Sources and provenance

#1. How this course was built

Like this platform's Marketing course, this course was researched from scratch, not compiled from an existing brief. Six research documents were built first, from live web research conducted at build time rather than from memory alone, each one fetching primary sources directly where a primary source existed and flagging plainly where it didn't: research/first-customer/customer-development.md, concierge-mvp-case-histories.md, founder-led-outreach.md, pricing-and-closing.md, post-sale-flywheel.md, and competitor-sweep.md. Every claim in those documents carries its own confidence marking, and every unconfirmed figure is listed twice — once inline, once again in a flagged-issues section at the end — a discipline this course's nine lessons inherited directly, using the same [Established] / [Directional] / [Speculative] tag system the rest of this platform's research-backed courses use.

The competitor-sweep.md document is a different kind of research file from the other five: it doesn't itself supply lesson citations so much as it audits what Y Combinator's own library, First Round Capital, Lenny Rachitsky's newsletter, Indie Hackers, Traction, and The Mom Test already cover on this exact topic, and names the real gap those six sources leave — no single one of them bridges customer-development validation (does a real problem exist) with acquisition tactics (how do you actually find the person who has it), and none of them honestly accounts for a founder with no warm network at all. That gap is why this course exists in the shape it does, and it's why the Marlowe throughline — an explicitly invented example, stated once per lesson, never presented as a real company — runs through every module rather than switching to a new hypothetical business each time: a continuous, specific decision is more honest about how these frameworks actually combine than nine unrelated examples would be.

Two disciplines carried over directly from the underlying research and are worth stating plainly here rather than leaving implicit. First, a folk statistic repeated across dozens of SEO-shaped sites with no named study behind it — "63% of B2B revenue comes from referrals," "founder-led sales converts 2–3x higher than delegated outreach" — was checked against a primary source before use, and where no primary source existed, the course states the real, better-evidenced, more modest finding instead (see §3) or refuses the number outright and says so (see §4), rather than smoothing an unsourced figure into confident prose. Second, several real, well-documented founder stories in wide circulation — Airbnb's early photography, Stripe's "Collison installation" — carry specific numbers (a camera price, a host count, an exact percentage lift) that could not be traced to any primary disclosure anywhere in this research pass. Where that happened, the lessons present the event as verified and the number as folklore, explicitly, rather than repeating the number because it's the version everyone else uses.

