companies were scored or whether capital stopped after failure. Charlie Rose, 16:28–17:05

9. Salvation (and profit) in greentech — TED, 2007-03-03

TED provides the complete 17:32 video and official cue-level transcript. TED, full talk and transcript

Doerr moves from daughter Mary's challenge to a partnership learning effort involving about fifty scientists, policy advocates, entrepreneurs, and business leaders, then organizes the response into companies, individuals, policy, and radical innovation. [investor-stated · strong] TED, 0:30–2:27

Bob Epstein receives credit for the AB32 lobbying priority; Berkeley researchers and Gates Foundation funding receive credit for the artemisinin platform later associated with Amyris. Doerr says KPCB was investing $200 million across green technologies. [contemporaneous record · moderate] The amount is an announced firm program, not deployed cost or a personal check. TED, 6:13–7:57 and 9:36–11:53

10. Global Leadership — Charlie Rose, 2008-10-20

The full official transcript was accessible; Doerr’s terminal cue is 49:21. Doerr has only five turns, so attribution is confined to those turns. Charlie Rose, Global Leadership

During the financial crisis, Doerr distinguishes low-capital internet startups, longer-duration biotechnology, and capital-intensive green ventures. He expects tighter funding and credit but advises small companies to preserve R&D and accelerate products where possible. [contemporaneous record · strong] Charlie Rose, 17:34–26:48

His assertion that ideas are abundant while execution, communication, character, and teams are scarce repeats the 1997 and 1999 framework before the later books. [contemporaneous record · strong] Charlie Rose, 49:21

11. U.S. clean-energy competitiveness hearing — Senate, 2009-07-16

GovInfo supplies the complete official hearing record, including Doerr's oral and prepared statements and adversarial questioning. GovInfo, oral statement printed pp. 21–24 and prepared statement pp. 24–44

Doerr frames climate policy as industrial competition: a carbon cap and price should create durable demand, while utility incentives, efficiency standards, and research/development/deployment funding accelerate the response. [investor-stated · strong] GovInfo, printed pp. 21–24

Under Senator Lamar Alexander's questioning, Doerr calls nuclear necessary and carbon-free but not renewable, supports declining wind/solar/battery incentives, and says government should not pick winners and losers. [contemporaneous record · strong] The selection and sunset rules remain undefined. GovInfo, printed pp. 208–212

12. President’s Economic Recovery Advisory Board — White House, 2009-11-02

The complete 14-page transcript was recovered through the official White House meeting index; the mapped direct PDF is fragile rather than content-incomplete. White House, official meeting index

Doerr proposes Cash for Caulkers: recruit retailers and unemployed construction workers to scale residential retrofits, while retaining a carbon cap and price as the stronger long-run signal for private capital. [investor-stated · strong] White House transcript, printed pp. 5–6

He forecasts hundreds of thousands or as many as one million jobs within a year, but President Obama assigns the segment jointly to Doerr and Mark Gallogly and Doerr refers to committee outreach. [contemporaneous record · moderate] The forecast lacks a spending baseline, additionality model, verification regime, and later outcome in this route. White House transcript, pp. 2 and 5–6

13. Why the secret to success is setting the right goals — TED, 2018-04-10

TED provides the complete 11:41 video and official cue-level transcript. TED, full talk and transcript

Doerr sequences purpose before measurement: a compelling why supports an objective, measurable key results define completion, transparency and grading create commitment, and culture and leadership remain prerequisites. [investor-stated · strong] TED, 1:27–3:07, 5:13–7:17, and 9:32–9:55

Nuna's purpose belongs to Jini Kim, ONE's risk culture is voiced by Bono, Chrome's metric belongs to Sundar Pichai, and OKR invention belongs to Andy Grove. [retrospective witness · moderate] Doerr is transmitter rather than sole originator or operator. TED, 3:11–9:28 and 10:51–10:57

The 1975 Intel scene conflicts with Doerr's 1974 account elsewhere, and selected successes do not establish OKR efficacy across failures or ordinary adopters. [researcher inference · strong] TED, 1:01–1:35 Charlie Rose 1999, 6:32–8:18

14. Fareed Zakaria GPS — CNN, 2018-05-20

CNN supplies the complete official broadcast transcript, with speaker turns but no stable per-turn timestamps. CNN, final Zakaria/Doerr interview block

Doerr's five investment screens are technical excellence, ability to build a team, a large unserved market need, reasonable financing, and speed against incumbents. [investor-stated · strong] CNN, Doerr turn beginning “Well, first and foremost”

He describes the iPhone as a platform important enough for a $100 million dedicated fund, while offering risk-taker and reckless as alternatives to a pure skill narrative. [retrospective witness · moderate] The transcript supplies no vehicle name, company list, ownership, or return. CNN, iPhone-fund and luck exchanges

His privacy answer calls trust essential but defines no consent, minimization, enforcement, or pass threshold. [investor-stated · provisional] CNN, opening privacy exchange

15. How to decarbonize the grid and electrify everything — TED, 2020-10

TED supplies the complete 20:17 video and official cue-level transcript. TED, full talk and transcript

Hal Harvey—not Doerr—supplies the grid/electrification shortcut, follow the tons, regulator targeting, and continuous-improvement logic. Doerr supplies the capital, commercialization, and OKR campaign frame. [contemporaneous record · strong] TED, Harvey 1:03–2:46 and 9:34–13:35; Doerr 4:38–5:46 and 17:24–18:15

Doerr argues that physical climate companies require explicit underwriting of technical risk, cost, capital intensity, deployment time, policy-created demand, and later financing. [investor-stated · strong] TED, 3:03–5:46

The 55-billion-ton baseline later becomes 59 and then 74 after a dataset change; these figures cannot be treated as one like-for-like series. [contemporaneous record · strong] TED, 0:18–0:57 Climate One, data-source exchange

16. An action plan for solving the climate crisis — TED, 2021-10

TED supplies the complete 11:28 video and official cue-level transcript. TED, full talk and transcript

Doerr defines six climate objectives; Ryan Panchadsaram supplies four accelerants, example key results, technology-readiness estimates, and much of the measurement architecture. [contemporaneous record · strong] The plan is collaborative, not a Doerr-only operating system. TED, Doerr 0:16–1:25; Panchadsaram 1:31–3:39 and 5:18–7:16

Panchadsaram predicts EV price-performance superiority by 2024 and all newly purchased buses becoming electric by 2025; the 2026 review says the bus deadline was missed and moved to 2030. [contemporaneous record · strong] TED, 3:00–3:39 Climate One, 15:30 deadline exchange

Doerr shifts the unit of climate action from personal behavior to institutions, workplaces, and politics. [investor-stated · strong] The talk does not show that company-style OKRs create accountability among sovereign actors without shared authority. TED, 7:38–10:24

17. The Tim Ferriss Show #543 — 2021-11-04

The publisher provides the complete transcript but warns it may contain typos; evidence is anchored to named question/answer blocks. Tim Ferriss, full transcript

Doerr remembers founder ambition, better search, rapid usage growth, and a large internet thesis, while calling Google a controversial partnership decision with co-sponsors. [retrospective witness · strong] The source rejects unilateral Doerr authority but supplies no IC record, term sheet, or sponsor-role map. Ferriss, Google decision exchanges

Doerr reports roughly $1 billion invested from 2006 across 100 companies, most failures, with remaining value around $3 billion; he says physical ventures need more time and capital and decisive risks should be retired early. [retrospective witness · provisional] Ferriss, “favorite failure” exchange

Ritholtz records roughly 70 ventures for approximately the same capital and value. Different scopes or recall may explain the conflict, but no hit rate or fund return is computable. [retrospective witness · insufficient public record] Ritholtz, renewable-energy exchange

18. Masters in Business — page dated 2021-12-05

Ritholtz supplies the complete edited Bloomberg Radio transcript. The page is dated December 5, not the prior inventory's December 7; obvious editing/ASR defects remain a caveat. Ritholtz, page header and transcript

Doerr says meaningful pivots occur frequently, making the team's adaptive capacity more important than the static initial plan. [investor-stated · strong] Ritholtz, pivot and final venture-career questions

He says Kleiner favored Fisker's experienced automobile designer over Musk's software-first reinvention and made the wrong call. [retrospective witness · strong] We leaves sponsors, dissent, votes, terms, vehicles, and board responsibility unresolved. Ritholtz, memorable failures question

His Silicon Compilers CEO, 18-month Sun operator, and roughly 70 climate-venture claims require corporate records and a vehicle ledger before being treated as settled biography or economics. [retrospective witness · provisional] Ritholtz, Silicon Compilers/Sun and renewable-energy exchanges

19. Speed & Scale’s Reality Check — Climate One, recorded 2026-05-01

Climate One provides a complete indexed transcript and warns that automated and human transcription may contain errors. The composite gives Doerr a short opening exchange; Panchadsaram and others supply separate segments. Climate One, guest list, index, and transcript

Doerr concedes that progress occurred but the world is not on track, and says healthy OKRs should be at least 60% self-declared rather than wholly imposed. [investor-stated · strong] The latter is an authority boundary when moving company controls into public systems. Climate One, opening Doerr turns

Panchadsaram—not Doerr—explains the 52-KR scorecard, shift from a 59-billion-ton UN baseline to 74 billion tons using Climate TRACE, funding changes, and moved bus deadline. [contemporaneous record · strong] Improved observation cannot be read as a simple 15-gigaton same-method deterioration. Climate One, 05:30 onward and 15:30 deadline exchange

The Mary story reaches a roughly 20-year checkpoint without triumph; Doerr says she remains angry. [investor-stated · strong] Climate One, opening Mary exchange

20. OKRs Explained, with John Doerr — What Matters, undated

The route contains a complete 4:01 embedded course video and official transcript. It is promotional and undated, so its headcount and title claims are valid only for an unknown recording period. What Matters, embedded video and Transcript

Doerr assigns the original Intel system to Andy Grove and himself the messenger role; he presents visible goals, quarterly grading, and separation from bonuses and promotions as a social contract around priorities. [investor-stated · strong] What Matters, Intel and Google transcript paragraphs

The course dates the Google pitch to 1998, while Ferriss dates the first session to 1999. A pitch and first session may differ, but no opened contemporary record proves it. [retrospective witness · provisional] What Matters, Google paragraph Ferriss, date clarification

21. Keep Up the Momentum: John Doerr’s OKR tips — What Matters, undated

The route contains a complete 2:26 embedded course video and official transcript. What Matters, embedded video and Transcript

Doerr says no single implementation is correct: preserve what worked, diagnose failure, adapt to culture, allow time for skill development, and revise as the organization changes. [investor-stated · strong] What Matters, iteration paragraphs and numbered list

He credits Ryan Panchadsaram alongside himself for portfolio-company guidance; later climate appearances visibly confirm Panchadsaram's substantive metric and revision role. [investor-stated · strong] What Matters, portfolio-company paragraph Climate One, Panchadsaram turns

Stay the course can conflict with adaptation; the lesson provides no kill criterion for abandoning a failed goal system. [researcher inference · strong] What Matters, persistence and adaptation paragraphs

Recurring stories and operating frameworks

Teams are the scarce input; capital is not the product

Doerr's most durable spoken thesis is that ideas and money are abundant relative to leaders who can recruit and execute. It appears in 1997 Netscape recruiting, the 1999 networked-economy interview, the 2008 crisis panel, the 2018 five-screen framework, and the 2021 pivot discussion. [investor-stated · strong] Edge, Netscape turn Charlie Rose, 1999 at 4:54 Charlie Rose, 2008 at 49:21 CNN, five-factor answer Ritholtz, pivot question

The causal chain he proposes is technical or market shift → exceptional founder/team → recruiting and operating help → adaptation → execution. [investor-stated · moderate] The corpus does not provide the rejected-team denominator, false positives, or a controlled comparison with equally strong teams in weaker markets. [researcher inference · strong] Edge, venture-role turn Ritholtz, pivot discussion

VC use: write the team thesis as capabilities, evidence, missing roles, and adaptation behavior, then separately record who sourced, recruited, governed, and executed. [researcher inference · strong] Edge, Netscape attribution

Find a platform wave, then test the company inside it

The corpus repeatedly searches enabling waves: microprocessors, internet transactions, broadband, mobile/open systems, the iPhone, synthetic biology, and clean-energy infrastructure. [investor-stated · moderate] Edge, Internet and enabling-technology turns Charlie Rose, 2004 at 16:15–21:27 CNN, iPhone-fund exchange TED 2007, Amyris at 9:36–11:17

The stronger rule is not big wave = invest. Google combines product superiority, usage acceleration, founder ambition, and a large internet thesis; climate adds delivered cost, capital duration, technical risk, policy demand, and a later financing stack. [investor-stated · strong] Ferriss, Google decision turns TED 2020, 3:03–5:46

VC use: distinguish platform thesis from company thesis and declare what would show that a company captures value rather than merely participating in an important transition. [researcher inference · strong] Ferriss, Google rationale

Operating discipline travels through apprenticeship

Doerr consistently assigns OKR provenance to Andy Grove and presents himself as transmitter. The broader apprenticeship chain includes Intel operating work, company formation, board support, and later public codification. [retrospective witness · strong] TED 2018, 1:01–2:04 What Matters, Intel paragraphs Ferriss, Grove questions

The usable cycle is purpose → objective → measurable results → visible owner → review → learning or revision. [investor-stated · strong] Judgment, culture, and leadership remain prerequisites; measurement is not a substitute. TED 2018, 5:13–9:55

VC use: require a board scorecard to name the decision triggered by each result. Without an action rule, a metric becomes reporting theater. [researcher inference · strong] TED, OKR method and boundary

Networked diligence precedes system intervention

In climate and policy work, Doerr repeatedly assembles scientists, operators, policy specialists, institutions, and capital providers before reducing a problem to intervention levers. [investor-stated · strong] TED 2007, 1:29–2:27 White House, pp. 5–6 TED 2020, collaborator exchange

The network supplies domain knowledge, public-policy channels, distribution, research, financing, and accountability. [researcher inference · moderate] It risks elite-network confirmation bias unless dissent, affected stakeholders, and alternative channels are deliberately included. TED, networked climate-learning passage

Match risk retirement and capital duration to the technology

The 2006 oil-price threshold, 2008 sector-duration distinctions, and 2021 cleantech postmortems converge on one rule: physical ventures require earlier attack on decisive technical/commercial risks and a financing chain long enough to reach deployment. [investor-stated · strong] Charlie Rose, 2006 at 16:28–17:05 Charlie Rose, 2008 at 17:34–26:48 Ferriss, cleantech failure lesson Ritholtz, venture-capital role question

The corpus does not show reserve rules, stop decisions, or vehicle cash flows. [researcher inference · insufficient public record] Persistence can represent informed duration matching or delayed loss recognition; the absent ledger prevents distinguishing them. Ferriss, cleantech failure account Ritholtz, climate-finance boundary

Public policy is market formation—and a conflict surface

Doerr treats carbon pricing, efficiency rules, research funding, immigration, education, infrastructure, and deployment policy as innovation inputs. [investor-stated · strong] Charlie Rose, 2004 Senate hearing, printed pp. 21–24 White House transcript, pp. 5–6

Because Kleiner invested in affected sectors, advocacy also creates an incentive and disclosure problem. [researcher inference · strong] No talk supplies an affected-investment schedule, recusal rule, subsidy-incidence analysis, or test separating public value from investor benefit. Senate, Doerr testimony and QA

Intellectual evolution and biographical turning points

1974–1999 — operator apprenticeship becomes a people-and-platform thesis

The earliest interviews connect engineering, Intel operating work, KPCB apprenticeship, recruiting, and internet-platform interpretation. [retrospective witness · moderate] Edge, opening and venture-role turns Charlie Rose, 1999 at 6:32–8:18

The strongest continuity is the claim that teams and execution are scarcer than capital. The Intel chronology remains unresolved: 1999 direct speech says 1974, while the 2018 TED scene says 1975. [contemporaneous record · provisional] Charlie Rose, 1999 TED 2018, 1:01–1:35

2000–2004 — boom becomes a capital-discipline lesson

The 2002 interview is a genuine update: Doerr criticizes excess venture capital, acknowledges investing over the top, and argues for smaller funds and lower company burn. [contemporaneous record · strong] Charlie Rose, 2002 at 4:28–7:54 and 15:33–16:39

By 2004, the platform thesis broadens into a national innovation system of universities, immigration, R&D, connectivity, and policy. [contemporaneous record · moderate] Charlie Rose, 2004 at 3:53–6:56 and 16:15–25:23

2005–2009 — founder service and climate-market formation converge

Stanford makes operator experience the preferred route into venture. The Charlie archive shows clean-technology investing by roughly 2001, nine ventures by May 2006, and an oil-price underwriting threshold by December 2006; TED then supplies the family catalyst and four-lever public narrative. [contemporaneous record · strong] Stanford Charlie Rose, May 2006 Charlie Rose, December 2006 TED 2007

Mary's challenge may have intensified or personalized the strategy, but it did not mark the first public evidence of environmental interest or clean investment. [researcher inference · strong] Charlie Rose, May 2006 disclosure TED, 2007 catalyst account

2010–2018 — operating doctrine becomes public intellectual property

By 2018, Doerr turns Grove's management system into a public OKR curriculum and merges it with a five-factor venture screen. [investor-stated · strong] TED 2018 CNN, five-factor answer

The evolution adds purpose, transparency, stretch, and separation from compensation, but the public record lacks failed-adoption and abandoned-system denominators. [researcher inference · strong] TED, OKR cases and boundary

2020–2026 — climate ambition becomes a versioned scorecard

Later talks shift from opportunity rhetoric to physical quantities, sector objectives, KRs, capital-stack dependencies, and public scorekeeping. [documented behavior · strong] TED 2020 TED 2021 Climate One 2026

The evolution also exposes the method's hardest problem: baselines change, deadlines move, and much technical revision belongs to collaborators. Doerr's off-track admission is accountability evidence, not proof that the original deadlines or methodology were adequate. [researcher inference · strong] Climate One, revision and off-track assessment

Deal rationales and decision evidence

Google

Doerr's retrospective rationale combines superior beta search, rapid usage growth, founder ambition, and a non-consensus market view because Google was one of many search engines. [retrospective witness · strong] Ferriss, Google decision turns

The decision was a partnership action with co-sponsors, not a unilateral Doerr bet. [retrospective witness · strong] The talks do not resolve the firm check, vehicle, ownership bridge, IC vote, rejected alternatives, or cash flows. Ferriss, group-decision turn

The OKR chronology remains split between a 1998 pitch and a 1999 first session. [retrospective witness · provisional] What Matters, Google paragraph Ferriss, date clarification

Netscape

In 1997 Doerr emphasizes partnership-assisted recruiting and a market larger than a binary Microsoft-versus-Netscape contest. [contemporaneous record · moderate] Edge, Netscape turns

The case supports team construction and market-expansion reasoning near the decision period. It does not prove enduring competitive defense or allocate recruiting, governance, and financing among partners and company leaders. [researcher inference · strong] Edge, Netscape turns

Amazon

Amazon appears in the 1997 transaction-business taxonomy, when Doerr says the company was sacrificing current profit to invest in brand. [contemporaneous record · moderate] Edge, Internet-business turn

Later interviews celebrate Amazon's founders and market scale, but the spoken corpus does not supply the term sheet, partnership vote, sourcing route, board interventions, dilution, or realized fund economics. [retrospective witness · insufficient public record] Ritholtz, memorable startups question

Amyris and platform transfer

The 2007 TED talk treats antimalarial synthetic biology as a validated production platform that might transfer into biofuels through Amyris. [contemporaneous record · moderate] TED, 9:36–11:17

This is useful ex-ante evidence because it captures the analogy before the outcome. It does not disclose cost curves, commercialization milestones, financing requirements, governance, or a stop rule. [researcher inference · strong] TED, Amyris thesis

Fisker versus Tesla

Doerr later says Kleiner favored Fisker's experienced automobile designer over Musk's software-first reinvention and made the wrong call. [retrospective witness · strong] Ritholtz, memorable failures question

The case warns against treating an industry résumé as equivalent to systems architecture or adaptive execution. [researcher inference · moderate] It remains incomplete because we does not identify vehicles, sponsors, dissent, terms, timing, milestones, board roles, or the counterfactual feasibility of investing in Tesla. Ritholtz, Fisker/Tesla exchange

The iPhone platform vehicle

Doerr says he treated the iPhone as the world's most important platform and created a $100 million fund around it. [retrospective witness · provisional] CNN, iPhone-fund exchange

The appearance supports a dedicated-capital response to a platform shift, but not its economics. Vehicle name, commitments versus deployment, portfolio, partner attribution, ownership, reserves, distributions, and net return remain absent. [researcher inference · insufficient public record] CNN, iPhone-fund exchange

Contemporaneous evidence versus retrospective narrative

Forecast and contradiction scorecard

This table scores only what the processed spoken corpus can support as of 2026-08-23. Not evaluated means a fixed-definition outcome source was not processed; it does not mean success or failure.

The scorecard's lesson is not to stop forecasting. Deadline, definition, dataset, owner, and revision history must travel with the forecast. [researcher inference · strong] The corpus shows that memorable ambition without a versioned audit trail is hard to learn from. Climate One, deadline and baseline revisions

Criticisms, failures, and unresolved record

Success selection and hindsight

Google, Amazon, Netscape, Chrome, Nuna, ONE, and climate narratives are educationally rich but selected. The corpus contains no complete accepted/rejected opportunity set, ordinary-company sample, failed-OKR inventory, or performance comparison for the five screens. [researcher inference · strong] Repetition shows consistency of narrative, not causal efficacy. TED, selected OKR cases Ritholtz, memorable startups

Capital discipline versus cleantech persistence

Doerr's 2002 rule says excess capital lowers standards; his later cleantech account describes roughly $1 billion across a failure-heavy cohort. [researcher inference · strong] Charlie Rose, 2002 Ferriss Ritholtz

The record does not show whether the later program violated the earlier rule, correctly matched long-duration risk, or became concentrated in a few survivors. Vehicle cash flows, reserves, write-offs, and stop decisions are required. [researcher inference · insufficient public record] Ferriss, cleantech account Ritholtz, climate account

Metric discipline versus revision freedom

Doerr says goals should be measurable, transparent, graded, and revised through learning. The bus deadline moved from 2025 to 2030, emissions baselines changed, and the KR inventory changed from 55 to 52. [documented behavior · strong] TED 2021 Climate One

Revision may be rational; revision without an itemized bridge weakens accountability. [researcher inference · strong] The missing artifact is a ledger showing old target, result, reason, new target, owner, and decision consequence. Climate One, bus deadline and data-source changes

Policy advocacy and portfolio incentives

The talks place Doerr in clean-energy investing while he advocates carbon prices, subsidies, R&D, deployment programs, utility rules, and market-making policy. [contemporaneous record · strong] Charlie Rose, May and December 2006 Senate hearing

No processed appearance supplies an affected-investment schedule, recusal standard, or distributional analysis. [researcher inference · strong] Policy can create public value and private gains simultaneously; both require disclosure. Senate, testimony and QA

Collaborator and partnership attribution

Doerr credits partners, founders, Grove, Epstein, Harvey, Panchadsaram, public officials, researchers, and operators. Yet retrospective narratives compress collective work into we or one public investor. [researcher inference · strong] Edge TED 2007 TED 2020 Climate One 2026

The unresolved ledger must identify sourcing, sponsorship, diligence, approval, negotiation, board work, recruiting, follow-ons, exit influence, capital vehicle, and later narration by deal. [researcher inference · strong] Ferriss, partnership and cleantech attribution gaps

Highest-value retrieval gaps

• Original Google OKR pitch/session material resolving 1998 versus 1999.
• KPCB IC memos, term sheets, votes, partner roles, reserve decisions, and distributions for Google, Amazon, Netscape, Amyris, Fisker, and cleantech.
• A company-by-company bridge reconciling nine energy ventures in 2006, roughly 70 ventures, and roughly 100 companies.
• A complete list and outcome ledger for the iPhone platform fund.
• The Rice 2000 VHS and any dedicated Doerr oral history or public personal-papers collection.
• A versioned 2020–2026 climate data dictionary and KR changelog.
• A documented exit, realization, and loss-recognition discipline; the spoken record is rich on sourcing and execution but thin on selling and distributions.

Listen or read these five first

1. The Coach — Edge, 1997. Best early, pre-celebrity statement of the venture role, team construction, platform markets, partnership attribution, and contemporaneous Amazon/Netscape reasoning. [researcher inference · strong] Edge
2. John Doerr after the dot-com bust — Charlie Rose, 2002. Best evidence of a real update after failure: excessive capital, investing over the top, smaller funds, lower company burn, and a broadened public-purpose agenda. [researcher inference · strong] Charlie Rose
3. U.S. clean-energy competitiveness hearing — Senate, 2009. Best adversarial source: questioning forces nuclear, subsidy, carbon-price, deployment, and conflict boundaries that prepared speeches leave vague. [researcher inference · strong] GovInfo
4. The Tim Ferriss Show #543 — 2021. Best long-form synthesis of Grove, Google's partnership decision, founder/product evidence, the 100-company account, and early-risk retirement; pair it with Ritholtz because denominators conflict. [researcher inference · strong] Ferriss
5. Speed & Scale's Reality Check — Climate One, 2026. Best late counterweight to triumphal retrospection: the plan is off-track, a deadline moved, baselines changed, and Panchadsaram visibly owns technical revision. [researcher inference · strong] Climate One

A close sixth is Masters in Business for the Fisker/Tesla error, roughly 70-venture denominator, and detailed operating biography. [researcher inference · strong]

Model 1 — Let mission define the search space; let evidence select the company

• Rule: Start with a consequential problem, but do not let consequence substitute for underwriting. Mission determines where to search; team quality, product advantage, customer pull, delivered economics, financing, and deployability determine whether to invest. [researcher inference · strong] Speed & Scale sample, printed pp. xxx–xxxi and 7–9 CNN, five-screen answer
• Evidence status: This is a reconstructed selection model, not a verbatim Doerr checklist. The mission gate is directly stated in his climate work; the company gate is directly stated in his five investment screens and retrospectively illustrated by Google, DoorDash, and Fisker. [researcher inference · strong] TED, climate underwriting at 3:03–5:46 Ferriss, Google decision exchanges
• Provenance and origin: The model combines Andy Grove's insistence on consequential objectives with Doerr's later climate rule to pursue gigaton-scale problems and his older venture emphasis on large, unmet markets. Grove owns the OKR lineage; Ryan Panchadsaram and the Speed & Scale collaborators materially shaped the climate system. [investor-stated · strong] TED, Grove attribution at 1:01–2:04 Speed & Scale sample, collaborators and printed pp. xxii–xxxi
• Mechanism and reasoning: A mission gate directs scarce attention toward large externalities or unmet needs; a separate evidence gate prevents thematic enthusiasm from becoming a company conclusion. The second gate asks whether this team can deliver a meaningfully better product at a price customers will accept through a financeable path. [researcher inference · strong] DoorDash, customer and local-commerce thesis Speed & Scale sample, price-performance discussion
• Historical and vehicle context: Doerr's 2007 climate account describes a KPCB learning and investment program, while later Speed & Scale activity includes personal advocacy and philanthropy. Neither makes every climate company a Doerr investment or establishes a single vehicle, check, or return denominator. [researcher inference · strong] TED, KPCB program at 1:29–2:27 and 9:36–11:53 Speed & Scale sample, printed pp. xi–xv
• When to use: Use the model before a sector sprint, thesis fund, policy-linked theme, or company screen where urgency can overwhelm selectivity. It is especially useful when the social TAM is far larger than the startup's capturable market or when customer willingness to pay differs from public value. [researcher inference · strong]
• How to apply it: Write two memos. Gate A names the consequential outcome, affected users, system bottleneck, and why venture is appropriate. Gate B names decision-date product evidence, buyer, budget, unit economics, team gaps, financing chain, and alternatives. Kill the company thesis—not necessarily the mission—if no buyer has urgent pull, delivered cost cannot approach the incumbent, deployment cannot repeat, or the next financing cannot retire a named risk. [researcher inference · strong]
• Examples:
  • Documented application — Google: In 1999 Google claimed a relevance advantage, a 100-million-query daily market, and 50% monthly traffic growth; later Doerr testimony adds founder ambition and product superiority. These facts support a company screen inside the internet thesis, not hindsight use of the advertising model. [contemporaneous record · moderate] Google announcement Ferriss, Google exchanges
  • Documented application — DoorDash: The 2015 record exposed a crowded delivery market, nine operating markets, and a proposed logistics layer for local commerce; Doerr's later essay emphasizes merchants' concrete need. The signed retrospective is interested evidence and does not recover the IC memo. [researcher inference · moderate] Fortune, March 2015 Doerr essay
  • Counterexample — Fisker versus Tesla: Climate importance and an experienced auto designer appeared to reduce risk, yet Doerr later said Kleiner selected the wrong architecture and team. Mission value did not prove company capture or execution. [retrospective witness · strong] Speed & Scale sample, printed pp. 1–2 Ritholtz, memorable failures question
• Evidence: Four independent routes support the two-gate reconstruction: the official Google announcement supplies contemporaneous product and demand claims; CNN supplies Doerr's five screens; the official Speed & Scale sample supplies mission and price-performance; and the Fisker retrospective supplies a negative case. None supplies the rejected-company denominator. [researcher inference · strong] Google CNN Speed & Scale Ritholtz
• Limits, failure modes, and exception ledger: Likely misuse is mission-washing: converting an economy-wide problem into a startup TAM, treating policy support as durable demand, or waiving price and financing tests. Observable warnings are persistent green premiums, pilots without conversion, subsidy concentration, bespoke deployments, and an undefined next capital provider. No exception is granted for urgency; the public record lacks the full accepted/rejected funnel and must not be backfilled from winners. [researcher inference · strong]
• Evolution and contradictions: The early corpus emphasizes markets and platform shifts; the 2007–2026 corpus makes climate mission explicit. Doerr's later admission that most climate ventures failed contradicts any simple important problem = good investment reading and strengthens the evidence gate. [retrospective witness · strong] Edge, market and venture-role turns Ferriss, cleantech failure exchange
• Modern VC translation (as of 2026-08-23): For AI, climate, health, or defense, keep the public-value thesis and the company-value thesis on separate pages. Applicability signals are a measurable user outcome, repeat purchase, improving delivered cost, repeatable deployment, and a credible financing chain; likely misuse is calling every participant in a favored theme inevitable. Analyst-created hypothetical: an AI-grid startup passes Gate A by addressing power scarcity but fails Gate B if savings disappear after inference, interconnection, and integration costs. [researcher inference · strong]

Model 2 — Underwrite adaptive teams, then coach without taking false credit

• Rule: Treat adaptive team quality as the scarce input, then define the investor's contribution as recruiting, prioritization, governance, and service—not authorship of founder execution. [investor-stated · strong] Edge, venture-role and Netscape turns Charlie Rose, 1999 at 2:08–4:54
• Evidence status: The people-first rule is repeated contemporaneously from 1997 through 2018 and retrospectively in 2021. Claims about particular recruiting contributions are less complete because sponsor accounts rarely expose other partners, rejected candidates, founder agency, or marginal causality. [investor-stated · strong] CNN, five-factor answer Ritholtz, pivot and final venture-career questions
• Provenance and origin: Doerr links this model to operating apprenticeship at Intel and to KPCB's company-building culture under Eugene Kleiner and Tom Perkins. His later NVCA speech credits partners, founders, recruiters, and mentors; it does not support a lone-coach mythology. [retrospective witness · moderate] NVCA acceptance speech Ritholtz, Intel and KPCB entry exchanges
• Mechanism and reasoning: Plans change when markets, products, and technologies change; a team with honest disagreement, learning speed, role coverage, and recruiting ability can adapt. A useful investor supplies missing networks or judgment while leaving operating ownership with management. [investor-stated · strong] Ritholtz, pivot question Stanford, printable transcript p. 1
• Historical and vehicle context: KPCB, KPCB nominee entities, DGF vehicles, co-investors, and founders are separate actors. A Doerr board seat documents governance access; it does not prove sourcing, approval, check ownership, recruiting causality, or personal economics. [documented behavior · strong] Google S-1, management and principal holders DoorDash S-1, board and securities
• When to use: Use this model when the initial plan is likely to pivot, executive hiring is incomplete, or the board believes a famous founder or investor can substitute for a complete organization. It is strongest when team behavior can be observed under disagreement and weak evidence, not inferred from résumé prestige. [researcher inference · strong]
• How to apply it: Map product, engineering, go-to-market, finance, and operations owners; record one observed adaptation; test whether dissent changes decisions; name the recruit or intervention the investor can uniquely help deliver. Kill or reprice when the team conceals bad news, cannot recruit its missing complement, repeats a failed experiment without new evidence, or depends on the investor to operate the company. [researcher inference · strong]
• Examples:
  • Documented application — Netscape: Doerr's 1997 account uses Netscape team construction and executive recruiting to explain the venture role; later records preserve Clark, Andreessen, Joy, Barksdale, Doerr, and the partnership as distinct contributors. [retrospective witness · moderate] Edge, Netscape turn New Yorker, team and network account
  • Documented application — DoorDash: Doerr joined the board in March 2015, but Xu, Fang, Tang, and Moore remained the operating founders and DGF entities held the securities. This supports coaching and governance, not personal authorship of execution. [documented behavior · strong] DoorDash S-1 DoorDash proxy, footnote 7
  • Counterexample — Fisker: Kleiner favored an experienced automobile designer over a software-first system builder, and Doerr later called the choice wrong. Prestige and domain résumé were not sufficient evidence of adaptive system execution. [retrospective witness · strong] Ritholtz, memorable failures question
• Evidence: Edge and Charlie Rose provide contemporaneous repetition of the scarce-team thesis; Ritholtz supplies the pivot mechanism and adverse Fisker case; SEC filings bound Doerr's board and nominee roles at Google and DoorDash. The record supports the rule more strongly than any causal claim about one coach. [researcher inference · strong] Edge Charlie Rose Google S-1 DoorDash S-1
• Limits, failure modes, and exception ledger: Team quality is vulnerable to charisma, selection bias, and hindsight. A board can become intrusive while calling itself helpful, and we recruited can erase founders or other partners. Observable warnings are unresolved role gaps, reference variance, low hiring throughput, founder dependence on one executive, and investor interventions with no named owner or decision right. The absent rejected-team denominator prevents a base-rate efficacy claim. [researcher inference · strong]
• Evolution and contradictions: The model moves from 1997 recruiting to 1999 character, 2018 five screens, and 2021 adaptation. Its tension is that Doerr's Stanford advice valorizes founder/operator experience while his own path reached venture before a fully documented founder tenure; that career slogan is therefore not promoted to a separate model. [researcher inference · moderate] Stanford, p. 1 Ritholtz, career chronology
• Modern VC translation (as of 2026-08-23): For distributed and AI-native companies, test role clarity across model, product, security, distribution, and regulation, then write a board-support contract naming what the investor will and will not do. Applicability signals are evidence-based conflict, rapid learning, founder pull, and recruiting conversion; likely misuse is replacing a capability gap with a celebrity adviser. Analyst-created hypothetical: an AI infrastructure IC memo assigns each diligence and post-close workstream by person rather than saying the firm will help. [researcher inference · strong]