#2. Tier 1 — primary or near-primary sources

• Hovland, C., & Weiss, W., "The Influence of Source Credibility on Communication Effectiveness," Public Opinion Quarterly, 1951, and Hovland, C., Janis, I., & Kelley, H., Communication and Persuasion (Yale University Press, 1953) — the founding, peer-reviewed source-credibility research: expertise and trustworthiness as the two dimensions a listener scores a message on. Corroborated across five decades of subsequent research by Chanthika Pornpitakpan's 2004 meta-analytic review, "The Persuasiveness of Source Credibility," Journal of Applied Social Psychology. Cited in Why the First Customer Is Different, Founder-Led Outbound, and The Actual Conversation as the mechanistic reason a founder's own unscaled outreach outperforms a scripted or templated one.
• Fisher, Robert J., "Social Desirability Bias and the Validity of Indirect Questioning," Journal of Consumer Research 20(2), 1993, and the peer-reviewed 2022 qualitative-methods paper on the same effect in interview settings (PMC9749714) — the independently studied bias underneath why a friend, a warm introduction, or an emotionally invested conversation partner tends to say yes regardless of what they actually believe. Cited in Why the First Customer Is Different, The Mom Test, and The Actual Conversation.
• Granovetter, Mark, "The Strength of Weak Ties," American Journal of Sociology 78(6), 1973, and its large-scale causal replication, Rajkumar, K., Saint-Jacques, G., et al., "A causal test of the strength of weak ties," Science 377(6612), 2022 (a five-year, ~20-million-user randomized experiment on LinkedIn) — the evidence that the person most likely to produce a real opportunity is a moderately weak tie, not a close friend and not a stranger. Cited in Why the First Customer Is Different, The Mom Test, Finding Your First Buyer, and From One Customer to a Flywheel.
• Tversky, A., & Kahneman, D., "Judgment under Uncertainty: Heuristics and Biases," Science, 185(4157), 1974 (the rigged wheel-of-fortune anchoring experiment), and Ariely, D., Loewenstein, G., & Prelec, D., "Coherent Arbitrariness: Stable Demand Curves Without Stable Preferences," Quarterly Journal of Economics, 118(1), 2003 (the SSN-digit pricing experiment) — the peer-reviewed foundation for why the first number said in a sales conversation anchors every number that follows it, arbitrarily. Cited in Pricing Your First Sale and The Actual Conversation.
• Tversky, A., & Kahneman, D., "Belief in the Law of Small Numbers," Psychological Bulletin, 76(2), 1971 — the peer-reviewed finding that people, including trained researchers, systematically over-trust small samples. Cited in From One Customer to a Flywheel and Diagnosing a "No" as the discipline against reading one sale, or one rejection, as more than it is.
• Schmitt, P., Skiera, B., & Van den Bulte, C., "Referral Programs and Customer Value," Journal of Marketing 75(1), 2011 (a 2011 MSI/H. Paul Root Award-winning study tracking 5,181 referred customers against 4,633 non-referred customers at a German retail bank over 33 months) — the primary, disclosed source for the referred-customers-are-16%-more-valuable, 18%-stickier finding this course uses in place of the unsourced "63% of B2B revenue" folklore. Cited in How to Use This Course and, fully attributed, in From One Customer to a Flywheel.
• Eisenmann, Thomas R., & Mazzanti, Lisa, Quincy Apparel (A), Harvard Business School Case 815-067 (Feb. 2015, rev. 2022) — a case built from firsthand founder interviews and company financial data, the strongest evidence tier this course uses for a single company history. Cited in The Mom Test as the documented example of a founding team validating the problem correctly and the operational capability not at all.
• Steve Blank's own primary account, read directly from steveblank.com (specifically steveblank.com/2012/03/05/search-versus-execute) — the direct-quoted source for "a startup is a temporary organization designed to search for a repeatable and scalable business model," and for the Discovery/Validation staging this course's own module structure follows. Cited in Why the First Customer Is Different, The Mom Test, From One Customer to a Flywheel, and Diagnosing a "No".
• Eric Ries, The Lean Startup (Crown Business, 2011), and Ries's own MVP definition as published by leanstartup.co, the organization he founded — the primary source for "validated learning" and for the definition of a minimum viable product as a learning instrument, not a smaller product. Cited in Why the First Customer Is Different, The Mom Test, and Diagnosing a "No".
• Paul Graham, "Do Things That Don't Scale" (2013), read directly from paulgraham.com/ds.html, and Graham's own on-the-record confirmation of the term "Collison installation" and its continued live use in YC teaching (@paulg on X, June 2019 and August 2026) — the primary essay behind this course's entire "manual, founder-led, deliberately unscalable" argument. Cited in Why the First Customer Is Different, Finding Your First Buyer, Pricing Your First Sale, and From One Customer to a Flywheel.
• Brian Chesky's own first-person account of Airbnb's founders personally photographing early New York host listings, given on the Masters of Scale podcast, and Tony Xu's own first-person account of DoorDash's founders acting as the delivery drivers themselves, given in a November 2015 Stanford Graduate School of Business interview — direct, on-record founder quotes. Both cited in Why the First Customer Is Different and Finding Your First Buyer, with the event itself treated as verified and the specific numbers that circulate around each story (a camera price, a host count, a restaurant count) treated separately as unsourced and not repeated.
• Nick Swinmurn's own retelling of Zappos's 1999 origin, corroborated by independent trade press (WWD) — the founding case for what Eric Ries later named the "concierge MVP." Cited in Why the First Customer Is Different and Finding Your First Buyer.
• Manuel Rosso's Food on the Table, corroborated in Eric Ries's own blog (Startup Lessons Learned) — the case that gives the "concierge MVP" pattern its name. Cited in Why the First Customer Is Different and Finding Your First Buyer.
• First Round Capital, "Levels of PMF" (firstround.com/levels), built from named, attributed operator interviews — the source for treating a first sale as evidence at the "Nascent" stage of product-market fit, not proof of a repeatable model. Cited in Why the First Customer Is Different.
• Y Combinator's own institutional guidance: the YC Startup Library's pricing guidance, and Michael Seibel's own recorded, published lecture content on charging as a diagnostic filter — YC's own published position, not a paraphrase. Cited in Why the First Customer Is Different and Pricing Your First Sale.
• Yuri Sagalov, "The Counter-Intuitive Pricing of Enterprise Sales," published directly on Y Combinator's own blog — a named founder's own account, published on the platform's own site. Cited in Pricing Your First Sale and The Actual Conversation.
• Dharmesh Shah, "Giving Your Software Away For Free," OnStartups — a named co-founder's own published essay. Cited in Pricing Your First Sale.
• Bessemer Venture Partners, "Design partners: the pre-launch edge most AI founders ignore" (bvp.com) — a named venture firm's own published framework for a structured, time-boxed free pilot with a defined conversion trigger. Cited in Pricing Your First Sale and The Actual Conversation.
• Stripe's own documentation for freelancers and sole operators using Stripe Invoicing — a platform's own published documentation, confirming no prior processing history is required to invoice a first client. Cited in The Actual Conversation.
• Nathan Barry's own account of building ConvertKit, published directly on his blog (nathanbarry.com/30k) and on X — a primary, first-person founder account of turning one named customer into sales-conversation credibility. Cited in From One Customer to a Flywheel.
• Alex Turnbull / Groove's own blog (groovehq.com/blog/customer-development), documenting roughly 500 customer conversations run as a structured discovery exercise — a company's own published account, not a third-party retelling. Cited in From One Customer to a Flywheel.