Model 3 — Ride enabling platform waves, not product labels

• Rule: Search for enabling layers that expand many downstream markets, but require evidence that a company controls a valuable position in the layer. A fashionable product label is not a platform, and a wave is not a timing rule. [researcher inference · strong] Edge, internet-business taxonomy Ferriss, Google decision exchanges
• Evidence status: Doerr repeatedly names microprocessors, open systems, the internet, mobile, synthetic biology, clean infrastructure, and AI as enabling waves. The investment rule is reconstructed across sources; the claimed thirteen-year cadence remains unvalidated and is excluded as a predictive model. [investor-stated · moderate] Charlie Rose, 2004 at 16:15–21:27 Senate AI statement, pp. 1 and 4–5
• Provenance and origin: Intel exposed Doerr to the microprocessor as a general-purpose layer; Sun and Netscape extended the logic to open networked computing and the browser; later accounts apply it to the iPhone and AI. These are Doerr's remembered intellectual links, not proof he originated the platform concept. [retrospective witness · moderate] Ritholtz, Intel and innovation-wave questions Sun proxy, Doerr board chronology
• Mechanism and reasoning: An enabling layer reduces the cost of creation, distribution, or transaction for many complements. Value capture then depends on control points such as product superiority, developer or user adoption, distribution, data, standards, switching costs, and credible monetization—not merely exposure to the wave. [researcher inference · strong] Google announcement, relevance and usage claims Netscape S-1 mirror, business and ownership
• Historical and vehicle context: Netscape was associated with KPCB VII and related side vehicles; Google was a joint KPCB/Sequoia round through an unidentified KPCB fund; the announced iPhone fund lacks a public vehicle schedule in the transcript. Platform continuity does not imply one fund, sponsor, or return series. [documented behavior · strong] Netscape S-1 mirror, p. 48 Google announcement CNN, iPhone-fund exchange
• When to use: Use when a new technical capability may become infrastructure for multiple categories or when a company claims platform status before complement adoption exists. It is useful for distinguishing infrastructure, tool, marketplace, application, and distribution positions. [researcher inference · strong]
• How to apply it: Name the layer, its complements, the scarce control point, adoption evidence, capture mechanism, incumbent response, and migration risk. Kill or reclassify the platform thesis when usage is single-application, complements do not form, distribution remains hostage to another gatekeeper, switching costs are imagined, or monetization conflicts with ecosystem growth. [researcher inference · strong]
• Examples:
  • Documented application — Netscape: The decision thesis treated the browser as the internet client platform and server software as a monetization path before mature browser economics. The rapid IPO is outcome evidence, not a substitute for the 1994 underwriting. [researcher inference · moderate] Institutional Investor, Netscape passages Netscape filing mirror
  • Documented application — Google: Product relevance, compounding query usage, and a large internet thesis supported an enabling search position before mature advertising economics. The official 1999 announcement preserves decision-date claims. [contemporaneous record · moderate] Google announcement Ferriss, Google exchanges
  • Counterexample — Zynga: Doerr later framed the adverse lesson as failing to execute the platform transition from Facebook to mobile. Exposure to a large platform did not protect a dependent application from a control-point shift. [investor-stated · moderate] TechCrunch, Doerr's Zynga comments
• Evidence: Edge supplies a contemporaneous internet taxonomy; Google and Netscape issuer records provide product, market, security, and timing evidence; CNN documents the mobile-platform fund claim; the Senate statement supplies the later AI framing and its limitations. This breadth supports wave recognition, not the thirteen-year forecast. [researcher inference · strong] Edge Google CNN Senate
• Limits, failure modes, and exception ledger: Platform language invites category inflation, timing error, and value-capture blindness. Observable warnings are one dominant distributor, no independent complements, subsidy-only adoption, rising customer-acquisition dependence, or a value pool accruing elsewhere. The official record supplies no KPCB-wide wave entry/exit ledger, so hit rates and timing skill remain [researcher inference · insufficient public record].
• Evolution and contradictions: The platform frame expands from internet transactions and open systems to dedicated mobile capital and an ecosystem-wide AI argument. The 2023 thirteen-year cadence is a retrospective pattern with no selection rule, counterexample, or confidence interval; it is evidence of narrative organization, not a clock. [researcher inference · strong] Edge, 1997 Senate AI statement, p. 1
• Modern VC translation (as of 2026-08-23): In AI, ask whether the company owns a durable control point in compute, models, data, orchestration, security, distribution, or workflow—not whether it uses AI. Applicability signals are independent complements, organic developer or user pull, declining marginal adoption cost, and a defensible capture layer; likely misuse is funding wrappers as platforms. Analyst-created hypothetical: an agent framework fails the model if all demand, distribution, and economics remain controlled by one upstream model vendor. [researcher inference · strong]

Model 4 — Turn the thesis into measurable milestones and transparent review

• Rule: Translate a consequential objective into a small set of specific, time-bound, verifiable results; assign owners, expose progress, and review on a cadence tied to decisions. Metrics are a learning instrument, not a substitute for judgment or culture. [investor-stated · strong] WIRED, OKR method and hazards TED, OKR method at 1:01–2:42 and 10:06–10:57
• Evidence status: Doerr directly teaches the method and repeatedly attributes its origin to Andy Grove. Evidence establishes transmission and selected applications, not causal efficacy across the more than fifty organizations claimed or across abandoned and failed users. [retrospective witness · strong] What Matters, Intel and Google sections Penguin Random House, book description
• Provenance and origin: Grove's Intel practice is the named source; Doerr presents himself as messenger to Google and other organizations. Nuna's purpose belongs to Jini Kim, ONE's risk culture to Bono, and Chrome's user metric to Sundar Pichai. [retrospective witness · moderate] TED, cases at 3:11–9:28 Ferriss, Grove and Google questions
• Mechanism and reasoning: The objective creates direction; key results convert assumptions into observable outcomes; transparency coordinates dependencies; periodic grading exposes misses; and the review should trigger an explicit operating or capital decision. Without the last link, measurement becomes reporting theater. [researcher inference · strong] WIRED, transparency and review sections
• Historical and vehicle context: The method is post-investment operating support, not evidence of KPCB investment selection, fund economics, or formal board rights. Company adoption, management ownership, board use, and any link to follow-on financing must be established case by case. [researcher inference · strong] Google S-1, board and governance Measure What Matters publisher record
• When to use: Use when a thesis has a short or medium feedback loop, controllable drivers, and decisions that can be predeclared. Avoid forcing it onto discovery work, safety questions, or sovereign systems where authority, data definitions, and causal control are fragmented. [researcher inference · strong]
• How to apply it: State one objective, two to five outcome KRs, baseline, deadline, owner, data source, review cadence, and green/yellow/red action rule. Separate committed from stretch results. Kill a KR when it ceases to measure the intended outcome; kill or revise the thesis when repeated misses survive adequate execution and the predeclared evidence test. [researcher inference · strong]
• Examples:
  • Documented application — Google and Chrome: Doerr's remembered Google pitch introduced OKRs; the Chrome case kept users as the metric across three years, missed 20 million and 50 million, then reached 111 million against a 100-million target after product and distribution changes. [retrospective witness · moderate] WIRED, Google excerpt TED, Chrome at 8:09–9:28
  • Documented application — Nuna and ONE: Doerr uses Nuna to connect purpose to a low-profit Medicaid project and ONE to connect an audacious objective to risk acceptance. The purposes and decisions belong to Kim and Bono; the talk does not independently validate results. [retrospective witness · moderate] TED, Nuna/ONE at 3:11–8:08
  • Researcher-applied boundary — public climate scorecard: The 2021 plan offered 10 objectives and 55 KRs, but the 2026 review changed targets and data without a public formula-to-action ledger. Visibility improved; accountability remained incomplete. This is a public-governance application of the model, not a documented portfolio decision. [researcher inference · strong] 2021 plan, p. 2 2026 plan, p. 2
• Evidence: Direct teaching appears in TED, What Matters, WIRED's official excerpt, and the publisher record; the two official climate plans expose an observable implementation and its boundary. The sources are independent routes but largely selected by Doerr or his collaborators, so adoption is stronger than efficacy. [researcher inference · strong] TED What Matters 2021 plan
• Limits, failure modes, and exception ledger: Metrics can narrow attention, invite gaming, turn stretch goals into forecasts, or let boards weaponize misses. Observable warnings are proxy drift, targets without owners, grading without decisions, ever-green status, and management sandbagging. No public rejected/adoption denominator, board action log, or causal comparison was recovered; the model therefore claims disciplined observability, not guaranteed performance. [researcher inference · strong]
• Evolution and contradictions: Doerr's narrative evolves from Intel execution to Google scaling and then to global climate governance. In 2026 he says healthy OKRs should be at least 60% self-declared, an authority boundary absent from many earlier summaries. Company methods cannot be assumed to transfer unchanged to governments. [investor-stated · strong] Climate One, opening Doerr turns TED climate campaign, 17:24–18:15
• Modern VC translation (as of 2026-08-23): Use a visible hypothesis-and-milestone loop, but pair AI growth with gross-margin-per-inference, reliability, security, and customer outcome rather than a universal benchmark. Applicability signals are stable definitions, short feedback, controllable drivers, and a predeclared board action; likely misuse is dashboard governance without rights. Analyst-created hypothetical: two consecutive reliability misses trigger a security review and hiring change rather than an automatic CEO judgment. [researcher inference · strong]

Model 5 — Revise the plan without erasing the audit trail

• Rule: Change a target, method, or thesis when evidence changes, but preserve the original baseline, deadline, result, rationale, approver, and decision consequence. Revision is learning only if the prior miss remains scoreable. [researcher inference · strong] WIRED, review and revision discussion 2021 and 2026 plans
• Evidence status: Doerr directly supports course correction and revisable plans; the versioned audit-trail requirement is a researcher reconstruction prompted by observable target, baseline, and definition changes. It is not represented as a verbatim Doerr rule. [researcher inference · strong] Speed & Scale sample, printed pp. xvii–xxix Signed 2026 letter, pp. 2–3
• Provenance and origin: The model extends the OKR review cycle into an evidence-governance rule. Its strongest source is not a slogan but the contrast between the official 2021 and 2026 Speed & Scale plans plus the 2026 Climate One explanation of changed data and deadlines. [documented behavior · strong] 2021 plan Climate One, 05:30 and 15:30 exchanges
• Mechanism and reasoning: A preserved prior forecast makes learning possible: reviewers can separate thesis error, execution miss, metric defect, and changed environment. An erased or silently redefined target protects the narrator from falsification and destroys the evidence needed for the next decision. [researcher inference · strong]
• Historical and vehicle context: Climate plans are advocacy and coordination artifacts, not KPCB fund reports. Company boards, firm ICs, policy campaigns, and philanthropic programs have different authority; each needs its own approver and consequence field rather than one generic scorecard. [researcher inference · strong] Climate One, speaker and ownership boundaries KPCB chair announcement
• When to use: Use whenever a financing memo, board plan, technical milestone, policy target, or data series is revised. It is essential when a new measurement method makes the current number incomparable to the old one. [researcher inference · strong]
• How to apply it: Maintain a version table with original and current target, old and new definitions, actual result, status, reason, approver, and decision. Tag each item achieved, missed, redefined, merged, dropped, or not comparable. Do not overwrite historical actuals. Kill the revision if it lacks a bridge or merely moves the deadline without new mechanism, resources, or evidence. [researcher inference · strong]
• Examples:
  • Documented strategy reset — post-dot-com KPCB: In 2002 Doerr acknowledged that Kleiner invested over the top, contrasted that with not raising a billion-dollar fund or opening European offices, and argued for smaller funds and less capital per startup. The contemporaneous interview records the changed prescription, while the original fund plan and comparative return series remain absent. [contemporaneous record · strong] Charlie Rose, 4:28–7:54 and 15:33–16:39
  • Researcher-applied missed-deadline case — electric buses: The 2021 all-new-buses-electric-by-2025 KR was missed and moved to 2030. The 2026 interview acknowledges the move, but a complete cause/resource/action bridge is absent; this is a public-plan application, not a documented portfolio decision. [researcher inference · strong] TED 2021, 3:00–3:39 Climate One, 15:30
  • Researcher-applied method-change case — emissions baseline: The team moved from a 59-billion-ton UN-reported baseline to Climate TRACE's 74-billion-ton estimate because observation changed. Treating the 15-billion-ton difference as same-method deterioration would be false; treating it as no accountability issue would also be false. [researcher inference · strong] Climate One, data-source exchange
• Evidence: WIRED and TED establish the review logic; the two official plans and Climate One supply independently inspectable changes in deadlines, targets, and methodology. The record supports the audit need but does not reveal a full change log, owner map, data dictionary, or decision ledger. [researcher inference · strong] WIRED 2026 plan Climate One
• Limits, failure modes, and exception ledger: A rigid audit trail can punish legitimate discovery; a permissive one can normalize goalpost moving. Observable warnings are changed denominators with no bridge, deleted targets, retroactive baselines, unchanged status despite a new formula, and no decision owner. Retrieval exception: no public 2021→2026 itemized changelog, source workbook, status formula, or KR owner/action map was recovered after review of both official plans and the full 2026 transcript. [researcher inference · insufficient public record]
• Evolution and contradictions: The doctrine shifts from quarterly company OKRs to a global climate plan that explicitly admits multiple right answers and uncertain estimates. That broader scope increases the need for revision and simultaneously weakens common authority, comparable data, and enforcement. [investor-stated · strong] Speed & Scale sample, printed pp. xvii–xxix
• Modern VC translation (as of 2026-08-23): Preserve every original underwriting memo and attach a structured delta at each round. Applicability signals are rapid market or model change, new measurement quality, and decision-linked revisions; likely misuse is rewriting the thesis after every financing. Analyst-created hypothetical: an AI company changes from tokens served to verified task completion, retains both series, explains non-comparability, and predeclares what conversion or margin result unlocks the next check. [researcher inference · strong]

Model 6 — Retire fatal risks in sequence and match capital to the technology clock

• Rule: Identify the one or two risks that can kill the company, attack them in an evidence-producing sequence, and finance each step with capital whose duration and rights match the technology and deployment clock. Persistence is earned by risk retirement, not by mission or sunk cost. [investor-stated · strong] Ferriss, cleantech failure exchange Charlie Rose, 2008 at 17:34–26:48
• Evidence status: Doerr directly says physical ventures need more time and capital and that decisive risks should be retired early. Filing-based cases show durations, securities, capital escalation, and outcomes, but no public IC milestone, reserve, or stop ledger proves consistent application. [retrospective witness · provisional] Ritholtz, venture-capital role and renewable-energy exchanges Bloom S-1
• Provenance and origin: A 2006 oil-price threshold makes customer economics explicit; the 2008 crisis panel distinguishes internet, biotech, and green-company financing clocks; the 2021 interviews state the risk-retirement lesson after cleantech losses. [investor-stated · strong] Charlie Rose, 2006 at 13:33–17:05 Ferriss
• Mechanism and reasoning: Technical proof, manufacturability, unit cost, reliability, demand, regulation, project finance, and scale are different uncertainties. Sequencing converts capital into information; matching vehicle and syndicate duration prevents a technically valid program from dying between venture proof and deployment. [researcher inference · strong] TED, physical-company underwriting at 3:03–5:46 2026 plan, KRs 10.3–10.4
• Historical and vehicle context: KPCB main funds, dedicated cleantech programs, co-investors, family-affiliated Foris capital, project finance, government loans, and philanthropy are not fungible. Amyris and QuantumScape filings show why security, control, and conflict boundaries must accompany any duration argument. [documented behavior · strong] Amyris 13D/A, ownership footnotes QuantumScape Form 3
• When to use: Use for physical science, biotech, energy, infrastructure, regulated systems, or any company whose validation and deployment clocks exceed ordinary software iteration. It is also useful when a follow-on is being justified primarily by elapsed effort or mission importance. [researcher inference · strong]
• How to apply it: Build a risk ladder with hypothesis, cheapest decisive test, pass/fail threshold, cash and time to result, next capital source, and stop/partner/pivot action. Test the highest expected-loss risk early, but preserve dependencies: a cell result does not prove yield, a factory does not prove unit economics, and a pilot does not prove bankable deployment. Kill or syndicate when the next risk cannot be retired within the vehicle's reserves and remaining life. [researcher inference · strong]
• Examples:
  • Documented application/boundary — QuantumScape: The 2010 thesis concerned a manufacturable solid-state separator; the next-check evidence should include repeated multilayer cells, yield, cycle, safety, OEM validation, and factory cost. The Form 3's 18,349,929 KPCB XIV nominee shares document a later position, not initial risk retirement or personal ownership. [documented behavior · strong] QuantumScape S-1 Form 3
  • Mixed case — Bloom Energy: Bloom reached its IPO after sixteen years with real revenue and material-cost progress, roughly $1.5 billion gross private capital raised, and a roughly $2.3 billion accumulated deficit. This can reflect patient risk retirement or delayed loss recognition; absent milestone and cash ledgers prevent choosing. [contemporaneous record · strong] Bloom 2018 S-1
  • Counterexample — Fisker and Amyris: Fisker failed after large public and private capital while production and supplier risks remained; Amyris's valid synthetic-biology platform met scale-up, burn, governance, and related-party financing escalation before old equity was canceled. Neither record exposes a sponsor stop ledger. [researcher inference · strong] Fisker chronology Amyris 2024 8-K
• Evidence: Three direct-voice routes state duration, capital, and early risk retirement; SEC and court records supply four adverse or mixed cases. The spoken cleantech denominator conflicts at roughly 70 versus 100 ventures and lacks vehicle cash flows, so no hit rate or return validates the rule. [researcher inference · insufficient public record] Ferriss Ritholtz Bloom S-1
• Limits, failure modes, and exception ledger: Long duration can rationalize escalation of commitment; retire the largest risk first can ignore dependencies or cheap learning elsewhere. Observable warnings are repeated follow-ons with unchanged milestones, single-cell results presented as manufacturing proof, no reserve model, policy demand without offtake, and a next-round assumption without a named provider. Retrieval exception: no public KPCB cleantech IC memos, reserve schedule, stop decisions, write-off ledger, or reconciled company/fund cash-flow series was recovered. [researcher inference · insufficient public record]
• Evolution and contradictions: The category warning is contemporaneous by 2008; the sharper early-risk lesson is retrospective after losses. Doerr reports remaining cleantech value above invested capital while also saying most ventures failed, but conflicting counts and unaudited valuations cannot show whether persistence added value. [retrospective witness · provisional] Charlie Rose, 2008 Ferriss, cleantech exchange
• Modern VC translation (as of 2026-08-23): For frontier AI infrastructure, climate, and biotech, pair the technical roadmap with a financing-clock map and precommitted kill criteria. Applicability signals are discrete validation events, credible later capital, repeatable production, and bankable customers; likely misuse is treating abundant capital as technical evidence. Analyst-created hypothetical: a new battery check unlocks only after blinded multilayer yield and cycle distributions, not a best-cell demonstration. [researcher inference · strong]

Model 7 — Use networked diligence to turn uncertainty into intervention levers

• Rule: Before prescribing or investing in a complex system, assemble a deliberately diverse network of technical, operating, policy, financing, distribution, and affected-stakeholder knowledge; convert disagreements into a map of testable intervention levers. [researcher inference · strong] TED 2007, learning effort at 1:29–2:27 White House transcript, pp. 5–6
• Evidence status: Doerr directly describes a climate-learning network of about fifty scientists, advocates, entrepreneurs, and business leaders and later participates in committee and book-research networks. The conversion from network to decision ledger is reconstructed; marginal information value and dissent coverage are not public. [investor-stated · strong] TED 2007 Ritholtz, Speed & Scale research/writing question
• Provenance and origin: The model is consistent with KPCB's older network-and-recruiting practice but becomes explicit in Doerr's 2007 climate account. Named collaborators—not Doerr alone—supply important content: Bob Epstein on AB32, Hal Harvey on emissions and policy targeting, Panchadsaram on the climate scorecard, and committee participants on implementation constraints. [contemporaneous record · strong] TED 2007, Epstein and Berkeley/Gates at 6:13–11:53 TED 2020, speaker-separated transcript
• Mechanism and reasoning: Different experts reveal different failure surfaces; structured disagreement reduces single-discipline error and identifies where technology, distribution, finance, or policy can move the system. The output should be a causal map with evidence, owner, and falsifier—not a consensus narrative. [researcher inference · strong]
• Historical and vehicle context: Expert access can serve firm diligence, board work, public advocacy, or philanthropic program design. Those activities and their financing must remain separate; participation by a KPCB partner does not establish a fund investment, and book research does not establish investment approval. [researcher inference · strong] Senate AI statement, pp. 4–5 Speed & Scale sample, collaborators
• When to use: Use when the system spans disciplines, regulation, infrastructure, or multiple capital types and no single expert can see the full path. It is especially valuable before selecting an intervention rather than after the investment to validate a preferred answer. [researcher inference · strong]
• How to apply it: Build an expertise and stakeholder matrix; recruit independent and dissenting views; record each source's incentive, claim, evidence, and confidence; map bottlenecks and controllable levers; run pre-mortems; then name the decision that each new fact could change. Stop the network phase when marginal interviews cease changing the causal map, not when the room agrees. Kill a lever when its prerequisite or measurable effect fails. [researcher inference · strong]
• Examples:
  • Documented application — 2007 climate learning: Doerr says the partnership learned from about fifty scientists, policy advocates, entrepreneurs, and business leaders before organizing responses into companies, individuals, policy, and radical innovation. The source shows breadth, not interview identities or rejected conclusions. [investor-stated · strong] TED, 1:29–2:27
  • Documented application — 2009 recovery committee: Doerr says the committee conducted outreach inside and outside Washington and proposed linking carbon pricing, private capital, incumbent retailers, and idle construction labor. Other participants immediately added finance, training, standards, and demand constraints, demonstrating useful plural challenge. [contemporaneous record · moderate] White House transcript, pp. 5–6 and participant responses
  • Boundary case — Speed & Scale research: Doerr reports input from 100 leaders, 35 selected stories, about 1,000 checked data points, roughly 500 endnotes, and a small research/writing team. Selection criteria, negative cases, and decision impact remain unavailable, so network size is not diligence quality. [retrospective witness · moderate] Ritholtz, book research/writing question
  • Researcher-applied portfolio failure boundary — Amyris: The public origin story links Berkeley researchers and Gates-funded antimalarial work to Amyris's synthetic-biology platform; the issuer later documented scale-up progress, KPCB exposure, and then cancellation of old equity after capital and governance escalation. This shows why expert-network insight cannot substitute for manufacturing, capital-structure, and conflict verification; it is not a claim that Doerr contemporaneously applied this model to the deal. [researcher inference · strong] TED, Amyris provenance at 9:36–11:17 Amyris 2010 prospectus Amyris 2024 8-K
• Evidence: TED, an official White House transcript, the Speed & Scale sample, and Ritholtz independently document named collaborators and networked inquiry in three contexts. None supplies interview notes, saturation metrics, dissent logs, or a causal comparison to decisions made without the network. [researcher inference · moderate] TED White House Ritholtz
• Limits, failure modes, and exception ledger: Elite networks can reproduce shared assumptions, omit affected users, and launder advocacy as expertise. Observable warnings are repeated affiliations, unnamed dissent, interview counts without changed beliefs, and recommendations aligned with sponsor holdings. Retrieval exception: no complete 2007 participant list, interview protocol, rejected-lever log, or marginal-information record was recovered; the model therefore requires deliberate dissent and stakeholder coverage as an analyst extension. [researcher inference · insufficient public record]
• Evolution and contradictions: The practice broadens from company recruiting networks to climate, economic recovery, and AI ecosystem inquiry. Yet later public documents can be more prescriptive than transparent about the underlying disagreement set, and the 2023 AI statement develops opportunity recommendations more fully than governance constraints. [researcher inference · strong] Senate AI statement, pp. 1–5
• Modern VC translation (as of 2026-08-23): For AI, biotech, defense, or climate, maintain an evidence graph linking claims to technical experts, customers, regulators, financiers, and affected users. Applicability signals are cross-domain dependencies and high model uncertainty; likely misuse is using prestigious advisers as social proof. Analyst-created hypothetical: an AI-health deal cannot close until clinician, patient, payer, safety, privacy, and deployment evidence changes a named assumption or records a dissent. [researcher inference · strong]

Model 8 — Underwrite the ecosystem and disclose the investor's policy incentive

• Rule: Treat talent, research, compute or infrastructure, public funding, commercialization capital, regulation, distribution, and exit channels as part of the investment system; when advocating policy, disclose which vehicles or portfolio interests may benefit and test public welfare separately. [researcher inference · strong] Senate AI statement, pp. 4–5 White House transcript, pp. 5–6
• Evidence status: The ecosystem components are directly stated across Doerr's innovation and policy record. The incentive-disclosure rule is a researcher safeguard because the public statements identify him as an investor but do not provide a holdings schedule, recusal policy, or expected-benefit analysis. [researcher inference · strong] Senate AI statement, introduction and M&A section GovInfo hearing record, printed pp. 21–24
• Provenance and origin: The model appears by 2004 in Doerr's emphasis on universities, immigration, R&D, connectivity, and enabling policy; climate adds carbon pricing and deployment finance; the 2023 AI statement adds public compute/data, regional clusters, IPOs, and acquisitions. [contemporaneous record · strong] Charlie Rose, 2004 at 3:53–6:56 and 16:15–25:23 Senate AI statement, pp. 4–5
• Mechanism and reasoning: A startup can fail despite valid technology when a complement—trained labor, grid access, standards, buyer finance, project capital, or liquidity—is absent. Public rules can create demand and lower coordination costs, but the same rules can transfer value, entrench incumbents, or improve investor liquidity without improving innovation. [researcher inference · strong] White House transcript, retail/distribution proposal Senate AI statement, IPO/M&A discussion
• Historical and vehicle context: KPCB funds, DGF, family-affiliated holdings, the Doerr family foundation, Breakthrough Energy Ventures, philanthropy, government programs, and issuer capital are distinct. A favorable policy may benefit several without proving Doerr's personal check, control, or return. [documented behavior · strong] Amyris 13D/A, KPCB and Foris ownership QuantumScape prospectus, fund and family footnotes
• When to use: Use for regulated, infrastructure-dependent, research-intensive, or nationally strategic markets and whenever a thesis assumes a subsidy, procurement rule, acquisition path, or public input. It is also mandatory when an investor becomes a policy advocate in a sector where related vehicles may hold positions. [researcher inference · strong]
• How to apply it: Map every necessary complement, owner, lead time, funding source, and failure consequence. Run the company case with and without the proposed policy; identify beneficiaries, payers, concentration effects, sunset conditions, and recusal/disclosure needs. Kill or reprice when the thesis requires one unavailable complement, when subsidy unit economics never converge, or when the only credible exit depends on the advocated rule. [researcher inference · strong]
• Examples:
  • Documented application — 2009 clean-energy policy: Doerr argued that a carbon cap and price could create durable demand and pull private capital while R&D and deployment funding accelerated supply. The hearing also identifies him as an investor in the sector; no portfolio-interest schedule appears. [investor-stated · strong] GovInfo hearing, printed pp. 21–24
  • Documented application — Cash for Caulkers: The recovery proposal linked a durable carbon signal to incumbent retail distribution and underused construction labor. Other participants surfaced finance, training, standards, and access gaps, showing that the ecosystem map was incomplete. [contemporaneous record · moderate] White House transcript, pp. 5–6
  • Counterexample/incentive boundary — AI exits: The 2023 statement joins public compute, talent, R&D, clusters, IPOs, and M&A, but pro-acquisition policy also aligns with venture liquidity incentives and omits a portfolio schedule or concentration analysis. [researcher inference · strong] Senate AI statement, pp. 4–5
  • Researcher-applied portfolio boundary — Bloom Energy: Bloom's filing records sixteen years to IPO, material-cost progress, roughly $1.5 billion of gross private capital raised, and an approximately $2.3 billion accumulated deficit; Doerr's 2008 category analysis separately warned that capital-intensive green companies depended on later credit. The combination illustrates ecosystem and financing dependence, but does not prove Doerr applied this model at entry or that public policy produced a KPCB return. [researcher inference · strong] Bloom 2018 S-1 Charlie Rose, 2008 at 17:34–26:48
• Evidence: Charlie Rose supplies the early innovation-system view; official Senate and White House records document policy mechanisms and investor identity; issuer filings show why related vehicles cannot be inferred or merged. The sources support ecosystem underwriting and the need for disclosure, not policy effectiveness or net welfare. [researcher inference · strong] Charlie Rose GovInfo Amyris filing
• Limits, failure modes, and exception ledger: Ecosystem thinking can become an unfalsifiable excuse for company weakness or a lobbying narrative. Observable warnings are dependence on one subsidy, no unsubsidized cost path, missing affected stakeholders, exit policy framed only through startup liquidity, and undisclosed related holdings. Retrieval exception: the official policy routes provide no contemporaneous KPCB/DGF/family portfolio schedule, expected-gain analysis, recusal protocol, or bill-by-bill conflict ledger. [researcher inference · insufficient public record]
• Evolution and contradictions: The frame expands from national innovation inputs to climate market formation and AI competitiveness. Doerr acknowledges AI risks, but the detailed 2023 recommendations emphasize capacity and exits more than enforceable safeguards; opportunity and investor incentives are developed more fully than downside governance. [researcher inference · strong] Senate AI statement, pp. 1–5
• Modern VC translation (as of 2026-08-23): Add an ecosystem dependency graph and conflict statement to every regulated or infrastructure memo. Applicability signals are public inputs, constrained infrastructure, multiple capital layers, and policy-shaped demand; likely misuse is treating investor-friendly policy as neutral public interest. Analyst-created hypothetical: an AI-compute thesis reports which portfolio companies benefit from public compute and acquisition reform, then reruns returns and welfare without either policy. [researcher inference · strong]

Model 9 — Keep people, vehicles, and outcomes in separate ledgers

• Rule: Record who performed each decision role, which legal vehicle supplied or held capital, and what outcome metric is actually observed. Never convert a board seat, firm investment, nominee ownership row, company valuation, or collective we into personal credit or a realized return. [researcher inference · strong] Google announcement DoorDash proxy, ownership footnotes
• Evidence status: This is a researcher-constructed attribution model compelled by recurring filing structures and narrative ambiguity. Doerr's speeches credit mentors, founders, partners, and collaborators, but no public firm-wide decision-right or cash-flow ledger makes the separation complete. [researcher inference · strong] NVCA acceptance speech KPCB chair announcement
• Provenance and origin: The rule arises from three recurring conflicts: Doerr versus KPCB versus co-investors; main funds versus nominee, DGF, family, trust, and philanthropic vehicles; and company success or market value versus fund distributions. It is an audit discipline, not a claim that Doerr publicly taught this exact framework. [researcher inference · strong] Google S-1, principal-holder disclaimers QuantumScape prospectus, footnotes 7–13
• Mechanism and reasoning: Separate ledgers prevent three causal errors: collapsing many contributors into one narrator, treating nominal or shared voting power as personal economics, and deriving returns from a later valuation. The discipline makes uncertainty visible and permits later records to update one layer without rewriting the others. [researcher inference · strong]
• Historical and vehicle context: Public records expose KPCB VII and side vehicles at Netscape; KPCB Holdings as nominee at Google and DoorDash; DGF entities at DoorDash; KPCB XIV and family/trust positions at QuantumScape; and KPCB XII versus Foris-related exposure at Amyris. None is interchangeable with John Doerr personally. [documented behavior · strong] Netscape S-1 mirror, p. 48 Amyris 13D/A
• When to use: Use in every investment case, portfolio statistic, attribution claim, policy conflict analysis, board history, and return statement. It is mandatory when a source says we, lists beneficial ownership, combines affiliates, or celebrates a company outcome without cash flows. [researcher inference · strong]
• How to apply it: Maintain three linked tables. The people table names source, sponsor, diligence, approval, negotiation, board, recruiting, follow-on, exit, and narrator. The vehicle table names issuer, legal holder, beneficial/pecuniary status, security, check, date, rights, and follow-ons. The outcome table separates company event, security proceeds, gross fund cash, net LP cash, and unrealized marks. Stop at unknown rather than impute. [researcher inference · strong]
• Examples:
  • Documented application — Google: The issuer announced a $25 million aggregate KPCB/Sequoia round and Doerr/Moritz board roles; the S-1 later listed 23,893,800 shares through KPCB Holdings with a pecuniary-interest disclaimer. Neither fact supplies KPCB's check split, Doerr's personal stake, distributions, or return. [documented behavior · strong] Google announcement Google S-1
  • Documented application — DoorDash: The S-1 documents Doerr's board role and DGF securities; the proxy combines DGF nominee, DGF II Associates, VVT, and controlled entities in one beneficial-ownership row. Founders executed, and a disclosed follow-on does not create a full DGF or personal return. [documented behavior · strong] DoorDash S-1 DoorDash proxy
  • Counterexample — Amyris: KPCB XII residual equity, Foris family-affiliated debt/equity/warrants, DIP claims, and successor equity are different exposures. Old equity was canceled, but a contradictory filing sentence and absent cash ledger prevent combining them into one Doerr outcome. [researcher inference · strong] Amyris 13D/A Amyris 2024 8-K
• Evidence: Primary issuer announcements and SEC filings across Google, DoorDash, Netscape, QuantumScape, and Amyris independently repeat the legal and attribution problem. Direct interviews repeatedly use collective we without IC, term-sheet, sponsor, or cash-flow detail. [researcher inference · strong] Netscape S-1 mirror QuantumScape Form 3 Ferriss
• Limits, failure modes, and exception ledger: Legal ownership tables can still omit internal economics, carry, side letters, opportunity allocation, and distributions; role maps can overstate named actors and omit invisible work. Observable warnings are aggregate affiliate rows, pecuniary disclaimers, we without names, valuation-derived returns, and the same person treated as source, approver, board, and executor. Retrieval exception: no KPCB-wide opportunity-allocation, IC-vote, carry, reserve, distribution, or succession-right ledger was recovered. [researcher inference · insufficient public record]
• Evolution and contradictions: Early public narratives often personalize partnership activity; later filings expose increasingly complex nominee, growth-fund, family, and trust structures. Doerr's own later collaborator credit is a corrective, but selected success stories and aggregate company/job/market-cap claims still compress founders, co-investors, and vehicles. [researcher inference · strong] NVCA speech Senate AI statement, p. 1
• Modern VC translation (as of 2026-08-23): Put an attribution ledger beside every IC case, case study, and performance claim, especially in multi-stage firms and SPV-heavy portfolios. Applicability signals are shared sponsorship, nominee structures, cross-fund follow-ons, and policy advocacy; likely misuse is ranking investors by famous company market caps. Analyst-created hypothetical: report Fund X invested; Partner A sourced; Partner B diligenced; Partner C joined the board; founders executed; proceeds unknown until primary evidence permits more. [researcher inference · strong]

Evidence gaps and candidate models

The supported inventory therefore contains nine models. Every model has exactly 12 required fields, at least three grounded cases including a counterexample or boundary case, and at least four claim-supporting citation placements spanning two or more independent underlying sources. Candidate gaps are excluded rather than promoted from slogans, collaborator testimony, single-source taxonomy, or missing decision records. [researcher inference · strong]

What the evidence supports

The strongest defensible thesis is narrower than “great picker.” In the Internet-era cases with the best public record, Doerr repeatedly helped convert a technical or market discontinuity into an institutional company-building program: source through operator and founder networks, test product or technical superiority, recruit missing executives, secure partnership capital, take governance responsibility, and make operating priorities inspectable. Netscape, Amazon, and Google support different links in that chain; none supports a lone-genius account. [researcher inference · moderate] edge thesis · Netscape case · Amazon case · Google case

The best evidence is for actions and institutional roles, not personal alpha. Amazon's filing fixes a preferred issuance, two KPCB vehicles, a board date, and proportional-interest disclaimer; Google's announcement fixes a joint KPCB/Sequoia round and two board appointments; Netscape's filing mirror fixes a later nominee block but leaves entry terms disputed. Those records show capital, governance, and ownership boundaries. They do not expose complete cash flows, fees, carry, reserves, distributions, or Doerr-personal returns. [contemporaneous record · strong] Amazon S-1/A, transactions and principal holders · Google announcement, June 7, 1999 · Netscape S-1 mirror, printed p. 48

Google is the likeliest dominant public economic outlier, but its contribution cannot be measured as a return. Excluding it leaves Amazon, Netscape, and Sun as historically material demonstrations of transaction, browser-platform, and network-computing theses; it also eliminates any defensible economic ordering because the remaining cash-flow series is absent. The favorable claim that survives the exclusion test is about a repeatable company-building mechanism, not a quantified economic edge. [researcher inference · strong] three-way Pareto analysis and Google exclusion

The record also supports genuine updating. In 2002 Doerr said KPCB had invested “over the top,” rejected a proposed billion-dollar fund and European expansion, and argued for smaller funds and lower company burn. In 2026 he said the climate program remained off track, while Ryan Panchadsaram explained a moved bus deadline and a changed emissions baseline. These are unusually useful because they preserve rejected alternatives or revised targets; neither supplies a complete fund stop policy or a versioned climate data ledger. [investor-stated · strong] Charlie Rose, 2002 at 16:06–21:18 · Climate One, 2026 assessment and revision exchange

The failure record materially narrows the doctrine. Fisker tests industry-resume heuristics, physical integration, financing duration, and conflict-constrained choice; Amyris tests scale-up, burn, related-party finance, and governance escalation; Bloom shows that technical and commercial progress can coexist with long duration and large accumulated losses; Excite/@Home, Zynga, Weitek, and MiaSolé expose distribution, governance, pass, and capital-intensity errors. These cases support staged risk retirement, explicit financing clocks, and re-underwriting. They do not form a matched causal cohort. [researcher inference · strong] failure and miss postmortems

The public opportunity denominator is deliberately bounded: 49 searched routes equal 46 eligible company candidates plus three explicit passes; 14 candidates are sufficiently reconstructed, four partial, and 28 unresolved. Thus 14/46, or 30.4%, measures reconstruction coverage—not hit rate, opportunity coverage, or capital-weighted performance—and zero cases have audited economic reconstructions. [researcher inference · strong] count rules and mechanical reconciliation

What could falsify the thesis

The proposed edge predicts records that are observable in principle. A real sourcing edge should leave dated referrer and sponsor trails across winners, ordinary cases, and passes. A real picking edge should produce decision-date technical or customer evidence that discriminates selected companies from rejected alternatives. A real intervention edge should produce board-level recruiting, milestone, and follow-on actions that founders or independent participants can identify. A real portfolio edge should survive removal of Google in compatible net cash-flow data. [researcher inference · moderate] profile predictions and falsifiers

The strongest skeptical explanation remains elite KPCB access, founder execution, partnership labor, favorable platform timing, and one extraordinary Google outlier, later organized into a coherent teaching narrative. The favorable and skeptical accounts diverge on evidence that is currently private: complete funnels, contemporaneous IC records, board-labor logs, matched rejected deals, reserve decisions, vehicle cash flows, and ex-Google performance. [researcher inference · strong] strongest case against greatness

Attribution and denominator audit

Attribution must be separated across source, sponsor, diligence, approval, negotiation, board work, recruiting, follow-on capital, exit influence, later narration, and legal vehicle. The package does this most effectively in Google, Amazon, Netscape, DoorDash, Bloom, QuantumScape, and Amyris. It also records contrary evidence: Tom Perkins—not Doerr—held the visible Compaq governance role, and the Weitek pass was attributed by a witness to Perkins. [researcher inference · strong] people, capital, and influence map · Weitek oral history, p. 8

The cleantech record cannot be collapsed into one portfolio statistic. Direct and sponsor accounts use nine, 54, roughly 70, or roughly 100 companies/ventures and multiple deployment/value figures with different scopes; none supplies a fund-by-fund cash-flow schedule. The correct conclusion is not that cleantech “returned 3×,” but that public claims are unaudited, scope-conflicted, and not convertible into KPCB or Doerr returns. [researcher inference · strong] cleantech denominator reconciliation

Transfer limits for modern VCs

The transferable core is an audit discipline: define a consequential search space, identify the cheapest evidence that retires a fatal risk, underwrite whether the team can update, name the vehicle and beneficiary, translate the thesis into measurable milestones, and preserve the original plan when it changes. This is a researcher synthesis of nine models, not a claim that Doerr stated the sequence verbatim. [researcher inference · strong] mental-model field guide · modern translation dated 2026-08-23

Several features do not transfer automatically:

• KPCB's 1990s brand, two-sponsor partnership funnel, recruiting network, and public-market environment were institutional advantages. A new fund cannot imitate them by adopting slogans or OKRs. [researcher inference · strong] historical environment
• Objective-and-key-result discipline works only when the metric is decision-relevant, versioned, and paired with judgment. The 2021-to-2026 climate changes show why baselines, owners, misses, and reasons must remain visible. [researcher inference · strong] revision model
• Capital-intensive climate and deep-technology companies require a financing clock aligned with technical, manufacturing, regulatory, and adoption milestones. Bloom, QuantumScape, Fisker, and Amyris show that “stay the course” without fresh evidence is not a risk rule. [researcher inference · strong] technology-clock model
• Policy can create markets and also create conflicts. Investors advocating subsidies, procurement, or regulation should disclose portfolio and vehicle exposure instead of presenting category advocacy as neutral evidence. [researcher inference · strong] ecosystem-and-incentive model
• “Servant VC” is an aspiration, not a power analysis. Board authority, financing leverage, hiring influence, dissent, founder agency, and attribution still require separate records. [researcher inference · strong] team attribution

The modern decision test is therefore behavioral. Before investing, record the source, sponsor, alternatives, decision-date facts, security, ownership target, reserve plan, kill criteria, milestone owner, and next check. Before following on, re-underwrite what changed. After outcome, keep company outcome, security outcome, fund cash flow, LP return, carry, and personal wealth in separate ledgers. [researcher inference · strong] reusable checklists

Unresolved evidence

The missing record is decision-relevant, not decorative:

1. Executed Google and Netscape financing documents, closing cap tables, allocation schedules, follow-ons, distributions, fees, and carry remain unavailable. [researcher inference · insufficient public record] uncertainty ledger
2. No complete KPCB or Doerr accepted/rejected opportunity ledger, matched ordinary cohort, or partner-attributed loss ledger is public. [researcher inference · insufficient public record] search and saturation receipt
3. The cleantech figures cannot be reconciled by company, vehicle, date, cost, realization state, or valuation method. [researcher inference · insufficient public record] cleantech reconciliation
4. Complete Measure What Matters chapters and complete Speed & Scale endnotes/cases were not available through legitimate reviewed routes. [researcher inference · insufficient public record] writings unresolved record
5. The public climate scorecards lack a data dictionary, source workbook, owner map, status formula, and itemized 2021-to-2026 changelog. [researcher inference · insufficient public record] 2021 plan, p. 2 · 2026 plan, p. 2
6. The canonical Edge transcript timed out on final live reopen, the patent detail route failed, the QuantumScape Form 3 renderer failed, and the Netscape filing remains a noncanonical mirror. Existing captures support bounded claims, but none justifies expanding them. [researcher inference · strong] source access barriers
7. Fees, carry, LP rights, fund votes, opportunity allocation, board labor, reserve decisions, and founder evaluations have insufficient public record across the career. [researcher inference · insufficient public record] denominators and attribution boundary

These gaps are the boundary of the conclusion, not a license to fill it with valuations or retrospective confidence. The package can teach a documented operating method and its failure modes; it cannot establish career-wide personal alpha. [researcher inference · strong] ranked retrieval priorities

Central-claim audit trail

The central audit reopened 90 high-risk claim groups across biography, deals, failures, edge, peers, modern premises, models, and package mechanics. It independently reopened or live-tested 34 primary, direct-voice, issuer, court, government, university, or institutional routes. The final disposition below incorporates every correction before acceptance. [documented behavior · strong]

Ordinary-claim audit trail

The ordinary sample was fixed before inspection. A claim unit was a complete paragraph, bullet, or table row with an externally checkable assertion and external route. Central claims were excluded. The auditor took the first and last eligible claim from each of the seven then-existing core documents, producing 14 claims; fixed seed 230823 selected writing item 11 because (230823 mod 11 = 10) and talk item 13 because (230823 mod 21 = 12). The initial result was 12 PASS and 4 CORRECT; after the four narrow repairs, all 16 pass. [documented behavior · strong]

Review-native claims checked after drafting

To prevent the review itself from escaping review, the first material claim under each of its five required substantive sections was reopened after drafting. [documented behavior · strong]

Chronology and deal cross-check record

The final chronology was checked against independent institutional, issuer, filing, archive, and direct-voice records. Sequence is reported as sequence; causal claims remain labeled inference. [documented behavior · strong] sourced chronology

Twenty-point acceptance evidence

All twenty evidence gates pass on the corrected package. This is acceptance of the documented learning architecture, not a claim that missing private economics, IC records, or archival material have become knowable. [researcher inference · strong] unresolved evidence

Books and signed writings

Retrieved and opened 2026-08-23. Priority: ★★★ core retrieval, ★★ useful corroboration/context, ★ optional lead. Reliability: A primary/official/direct voice, B credible reporting or retrospective witness, C discovery/index metadata. “Full” means the opened route exposed the complete text, transcript, filing, PDF, or playable recording; “partial” and “metadata-only” are literal access boundaries.

• ★★★ | When John Doerr Brought a “Gift” to Google's Founders | https://www.wired.com/story/when-john-doerr-brought-a-gift-to-googles-founders/ | 2018-04-24 | authorized book excerpt | Direct retrospective on Google's OKR adoption and an $11.8 million/~12% KPCB investment claim; reconcile the retrospective amount with the contemporaneous aggregate financing announcement before using it. (reliability: A for Doerr's account, B for historical claim; access: full; locator: opening and Google investment/OKR passages)

• ★★★ | John Doerr on addressing climate change | https://www.fastcompany.com/90694083/john-doerr-on-addressing-climate-change-speed-and-scale-excerpt | 2021-11-08 | authorized book excerpt | Direct excerpt from Speed & Scale on the cleantech record and climate plan; figures remain author assertions unless separately audited. (reliability: A for authorship, B for unaudited historical claims; access: full; locator: excerpt paragraphs)

• ★★★ | Speed & Scale Action Plan | https://speedandscale.com/wp-content/uploads/2021/11/Speed-Scale-Action-Plan-Poster-with-OKRs.pdf | 2021-11 | official action-plan poster | Dated baseline of the climate objectives and KRs later updated in 2026. (reliability: A for program design; access: full; locator: PDF p. 2, objectives 1–10; p. 1 is the cover)

• ★★★ | Let’s Build, Friends, Build | https://speedandscale.com/wp-content/uploads/2026/04/Lets-Build-Friends.pdf | 2026-04 | signed public letter | Current signed statement on AI, energy, permitting, and climate priorities. (reliability: A for Doerr's stated position; access: full; locator: PDF pp. 1–3, signature on p. 3)

• ★★★ | Prepared statement for the Senate AI Insight Forum | https://www.schumer.senate.gov/imo/media/doc/John%20Doerr.pdf | 2023-10-24 | official submitted statement | Direct policy recommendations on AI, talent, research, and public-interest infrastructure. (reliability: A; access: full; locator: PDF pp. 1–5)

• ★★★ | Measure What Matters | https://www.penguinrandomhouse.com/books/546304/measure-what-matters-by-john-doerr-foreword-by-larry-page/ | 2018-04-24 | authored book/publisher record | Core direct doctrine and retrospective OKR history; publisher page gives authoritative publication metadata and a sample route, not the full book. (reliability: A for authorship/metadata; access: partial; locator: Product Details, description, sample link; Google Books preview was also checked)

• ★★★ | Speed & Scale | https://www.penguinrandomhouse.com/books/688191/speed-and-scale-by-john-doerr/ | 2021-11-09 | authored book/publisher record plus official publisher sample | Core climate framework and later-period direct voice; the official sample exposes the introduction, collaborator credits, cleantech retrospective, and opening action-plan material, but not the full book. (reliability: A for authorship/metadata and sample text; access: partial; locator: publisher Product Details and description; sample PDF pp. 7–8 and printed pp. xi–xxxi, 1–9; Google Books limited preview exposes the table of contents but not continuous full text)

• ★★★ | Speed & Scale: 2026 Action Plan | https://speedandscale.com/wp-content/uploads/2026/04/Speed-Scale-2026-Action-Plan.pdf | 2026 | official two-page action-plan update | Current, testable update to the climate program with explicit objectives and key results. (reliability: A for program claims; access: full; locator: PDF pp. 1–2 and numbered KRs)

• ★★★ | DoorDash: delivering the future of local | https://www.kleinerperkins.com/perspectives/doordash-delivering-the-future-of-local/ | 2020-12-09 | signed firm essay | Doerr's own sourcing, diligence, board, and company-building recollection; cross-check with filings and participants. (reliability: A for Doerr's account, B for historical claims; access: full; locator: paragraphs)

• ★★★ | John Doerr NVCA lifetime achievement award acceptance speech | https://www.kleinerperkins.com/perspectives/john-doerr-nvca-lifetime-achievement-award-acceptance-speech/ | 2019-03-11 | signed acceptance speech | Direct firm-history, team, fund-count, capital-raised, apprenticeship, and values claims. (reliability: A for speech, B for unaudited firm-history claims; access: full; locator: paragraphs beginning “Since 1980,” partner acknowledgments, final principles)

• ★★★ | A New Role as KPCB Chair | https://www.kleinerperkins.com/perspectives/a-new-role-as-kpcb-chair/ | 2016-03-31 | signed role-transition statement | Defines Doerr's announced chair, investment-committee, family-foundation, and succession boundaries. (reliability: A for self-described role; access: full; locator: opening and leadership/team paragraphs)

Talks, interviews, testimony, video, and oral direct voice

• ★★★ | How To Be a Venture Capitalist | https://stvp.stanford.edu/videos/how-to-be-a-venture-capitalist/ | 2005-02-02 | Stanford video/transcript fragment | Early direct articulation of operator-first venture preparation; the complete mapped artifact is a 0:19 clip, not a full-length talk. (reliability: A for institutional metadata/direct speech; access: full as published; locator: 0:00–0:19 video and five-paragraph official printable transcript: https://ecorner.stanford.edu/wp-content/uploads/sites/2/2005/02/1281.pdf)
• ★★★ | Salvation (and profit) in greentech | https://www.ted.com/talks/john_doerr_salvation_and_profit_in_greentech | recorded 2007-03-03; published 2007-05-27 | TED video/official transcript | Contemporary cleantech rationale, networked diligence, policy, Amyris platform transfer, a $200 million firm-program claim, and dated forecasts. (reliability: A for direct speech, B for unverified figures; access: full; locator: 17:32 video and official cue-level transcript)
• ★★★ | The Coach | https://www.edge.org/conversation/john_doerr-the-coach | 1997-01-06 | Edge interview/transcript | Full early account of Intel, KPCB recruitment, team construction, venture categories, Amazon/Netscape reasoning, coaching, and partnership. (reliability: A for direct speech, B for unaudited figures/forecasts; access: full through canonical and legacy routes; locator: canonical Doerr turns and legacy Digerati pp. 1–8 at https://www.edge.org/digerati/doerr/)
• ★★★ | President's Economic Recovery Advisory Board meeting transcript | https://obamawhitehouse.archives.gov/files/documents/PERAB_Meeting_11%202%2009_Transcript.pdf | 2009-11-02 | official White House meeting transcript | Primary record of green-jobs, retrofit-distribution, carbon-price, and policy-team recommendations. (reliability: A for meeting/direct speech, B for forecasts; access: full through official index despite fragile direct file route; locator: printed/PDF pp. 5–6 and p. 9; index: https://obamawhitehouse.archives.gov/administration/eop/perab/meetings/11-02-09)
• ★★★ | Ensuring and Enhancing U.S. Competitiveness While Moving Toward a Clean Energy Economy | https://www.govinfo.gov/content/pkg/CHRG-111shrg95163/pdf/CHRG-111shrg95163.pdf | hearing 2009-07-16; official volume printed 2016 | Senate hearing transcript/prepared statement | Primary policy testimony with adversarial questioning, nuclear/subsidy boundaries, and capital forecasts. (reliability: A for official record/direct speech, B for unverified figures; access: full; locator: oral statement printed pp. 21–24, prepared statement pp. 24–44, counterarguments pp. 180–184, Q&A pp. 208–219)
• ★★★ | Fareed Zakaria GPS transcript | https://transcripts.cnn.com/show/fzgps/date/2018-05-20/segment/01 | 2018-05-20 | official broadcast transcript | External questioning on biography, five venture screens, iPhone platform fund, luck/recklessness, privacy, and OKRs. (reliability: A for official transcript/direct speech, B for retrospective history; access: full; locator: final Zakaria/Doerr speaker-turn block; no stable per-turn timestamps)
• ★★★ | Why the secret to success is setting the right goals | https://www.ted.com/talks/john_doerr_why_the_secret_to_success_is_setting_the_right_goals | recorded 2018-04-10; published 2018-06-11 | TED video/official transcript | Compact OKR provenance, purpose, implementation, case, limitation, and collaborator attribution. (reliability: A for direct speech, B for retrospective company claims; access: full; locator: 11:41 video and official cue-level transcript)
• ★★★ | OKRs Explained, with John Doerr | https://www.whatmatters.com/okrs-explained/why-okrs-john-joerr | undated; retrieved 2026-08-23 | official course video/transcript | Grove-to-Doerr provenance, Google adoption, transparency, grading, purpose, and judgment/culture boundary. (reliability: A for doctrine, B for history/adoption; access: full; locator: complete 4:01 embedded video, page Transcript paragraphs, and course headings)
• ★★★ | Keep Up the Momentum: John Doerr's OKR tips | https://www.whatmatters.com/okrs-explained/john-doerr-okr-tips | undated; retrieved 2026-08-23 | official course video/transcript | Guidance on cultural adaptation, patience, iteration, revision, and Ryan Panchadsaram's portfolio-support role. (reliability: A for doctrine, B for promotional benefits; access: full; locator: complete 2:26 embedded video, page Transcript paragraphs, and numbered list)
• ★★★ | How to decarbonize the grid and electrify everything | https://www.ted.com/talks/john_doerr_and_hal_harvey_how_to_decarbonize_the_grid_and_electrify_everything | 2020-10 | TED video/conversation with official transcript | Climate doctrine with Hal Harvey, physical-venture underwriting, policy targeting, measurement, and collaborator attribution. (reliability: A for direct speech, B for unverified figures/forecasts; access: full; locator: 20:17 video and official cue-level transcript)
• ★★★ | The Tim Ferriss Show #543 — John Doerr transcript | https://tim.blog/2021/11/04/john-doerr-transcript/ | 2021-11-04 | podcast/publisher transcript | Biography, Grove, Google decision evidence/group attribution, cleantech failure lessons, the 100-company denominator, OKRs, and climate. (reliability: A for direct speech, B for edited transcript/unaudited figures; access: full; locator: named question headings and turns; publisher typo warning retained)
• ★★★ | Masters in Business: John Doerr | https://ritholtz.com/2021/12/transcript-john-doerr/ | page dated 2021-12-05; prior SCOUT/broadcast inventory 2021-12-07 unresolved | Bloomberg Radio/Ritholtz interview transcript | Career, pivots, teams, Fisker versus Tesla, climate finance, and a roughly $1 billion/~70-venture denominator conflicting with Ferriss. (reliability: A for direct speech, B for edited transcript/unaudited figures; access: full; locator: question-and-answer blocks; transcript-defect caveat retained)
• ★★★ | An action plan for solving the climate crisis | https://www.ted.com/talks/john_doerr_and_ryan_panchadsaram_an_action_plan_for_solving_the_climate_crisis | 2021-10 | TED video/conversation with official transcript | Speed & Scale thesis, dated targets, climate-justice argument, and Ryan Panchadsaram's substantive collaborator voice. (reliability: A for identified speech, B for unverified figures/forecasts; access: full; locator: 11:28 video and official cue-level transcript)
• ★★★ | John Doerr and Ryan Panchadsaram: Speed & Scale's Reality Check | https://www.climateone.org/audio/john-doerr-and-ryan-panchadsaram-speed-scales-reality-check | recorded 2026-05-01 | Climate One podcast/live-event composite with indexed transcript | Current progress review, off-track admission, baseline/KR revisions, missed bus deadline, capital stack, and collaborator attribution. (reliability: A for identified speech, B for transcript/unverified figures; access: full; locator: timestamp index, named turns, and downloadable transcript; transcript-error warning retained)
• ★★★ | John Doerr — Charlie Rose archive | https://charlierose.com/guests/2525 | seven unique air dates, 1999-05-31–2008-10-20; retrieved 2026-08-23 | broadcast archive index, video, and official cue-level transcripts | Live index exposes 12 cards representing seven programs: five whole-show/segment duplicate pairs plus 2006-05-22 guest-host and 2008-10-20 panel cards; preserves boom, bust, innovation-policy, cleantech, commodity-risk, and crisis-era views. (reliability: A for metadata/identified speech, B for transcript errors/unsourced figures; access: full index and seven program/segment transcripts; locator: dated cards and cue timestamps; direct program routes used across the synthesis: 1999-05-31 program, 2002-03-18 program, 2004-11-16 program, 2006-01-06 program, 2006-05-22 guest-host program, 2006-12-29 program, and 2008-10-20 panel program; 2006-05-22 uses only Doerr's questions/direct disclosures, not Friedman's answers)

Contemporary profiles, firm history, and archives

• ★★★ | Kleiner Perkins' John Doerr: High Tech, High Finance | https://www.institutionalinvestor.com/article/2btgh0ottddrhcc4r3sw0/portfolio/kleiner-perkins-john-doerr-high-tech-high-finance | 1996 | contemporary long-form profile | Earliest high-yield biography found: family, education, Intel, KPCB recruitment, Silicon Compilers, network strategy, team, and contemporaneous claims. (reliability: B, reported contemporaneously with named witnesses; access: full; locator: paragraphs 65–94 and remainder)
• ★★★ | The Networker | https://www.newyorker.com/magazine/1997/08/11/the-networker | 1997-08-04/11 | contemporary magazine profile | External long-form narrative of Doerr's operating style, network, deals, and contradictions near decision dates. (reliability: B; access: full digitized article, with possible archive transcription errors; locator: section paragraphs)
• ★★★ | Inside the Silicon Valley Money Machine | https://pages.stern.nyu.edu/~aljungqv/courses/KPCB%2010.26.98.pdf | 1998-10-26 | archived Fortune article PDF | Contemporary account of KPCB decision process, portfolio-network mechanics, side funds, Amazon, Netscape, and partner/team attribution. (reliability: B, journalism preserved by an academic mirror; access: full; locator: PDF pp. 1–8, especially pp. 4–5)
• ★★★ | Thomas J. Perkins oral history | https://digicoll.lib.berkeley.edu/record/218111/files/klienerperkins00perkrich.pdf | interview 2001 | Bancroft Library oral-history transcript | Founder-generation witness for Kleiner Perkins history and the division of board roles—Doerr at Sun, Perkins at Compaq, Caufield at AOL. (reliability: A for institutional provenance/direct witness, B for retrospective claims; access: full; locator: transcript index and role-allocation passage)
• ★★★ | Burt McMurtry oral history | https://digicoll.lib.berkeley.edu/record/218493/files/mcmurtry_burt.pdf | 2009 | oral-history transcript | Independent participant account of Intuit's financing and a useful warning against hero-centric VC attribution. (reliability: A for institutional provenance/direct witness, B for retrospective claims; access: full; locator: 03-00:26 and 03-00:47:49)
• ★★★ | Weitek oral history panel | https://archive.computerhistory.org/resources/access/text/2012/10/102746380-05-01-acc.pdf | 2012-02-15 | Computer History Museum oral-history transcript | Third-party primary recollection of Doerr/KP passing on Weitek, useful as a documented miss/decision example. (reliability: A for witness record, B for retrospective claim; access: full; locator: CHM Ref X6384.2012, p. 8 of 38)
• ★★★ | SPARC oral history panel | https://archive.computerhistory.org/resources/access/text/2012/04/102745979-05-01-acc.pdf | 2011 | Computer History Museum oral-history transcript | Witness account of Doerr connecting Sun and Silicon Compilers; corroboration rather than Doerr direct voice. (reliability: A for witness record, B for retrospective claim; access: full; locator: CHM Ref X6182.2011, around p. 16 of 67)
• ★★★ | Brook Byers oral history | https://digicoll.lib.berkeley.edu/record/218191/files/byers_brook.pdf | interviews 2002–2005; published 2006 | oral-history transcript | Partner testimony on Doerr's recruitment, internal behavior, and sourcing; relationship bias requires explicit attribution. (reliability: A for institutional provenance/direct witness, B for retrospective claims; access: full; locator: search Doerr, reputation-risk passage)
• ★★ | Department of Electrical Engineering presentation by John Doerr | https://archives.library.rice.edu/repositories/2/top_containers/10057 | 2000-03-31 | Rice/Baker Institute VHS catalog record | High-value direct archive lead; media was not digitized or reviewed in this pass. (reliability: A for metadata only; access: catalog-only/request archive; locator: Baker Institute records, Box 3, exact title/date)
• ★★★ | The Google Guys | https://www.gq.com/story/google-larry-sergey | 2005 | reported company history | Detailed participant-based account of the founders' two-firm condition, Ram Shriram/Ron Conway mediation, equal VC split, and CEO search. (reliability: B; access: full; locator: financing/mediation paragraphs and Bill Campbell passage)
• ★★ | Google Is Searching for Success | https://www.forbes.com/1999/10/04/feat.html | 1999-10-04 | contemporary company profile | Decision-era search competitors, traffic and monetization questions, and founder/company context; issuer claims remain unverified. (reliability: B; access: full; locator: opening, competition, monetization, and executive-team paragraphs)
• ★★ | The dotcom king | https://www.theguardian.com/technology/1999/dec/19/internet.observerfocus | 1999-12-19 | retrospective newspaper profile | Reports a conflicting Netscape $18 million/15% financing account; retained as a conflict, not preferred over the filing or Fortune archive. (reliability: B; access: full; locator: Netscape financing paragraph)
• ★★ | Execution is Everything | https://25iq.com/2018/10/28/execution-is-everything/ | 2018-10-28 | 25iq synthesis/source-trail guide | Mandatory 25iq check; useful pointers to Doerr interviews and execution themes, but secondary and not independent evidence for quoted claims until originals are opened. (reliability: C as synthesis, B for linked-source discovery; access: full; locator: numbered extracts and outbound links)

Criticism, setbacks, litigation, and adverse accounts

• ★★★ | “Johnny Appleseed” for a Risky Field | https://www.washingtonpost.com/archive/politics/2002/11/13/johnny-appleseed-for-a-risky-field/b1e42b09-b4ba-4f91-a60a-f2d0c05216ee/ | 2002-11-13 | investigative profile/adverse case | Reports the At Home merger and KPCB distributions while separately reporting that Doerr did not sell his personal holding; those facts must not be conflated. (reliability: B; access: metered, full availability varies; locator: At Home merger/distribution and personal-holdings passages)
• ★★ | Cleanup Crew: Al Gore, John Doerr, David Blood | https://fortune.com/2008/02/12/al-gore-kleiner/ | 2008-02-12 | contemporary strategy profile | Captures the public cleantech buildout and participant claims near the pivot. (reliability: B; access: partial/paywalled; locator: accessible paragraphs and metadata)
• ★★★ | Why the Clean Tech Boom Went Bust | https://www.wired.com/2012/01/ff_solyndra/ | 2012-01 | critical sector history | Independent adverse context for capital intensity, policy, market, and venture-model mismatch; not a KPCB-only performance study. (reliability: B, reported sector history; access: full; locator: section headings/paragraphs)
• ★★ | John Doerr's Plan to Reclaim the Venture Capital Throne | https://www.forbes.com/sites/connieguglielmo/2013/05/07/john-doerrs-plan-to-reclaim-the-venture-capital-throne/ | 2013-05-07 | critical/strategic profile | Contemporary questioning of KPCB's competitive position and Doerr's response. (reliability: B; access: partial; locator: accessible Forbes paragraphs)
• ★★ | John Doerr on Zynga investment; Fisker remarks | https://techcrunch.com/2013/07/23/john-doerr-on-zynga-investment-we-own-60m-shares-so-i-want-that-stock-back-above-10-where-god-intended-it/ | 2013-07-23 | event report with direct remarks | Dated adverse-case voice on two visible troubled investments; the quoted share count is not a fund return. (reliability: B with direct remarks; access: partial; locator: event-report paragraphs)
• ★★★ | Pao v. Kleiner Perkins Caufield & Byers LLC, A136090 | https://www.casemine.com/judgement/us/5914fb65add7b049349ad2f7 | 2013-06-26 | California Court of Appeal opinion mirror | Primary legal text on arbitration, allegations, employment, fund/Managing-LLC structure, carried interest, and decision authority; opinion was unpublished and the route is a mirror. (reliability: A for court text if authenticated, B for mirror; access: full; locator: Factual and Procedural Background, especially paragraphs 117–140, and holding)
• ★★★ | Pao complaint, verdict, and docket retrieval set | https://www.morelaw.com/verdicts/case/CA/77241/ | complaint 2012-05-10; verdict 2015-03-27 | legal-document retrieval set | Docket route for CGC-12-520719; complaint allegations and the signed defense verdict remain separate records. Complaint mirror: https://www.scribd.com/document/257792940/Ellen-Pao-original-complaint-against-Kleiner-Perkins-May-2012; verdict route: http://www.sfsuperiorcourt.org/sites/default/files/pdfs/Pao%20Verdict.pdf. (reliability: A for underlying documents, C for hosting; access: mirrored/partial; locator: case number, complaint allegations, signed verdict questions)
• ★★★ | A Message from KPCB Partner, John Doerr | https://techcrunch.com/2012/05/30/kleiner-perkins-partner-john-doerr-speaks-out-on-lawsuit-our-firm-does-not-discriminate-based-on-gender/ | 2012-05-30 | defense statement preserved by press | Full text of Doerr/KPCB's response to Pao's allegations; the dead original and Forbes echo are not counted separately. (reliability: A for quoted statement, B for hosting; access: full; locator: reproduced statement beginning “A Message”)
• ★★ | Doerr testimony in Pao trial | https://www.courthousenews.com/billionaire-financier-takes-stand-in-gender-bias-case/ | 2015-03 | courtroom reporting | Reports Doerr's testimony; no official trial transcript was found, so use “reported/testified” language only. (reliability: B; access: full article route; locator: testimony and investigator-statement passages)
• ★★★ | Jury verdict in Ellen Pao v. Kleiner Perkins | https://techcrunch.com/2015/03/27/the-jury-has-reached-a-verdict-in-ellen-pao-vs-kleiner-perkins/ | 2015-03-27 | contemporaneous verdict report | Records the jury's rejection of Pao's claims and verdict-form answers; use alongside the form/transcript when retrieved. (reliability: B; access: full article route; locator: verdict table/answers)
• ★★ | Ellen Pao talks Kleiner, the trial, and hopes for a tech reset | https://www.axios.com/2017/12/15/ellen-pao-talks-kleiner-the-trial-and-her-hopes-for-a-tech-reset-1513305421 | 2017 | edited critic interview | Direct critic voice after the verdict; allegations, interpretations, and findings must remain separate. (reliability: A for Pao's stated view, B for editing/context; access: partial; locator: accessible interview paragraphs)
• ★★ | John Doerr on why he joined Bill Gates' billion-dollar energy fund | https://fortune.com/2016/12/21/john-doerr-on-why-he-joined-bill-gates-billion-dollar-energy-fund/ | 2016-12-21 | edited interview | Direct explanation of BEV strategy and failure tolerance; keep BEV distinct from KPCB funds. (reliability: A for Doerr remarks, B for context; access: partial/paywalled; locator: accessible Q&A paragraphs)
• ★★★ | How the Kleiner Perkins Empire Fell | https://fortune.com/longform/kleiner-perkins-vc-fall/ | 2019 | critical firm history | Detailed adverse account of strategy drift, organization, succession, cleantech, and competitive decline; requires response/primary checks. (reliability: B; access: partial/paywalled; locator: section headings/accessible paragraphs and correction)
• ★★★ | Sandys v. Pincus | https://law.justia.com/cases/delaware/supreme-court/2016/157-2016.html | 2016-12-05 | Delaware Supreme Court opinion | Pleading-stage decision finding reasonable doubt about some directors' independence, including ties involving Doerr/Gordon; not a finding of wrongdoing. (reliability: A for court text; access: full; locator: independence analysis and holding)
• ★★★ | Orgone Capital III, LLC v. Daubenspeck | https://law.justia.com/cases/federal/appellate-courts/ca7/18-1815/18-1815-2019-01-07.html | 2019-01-07 | Seventh Circuit opinion | Fisker-related claim was time-barred; the court expressly did not decide why Fisker failed or the defendants' merits responsibility. (reliability: A; access: full; locator: limitations holding and merits disclaimer)
• ★★★ | Alphabet shareholder derivative complaint | https://www.courthousenews.com/wp-content/uploads/2019/10/Alphabet.pdf | 2019-10-22 | filed federal complaint | Alleges oversight and demand-futility facts involving Doerr; a complaint is not proof and later disposition must control. (reliability: A for filed allegations only; access: full; locator: ¶20/p.7, ¶¶93–96/pp.28–29, counts pp.31–32)
• ★★★ | Alphabet Announces Settlement of Shareholder Derivative Litigation | https://www.sec.gov/Archives/edgar/data/1652044/000119312520254587/d87812dex991.htm | 2020-09-25 | SEC-filed company release | Records a derivative settlement and governance commitments without admissions; Doerr was one of multiple defendants. (reliability: A for company disclosure; access: full; locator: settlement terms and no-admission language)
• ★★★ | Roth v. Foris Ventures, LLC | https://law.justia.com/cases/federal/appellate-courts/ca9/22-16632/22-16632-2023-11-13.html | 2023-11-13 | Ninth Circuit opinion | Reverses/remands a Section 16(b) pleading dispute involving an Amyris investor group; it does not establish liability. (reliability: A; access: full; locator: disposition and remand)
• ★★★ | Alphabet derivative final judgment/settlement order | https://docs.justia.com/cases/federal/district-courts/california/candce/3%3A2021cv09388/388891/127 | 2025 | federal court order mirror | Records settlement disposition without an adjudication that Doerr committed the alleged conduct. (reliability: A for court text, B for mirror; access: full; locator: approval, release, and no-admission provisions)
• ★★ | Excite Inc. to Be Bought by At Home | https://www.washingtonpost.com/archive/business/1999/01/20/excite-inc-to-be-bought-by-at-home/21d67e5c-1a27-4513-a571-4daf9ef75938/ | 1999-01-20 | contemporary transaction report | Decision-era record of the $6.7 billion stock merger and access-plus-content thesis. (reliability: B; access: full archive route; locator: opening transaction and strategy paragraphs)
• ★★ | How cleantech tarnished Kleiner and VC star John Doerr | https://www.livemint.com/Companies/NvrjRq1uNhQITiHUnblzGI/How-cleantech-tarnished-Kleiner-and-VC-star-John-Doerr.html | 2012-10-29 | Reuters syndication/adverse account | Reports KPCB backing and Doerr-linked rescue context for MiaSolé; do not infer personal capital or governance. (reliability: B; access: full syndication; locator: MiaSolé and Kleiner passages)
• ★★ | MiaSolé sold for US$30 million | https://www.pv-magazine.com/2012/10/01/miasole-sold-for-us30-million_10008700/ | 2012-10-01 | contemporary trade report | Records factory reorganization, buyer search, market-price pressure, and the reported consideration, later disputed by the buyer. (reliability: B; access: full; locator: paragraphs 174–187)
• ★★ | MiaSolé takeover by Chinese firm OK'd | https://www.sfgate.com/business/article/miasole-takeover-by-chinese-firm-okd-4181378.php | 2013-01-09 | contemporary press | Reports more than $550 million of financing, a shareholder-document $30 million figure, and cheap-panel context. (reliability: B; access: full; locator: acquisition and financing paragraphs)
• ★★ | Hanergy Completes Acquisition of MiaSolé | https://www.hanergyamerica.com/news/newsmiasole.html | 2013-01-09 | buyer release | Records 100% acquisition, more than $550 million aggregate investor financing, and named KPCB participation; interested source. (reliability: B; access: full; locator: opening bullets and acquisition paragraphs)
• ★★ | Hanergy says MiaSolé deal terms undisclosed | https://usa.chinadaily.com.cn/business/2013-01/10/content_16101523.htm | 2013-01-10 | reported buyer response | Preserves the dispute that $30 million represented creditor repayment rather than disclosed acquisition consideration. (reliability: B; access: full; locator: buyer-response paragraphs)
• ★★ | DoorDash, Grubhub, and Caviar Hit With Labor Suits | https://www.wired.com/2015/09/doordash-grubhub-caviar-labor-suits/ | 2015-09-30 | contemporary adverse report | Corroborates worker-classification risk after the Series B; not proof the March IC saw the litigation. (reliability: B; access: full; locator: opening and classification paragraphs)
• ★★★ | Fisker financing chronology / Delaware opinion | https://www.ded.uscourts.gov/sites/ded/files/opinions/13-2100_1.pdf | 2015 | federal court opinion | Complaint-derived round, DOE, milestone, and warning chronology; allegations are not findings. (reliability: A for court text, provisional for recited allegations; access: full; locator: PDF pp. 3–6)
• ★★ | Fisker bankruptcy sale / Karma opinion | https://www.deb.uscourts.gov/sites/deb/files/opinions/fisker-emerald-v-karma-motion-dismiss-opinion-and-order_0.pdf | 2017 | bankruptcy court opinion | Records the court-approved asset sale and successor structure. (reliability: A for court text; access: full; locator: opening facts)
• ★★ | Kleiner Perkins vets close $350 million fund | https://www.axios.com/2018/06/12/kleiner-perkins-vets-close-350-million-fund | 2018-06-12 | reported new-fund strategy | Separates G2VP from KPCB and preserves direct explanation of physical-cleantech capital/time mismatch. (reliability: B with direct remarks; access: full; locator: fund and strategy paragraphs)
• ★★ | John Doerr on climate investing and Speed & Scale | https://www.forbes.com/sites/alexkonrad/2021/11/09/john-doerr-speed-scale-climate/ | 2021-11-09 | edited Q&A | Preserves self-reported $3.2 billion stake value and Beyond Meat 80× original-cost claim; unaudited and scope-undefined. (reliability: A for direct remarks, B for figures; access: full; locator: cleantech performance Q&A)
• ★★ | QuantumScape Securities Litigation settlement administration | https://www.quantumscapesecuritieslitigation.com/ | settlement administration route retrieved 2026-08-23 | official settlement notice portal | Records $47.5 million settlement administration and no-admission boundary; not proof the 2010 science thesis was fraudulent. (reliability: A/B for court-approved settlement notice; access: full; locator: settlement summary and notice)

Firm, deal, board, filing, and denominator records

• ★★★ | U.S. Patent 4,096,582, magnetic-bubble circuits | https://patents.justia.com/patent/4096582 | filed 1974-05-30; issued 1978-06-20 | patent record/full text | Names L. John Doerr III as co-inventor and Monsanto as assignee, establishing pre-Intel engineering work. (reliability: A for patent record; access: full; locator: filing history, inventors, assignee, claims)
• ★★★ | John Doerr — Kleiner Perkins profile and attributed portfolio | https://www.kleinerperkins.com/people/john-doerr/ | live page retrieved 2026-08-23 | current firm biography/portfolio catalog | Current role and a broad company list with partnered-since dates and partner attribution; promotional, self-selected, and not a denominator or return record. (reliability: A for current self-description, B for historical claims; access: full; locator: profile intro, portfolio rows, embedded company narratives)
• ★★★ | Amazon: Reimagining commerce | https://www.kleinerperkins.com/perspectives/amazon/ | live page retrieved 2026-08-23 | firm case study | Detailed sourcing, recruiting, business-development, and team claims; cross-check with company participants and filings. (reliability: B, interested retrospective; access: full; locator: paragraphs and Partners field)
• ★★★ | Google: Organizing the world's information and making it searchable | https://www.kleinerperkins.com/perspectives/google/ | live page retrieved 2026-08-23 | firm case study | Detailed intro, diligence, recruiting, OKR, coaching, and board claims; preserve Sequoia and founder/team roles. (reliability: B, interested retrospective; access: full; locator: paragraphs and Partners field)
• ★★★ | Netscape: Bringing the internet to the world through the web browser | https://www.kleinerperkins.com/perspectives/netscape/ | live page retrieved 2026-08-23 | firm case study | Dates 1994 financing/board and offers recruiting/outcome narrative; requires financing and board records. (reliability: B, interested retrospective; access: full; locator: paragraphs and Partners field)
• ★★★ | Google Receives $25 Million in Equity Funding | https://googlepress.blogspot.com/1999/06/google-receives-25-million-in-equity.html | 1999-06-07 | contemporaneous company announcement | Primary decision-date evidence: aggregate $25 million round led by Sequoia and KPCB; Moritz and Doerr join board. It does not disclose the check split. (reliability: A; access: full; locator: headline/subheadline and opening announcement)
• ★★★ | Google Form S-1 | https://www.sec.gov/Archives/edgar/data/1288776/000119312504073639/ds1.htm | filed 2004-04-29 | SEC registration statement | Primary board, ownership, company, and offering record near IPO; use exact filing tables, not later return lore. (reliability: A; access: full; locator: Management, Principal and Selling Stockholders, Certain Relationships)
• ★★★ | Amazon.com Amendment No. 2 to Form S-1 | https://www.sec.gov/Archives/edgar/data/1018724/000089102097000603/0000891020-97-000603.txt | filed 1997-04-21 | SEC registration statement | Authoritative board, related-transaction, and KPCB ownership record; nominee attribution is not personal ownership. (reliability: A; access: full; locator: Management near line 2773, Related Transactions near 3043, ownership table/footnotes near 3135)
• ★★★ | Netscape Form S-1 | https://simson.net/ref/1995/netscape-s1.pdf | 1995; data as of 1995-05-31 | registration statement archival mirror | Reports KPCB VII and related-vehicle holdings in Doerr's row with a pecuniary-interest disclaimer; original SEC-hosted text was not recovered. (reliability: A for filing text, B for mirror; access: full; locator: printed p. 48, principal stockholders and footnotes)
• ★★★ | Sun Microsystems 2003 definitive proxy | https://www.sec.gov/Archives/edgar/data/709519/000119312503057278/ddef14a.htm | 2003 | SEC proxy | Dates Doerr's Sun directorship to 1982. (reliability: A; access: full; locator: printed p. 4, director biography)
• ★★★ | Compaq 2001 definitive proxy | https://www.sec.gov/Archives/edgar/data/714154/000089155401501345/d25040_def14a.htm | 2001 | SEC proxy | Identifies Tom Perkins—not Doerr—as KPCB's Compaq director, preserving sponsor/governance attribution. (reliability: A; access: full; locator: Perkins director biography)
• ★★★ | Bloom Energy Form S-1 | https://www.sec.gov/Archives/edgar/data/1664703/000119312518190488/d96446ds1.htm | 2018-06 | SEC registration statement | Dates Doerr's board service to May 2002 and gives a contemporaneous cross-board roster. (reliability: A; access: full; locator: management/director biography)
• ★★★ | Amyris 2023 annual report/proxy | https://www.sec.gov/Archives/edgar/data/1365916/000119312523097883/d383171dars.pdf | 2023 | SEC annual/proxy PDF | Separates Foris, VVT, Doerr, and Ann Doerr ownership/relationship footnotes; deemed beneficial ownership is not freely owned personal stock. (reliability: A; access: full; locator: printed p. 35, security ownership and vehicle footnotes)
• ★★★ | QuantumScape Form 3 | https://www.sec.gov/Archives/edgar/data/1811414/000120919120062067/xslF345X02/doc3.xml | 2020-12-07 | SEC ownership report | Exact KPCB vehicle and nominee/pecuniary-interest disclosure, useful for preventing personal-capital misattribution. (reliability: A; access: full; locator: ownership table and footnotes)
• ★★★ | Intuit 2003 definitive proxy | https://www.sec.gov/Archives/edgar/data/896878/000089161803005028/f93191dedef14a.htm | 2003 | SEC proxy | Independently dates Doerr's Intuit directorship to August 1990 and supplies role/governance context. (reliability: A; access: full; locator: director biography and board tables)
• ★★★ | DoorDash 2021 definitive proxy | https://www.sec.gov/Archives/edgar/data/1792789/000162828021008245/doordashinc-def14a.htm | 2021 | SEC proxy | Independently dates Doerr's board service and identifies KPCB-related beneficial ownership. (reliability: A; access: full; locator: director biographies, Principal Stockholders, related footnotes)
• ★★★ | Alphabet 2024 definitive proxy | https://www.sec.gov/Archives/edgar/data/1652044/000130817924000612/lgoog2024_def14a.htm | 2024-04-26 | SEC proxy | Primary dated record for Doerr's board tenure since 1999, title, independence, committee, and holdings context. (reliability: A; access: full; locator: printed pp. 4, 7–8, director biography/ownership tables)
• ★★★ | Kleiner Perkins is splitting up | https://www.axios.com/kleiner-perkins-is-splitting-up-1536933575-dccd3c28-2900-4ba4-ab23-b8e3e3f89a55.html | 2018-09-14 | reported firm reorganization with official email | Separates DGF I–III management and economics from the continuing venture practice, constraining attribution across the growth-team split. (reliability: B; access: full; locator: Ted Schlein email and DGF I–III mechanics)
• ★★ | John Doerr investment profile — NFX Signal | https://signal.nfx.com/investors/john-doerr | live page retrieved 2026-08-23 | discovery/denominator route | Page simultaneously displayed 121 “Investments on Record” and a 75-record table header, so neither figure is accepted without methodology. (reliability: C; access: full; locator: page count and table header)
• ★★★ | Amazon.com Form S-1 | https://www.sec.gov/Archives/edgar/data/1018724/0000891618-97-001309.txt | filed 1997-03-24 | SEC registration statement | Risk factors, Series A terms, voting agreement, Doerr board date, and decision-near company evidence. (reliability: A; access: full; locator: Risk Factors, Certain Transactions, Management, voting agreement)
• ★★ | Sun Microsystems — Pioneering open systems | https://www.kleinerperkins.com/perspectives/sun-microsystems/ | live page retrieved 2026-08-23 | firm case study | Reports a $1.7 million November 1982 investment, early sales, thesis, and Doerr board role; sponsor retrospective, not a cash ledger. (reliability: B; access: full; locator: investment, early-sales, and board paragraphs)
• ★★★ | United States v. Microsoft Corp. and Intuit Inc. | https://www.justice.gov/atr/case/us-v-microsoft-corp-and-intuit-inc | 1995 | DOJ case record | Official competition record for the challenged and abandoned Microsoft/Intuit acquisition. (reliability: A; access: full; locator: case summary and filings)
• ★★ | DoorDash raises Series B from Kleiner Perkins | https://fortune.com/2015/03/17/exclusive-doordash-raises-new-vc-from-kleiner-perkins/ | 2015-03-17 | contemporary financing report | Reports nine markets, prior financing, a nearly $600 million valuation, and an “around $35 million” round figure; reconcile with the filing's $45 million aggregate preference. (reliability: B; access: full/partial route; locator: opening financing and market paragraphs)
• ★★ | DoorDash Wants to Own the Last Mile | https://www.wired.com/2015/11/doordash-wants-to-own-the-last-mile/ | 2015-11-09 | contemporary operating profile | Corroborates crowded alternatives, logistics-platform thesis, commission example, and operating/labor risks after the Series B. (reliability: B; access: full; locator: competition, commission, logistics, and courier passages)
• ★★★ | DoorDash Form S-1 | https://www.sec.gov/Archives/edgar/data/1792789/000119312520292381/d752207ds1.htm | filed 2020-11-13 | SEC registration statement | Series B $5.68 price/$45 million aggregate preference, board/designee, DGF ownership, and company risk record. (reliability: A; access: full; locator: Preferred Stock, Management, Principal Stockholders)
• ★★★ | DoorDash Form 3 | https://www.sec.gov/Archives/edgar/data/1792789/000089924320033051/xslF345X02/doc3.xml | filed 2020-12-08 | SEC ownership report | Separates DGF II/DGF Founders II Series B, C, and H securities and nominee/shared-control footnotes. (reliability: A; access: full; locator: ownership table and footnotes)
• ★★★ | Bloom Energy draft registration statement | https://www.sec.gov/Archives/edgar/data/1664703/000095012316015543/filename1.htm | submitted 2016 | SEC draft registration statement | Company history identifies KPCB as first venture investor in 2002; supports entry-date reconciliation. (reliability: A; access: full; locator: History and financing passages)
• ★★★ | Bloom Energy 2019 Form 10-K | https://www.sec.gov/Archives/edgar/data/1664703/000166470320000013/be-20191231.htm | filed 2020-03-31 | SEC annual report | Accounting-correction, risk, loss, governance, and operating follow-through for the mixed case. (reliability: A; access: full; locator: restatement/control, risk factors, financial statements)
• ★★★ | Coursera 2022 definitive proxy | https://www.sec.gov/Archives/edgar/data/1651562/000119312522091748/d331267ddef14a.htm | filed 2022-03-30 | SEC proxy | Decomposes KPCB XIV nominee shares from WindyHill, trusts, VVT, KPIC, Lupum, and Portico Libre. (reliability: A; access: full; locator: principal-holder table, footnote 6)
• ★★★ | Amyris IPO prospectus | https://www.sec.gov/Archives/edgar/data/1365916/000119312510218728/d424b4.htm | 2010-09-28 | SEC 424(b)(4) | Ex-ante platform thesis, scale-up risks, losses, Doerr board role, and KPCB-affiliated ownership. (reliability: A; access: full; locator: risk factors lines 686–733, management 3729–3731, ownership 4737–4851)
• ★★★ | Amyris Schedule 13D/A | https://www.sec.gov/Archives/edgar/data/1365916/000119312523208226/d539259dsc13da.htm | filed 2023-08-11 | SEC ownership filing | Separates Foris common, debt-conversion, and warrant shares from VVT, Clarus, Perrara, KPCB XII, and direct interests. (reliability: A; access: full; locator: lines 76–93 and 139–142)
• ★★★ | Amyris 2022 Form 10-K | https://www.sec.gov/Archives/edgar/data/1365916/000136591623000023/amrs-20221231.htm | filed 2023-03-16 | SEC annual report | Cash burn, liquidity, debt, and $80 million Foris secured-note evidence. (reliability: A; access: full; locator: liquidity and Issuance of Foris Senior Note)
• ★★★ | Amyris Chapter 11 / DIP release | https://www.sec.gov/Archives/edgar/data/1365916/000136591623000094/a20230810xex991pressrelease.htm | 2023-08-10 | SEC-filed issuer release | Records Chapter 11 filing and Foris commitment of up to $190 million in DIP financing. (reliability: A for issuer disclosure; access: full; locator: opening restructuring and DIP paragraphs)
• ★★★ | Amyris plan effective / equity cancellation | https://www.sec.gov/Archives/edgar/data/1365916/000119312524132953/d804267d8k.htm | 2024-05-07 | SEC Form 8-K | Old-equity cancellation, DIP conversion into new common, the internal Items 1.02/3.02 recipient conflict, and successor directors. (reliability: A; access: full; locator: Items 1.02, 3.02–3.03, and 5.01–5.02)
• ★★★ | Zynga IPO prospectus | https://www.sec.gov/Archives/edgar/data/1439404/000119312511343682/d198836d424b4.htm | 2011-12-16 | SEC 424(b)(4) | $10 IPO, voting structure, and KPCB board designation. (reliability: A; access: full; locator: cover, risk factors, voting agreement)
• ★★★ | Zynga 2013 definitive proxy | https://www.sec.gov/Archives/edgar/data/1439404/000119312513171361/d497223ddef14a.htm | filed 2013-04 | SEC proxy | Doerr board date/grant and approximately 60.1 million shares across KPCB XIII/DGF entities. (reliability: A; access: full; locator: director biography and principal-stockholder footnotes)
• ★★★ | Take-Two completes Zynga acquisition | https://www.sec.gov/Archives/edgar/data/946581/000119312522160004/d306239dex999.htm | 2022-05-23 | SEC-filed issuer release | Final $3.50 cash plus 0.0406 Take-Two-share consideration and close date. (reliability: A; access: full; locator: paragraphs 20–25)
• ★★★ | QuantumScape Form S-1 | https://www.sec.gov/Archives/edgar/data/1811414/000119312520320220/d53940ds1.htm | filed 2020-12-17 | SEC registration statement | Legacy board date, Series A class terms, single-layer versus 100-plus-layer risk, business, and commercialization warnings. (reliability: A; access: full; locator: Business, Management, Risk Factors, preferred-stock table)
• ★★★ | QuantumScape investor presentation | https://www.sec.gov/Archives/edgar/data/1811414/000121390020026128/ea126929ex99-3_kensington.htm | filed 2020-09-03 | SEC-filed issuer presentation | Public-transaction forecasts and claims, used only after the 2010 freeze and labeled issuer projections. (reliability: A for filed artifact, B for forecasts; access: full; locator: technology, production, and forecast slides)
• ★★★ | QuantumScape 2021 prospectus | https://www.sec.gov/Archives/edgar/data/1811414/000119312521182920/d190995d424b3.htm | filed 2021-06-08 | SEC 424(b)(3) | Separates KPCB XIV Associates, VVT, trusts, and side entities in the Doerr ownership row. (reliability: A; access: full; locator: lines 82–86, footnotes 7–13)

Institutional biography and historical context

• ★★★ | John Doerr — President's Economic Recovery Advisory Board biography | https://obamawhitehouse.archives.gov/administration/eop/perab/members/doerr/ | archived Obama administration page | official frozen biography | Independent institutional route for Intel/KPCB chronology, patents, degrees, public roles, and then-current boards; biography itself likely supplied/approved by subject. (reliability: A for appointment/current-at-time facts, B for biography; access: full; locator: biography paragraphs 143–147)
• ★★★ | Rice ECE Hall of Fame — John Doerr | https://eceweb.rice.edu/alumni-giving/hall-fame | current institutional record | university alumni register | BSEE 1973 and MSEE 1974. (reliability: A for Rice degrees/dates; access: full; locator: John Doerr row)
• ★★★ | HBS Alumni Achievement Awards — John Doerr | https://www.alumni.hbs.edu/news-insights/alumni-achievement-awards/2008 | 2008 | university alumni record | MBA 1976. (reliability: A for HBS degree/date; access: full; locator: John Doerr heading)
• ★★ | Rice engineering alumnus Doerr elected to National Academy of Engineering | https://news.rice.edu/news/2026/rice-engineering-alumnus-doerr-elected-national-academy-engineering | 2026-02-12 | university announcement | Current institutional biography, St. Louis origin, education, career, philanthropy, and NAE recognition. (reliability: A for Rice/NAE institutional facts, B for career summary; access: full; locator: opening and recognition paragraphs 13–24)
• ★★★ | About NewSchools Venture Fund — Our Story | https://www.newschools.org/our-story/ | live history retrieved 2026-08-23 | nonprofit institutional history | Names Kim Smith, John Doerr, and Brook Byers as 1998 founders; use for education/nonprofit chronology. (reliability: A for current institutional self-history, B for retrospective details; access: full; locator: founding/history paragraphs)
• ★★ | NewSchools Venture Fund Ten-Year Report | https://www.newschools.org/wp-content/uploads/2016/01/10YearReport4.pdf | covers 1998–2008; hosted 2016 | nonprofit retrospective report PDF | Historical program, team, and capital context for the organization's first decade; outcomes are self-reported. (reliability: B; access: full; locator: PDF pp. 1–35, founding timeline and metrics)
• ★★ | Introducing the Stanford Doerr School of Sustainability | https://news.stanford.edu/stories/2022/05/stanford-doerr-school-sustainability-universitys-first-new-school-70-years-will-accelerate-solutions-global-climate-crisis | 2022-05 | university announcement | Primary institutional record for the $1.1 billion lead gift and school formation; philanthropy is not a venture result. (reliability: A for Stanford/gift announcement; access: full; locator: opening and gift/governance paragraphs)
• ★★★ | Venture Capital and Cleantech: The Wrong Model for Energy Innovation | https://energy.mit.edu/wp-content/uploads/2016/07/MITEI-WP-2016-06.pdf | 2016-07 | MIT Energy Initiative working paper | Independent quantitative context for cleantech venture outcomes and capital-intensity/model mismatch; supports sector baselines, not KPCB-specific cash flows. (reliability: A/B, academic working paper; access: full; locator: PDF tables, methods, and results)

Coverage assessment

Breadth and source-class result

• 118 annotated source rows were retained: 11 books/signed writings, 15 talks/interviews/testimony routes, 13 contemporary profiles/archive works, 31 criticism/adverse/legal works, 40 firm/deal/filing/denominator records, and 8 institutional/context routes. Of these rows, 107 are checked and 11 remain open. This is a row count, not an “independent source” claim: related Stanford clips, filing families, and excerpt/publisher routes can share an underlying work.
• Core primary/direct evidence now includes authored books and signed writings, Edge/Ritholtz/What Matters interviews, Stanford/TED/Charlie Rose direct voice, White House and Senate records, a patent, issuer announcements, SEC filings/proxies, court opinions and settlements, and institutional archives.
• External challenge evidence includes contemporary 1996–2002 profiles, cleantech and troubled-company reporting, Pao plaintiff/defense/verdict materials, derivative complaints and dispositions, participant oral histories, and the 2019 critical firm history.

Canonical query battery and name variants used

• Names/roles: John Doerr, L. John Doerr, Louis John Doerr III, Kleiner Perkins, KPCB, KP, chair, general partner, board, director, Intel, Silicon Compilers.
• Vehicles/entities: KPCB Holdings, KPCB fund, Managing LLC, side fund, Doerr Capital, family foundation, Breakthrough Energy Ventures, Form D, ADV, proxy, S-1.
• Deals/teams: Amazon, Google, Intuit, Netscape, Sun, Compaq, Symantec, Zynga, Fisker, DoorDash, cleantech; paired with term sheet, investment memo, board, ownership, recruit, failure, return, IRR, MOIC, and partner names.
• Writings/direct voice: book, memo, letter, essay, speech, podcast, transcript, oral history, interview, TED, Stanford, Senate testimony, and date/decade variants from the 1970s through 2026.
• Mandatory adverse battery: criticism, underperformance, decline, failure, Fisker, Zynga, cleantech bust, Ellen Pao, discrimination, lawsuit, complaint, verdict, SEC, DOJ, fraud, and conflict of interest.

Denominators, attribution, and what is not yet knowable

• The live Kleiner Perkins profile is a broad self-selected attributed portfolio, not a complete opportunity, portfolio, vehicle, failure, or cash-flow denominator.
• The opened filings identify particular board roles and entity holdings at specific dates. They do not establish personal check sizes, partner carry, realized proceeds, fund IRR, or sole decision credit.
• Google's contemporaneous announcement discloses one aggregate $25 million Sequoia/KPCB round. No opened primary record splits that amount between firms; later claims that assign a personal or firm check must remain unresolved.
• Firm case studies are valuable retrieval maps, but must not silently convert KPCB/KP team capital, founder execution, co-investor work, company boards, or later outcomes into a John Doerr-only record.
• No audited KPCB fund cash-flow series, LP letter corpus, investment-committee memo set, anti-portfolio, or complete cleantech vehicle schedule was found in this pass.

Archives, oral history, and access barriers

• Direct archive routes opened include Edge's full transcript and alternate node, CHM Weitek/SPARC transcripts, Bancroft Perkins/McMurtry/Byers transcripts, Rice's VHS catalog, SEC EDGAR, the Google press archive, GovInfo, and Stanford's clip/transcript route.
• No dedicated John Doerr oral history, personal-papers collection, IC/LP correspondence archive, or digitized copy of the Rice 2000 VHS was identified. The Rice item is catalog-only/request-access; oral histories by Perkins, McMurtry, and Byers are contextual witness evidence, not Doerr direct voice.
• Access barriers preserved rather than papered over: the Washington Post route is metered and availability varies; the 1998 Fortune article is full only through an academic archive mirror, while later Fortune items are partial/paywalled; the Netscape S-1 is full on an archival mirror because an official SEC route was not recovered; Pao trial documents are mirrored/partial and no official trial transcript surfaced; some podcast transcripts are edited or automated.
• Books were counted once per underlying work. Publisher metadata, authorized excerpts, and action-plan documents remain separate rows only where they expose distinct content; previews were not padded into separate rows.

Criticism/lawsuit result and legal boundary

• The adverse sweep surfaced Pao's complaint, KPCB/Doerr's response, a procedural arbitration opinion, trial reporting, the 2015 defense verdict, and Pao's later critic account, alongside At Home, cleantech/Fisker, Zynga, Amyris, and derivative-governance records.
• Pao's allegations, KPCB/Doerr's denial, reported testimony, and the defense verdict are distinct evidence classes. The Alphabet complaint likewise records allegations; the settlement records include no-admission language. Sandys was a pleading-stage independence decision, Orgone was time-barred without a merits decision on Fisker, and Roth was a remand without liability.
• Targeted SEC/DOJ/enforcement searches did not surface a Doerr-specific enforcement action in the opened results. This is a search result, not proof none exists; later legal work should use PACER/state dockets, regulator databases, exact entities, and date-bounded queries.

Historical context and completed-peer comparisons

• Industry context retained: Perkins, McMurtry, and Byers oral histories for partner-generation context; the 1996–98 contemporary profiles for the KPCB network model; and MIT's 2016 cleantech working paper for the venture/energy model mismatch.
• Vinod Khosla (investors/026-vinod-khosla/) is the strongest completed peer: same institutional platform, overlapping Sun/KPCB/cleantech history, but later a different firm/vehicle architecture. Compare attribution and cleantech portfolios without merging partners or funds.
• Michael Moritz (investors/027-michael-moritz/) is the second defensible completed peer: contemporaneous co-lead/co-board evidence in Google's 1999 round and a competing partnership model at Sequoia. The joint announcement is especially useful for preventing unilateral Google credit.
• Bob Kagle (investors/025-bob-kagle/) is the third completed peer: the same internet-era power-law and board-service questions, but Benchmark's equal-partner architecture and eBay outlier provide a contrasting institutional model. Compare all three without implying compatible personal-return evidence.

Saturation and highest-value retrieval leads

• The first attempted stop sequence was restarted after a broad criticism query surfaced Doerr's May 30, 2012 response to the Pao lawsuit and a separate filed Alphabet derivative complaint; both are represented above. The final three searches after that restart were:
  1. "A Message from KPCB Partner, John Doerr" "false allegations" — only the already captured Forbes preservation of the 2012 statement; no new material fact.
  2. "Mutually Exclusive Magnetic Bubble Propagation Circuits With Discrete Elements" "L. John Doerr III" — no results; the patent family had already been captured from Patent 4,096,582; no new material fact.
  3. "John Doerr" "CHM Ref: X6384.2012" Weitek — only the already captured CHM Weitek panel, p. 8; no new material fact.
• Three consecutive final searches produced no new material fact; the pass reached practical web saturation subject to the access gaps below.
• Highest-value next retrievals are: Silicon Compilers formation/exit records; Intel employment/patent records; KPCB Form D/ADV and Managing-LLC documents; Google/Amazon/Netscape cap tables and board records; trial transcript/verdict form/exhibits in Pao; participant interviews for Amazon/Google/Intuit/Netscape; fund/LP documents and cash flows; the missing CHM Doerr interview; and a complete cleantech portfolio/vehicle ledger including adverse outcomes.

Source notes