#3. Tier 2 — credible secondary and trade sources

• Rob Fitzpatrick, The Mom Test (self-published, August 2013) — a single-author practitioner framework, but Fitzpatrick co-ran a consultancy (Founder Centric) built specifically around this interview discipline, which is real, disclosed operating provenance rather than an armchair theory. The tag it gets is not uniform across the lessons that cite it, and that's worth stating plainly rather than smoothing over. Why the First Customer Is Different and The Actual Conversation flag the framework itself [Speculative], on the reasoning that a single self-published, non-peer-reviewed book is the right tier for the source regardless of how sound any one claim inside it turns out to be — even where the mechanism it names (people are too polite to be honest to your face) maps onto the independently studied social-desirability-bias literature in §2. The Mom Test (the whole lesson), From One Customer to a Flywheel, and Diagnosing a "No", by contrast, tag Fitzpatrick's specific operative claims — the ask-about-past-behavior interview rules and the compliments/fluff/ideas versus facts/commitment/advancement distinction — [Established], on the reasoning that those particular distinctions converge identically across the book's own text and every independent summary checked for this course, which is a claim about the specific claim rather than the source's tier. Both readings are individually defensible, but this course did not apply them consistently, and a reader should not expect one tag to hold across every lesson that cites this source. Cited in Why the First Customer Is Different, The Mom Test (the whole lesson), The Actual Conversation, From One Customer to a Flywheel, and Diagnosing a "No".
• Gabriel Weinberg & Justin Mares, Traction (2015) — reconstructed from multiple convergent third-party summaries (Dan Silvestre, ReadingGraphics), not verified against the book's primary text, the same reason Fitzpatrick's Mom Test above carries a [Speculative] tag rather than [Established]. The book's own centerpiece, the Bullseye Framework, is already cited on this platform's Marketing course and isn't re-taught here. What this course draws from it instead: the 50/50 time-allocation rule (split time between building and finding traction from day one, not sequentially) and the 19 named traction channels as brainstorming input for a founder still at zero customers. Both tagged [Speculative] for the same not-verified-against-primary-text reason. Cited in Finding Your First Buyer.
• April Dunford, Obviously Awesome — a named practitioner's positioning methodology, verified via her own site and book summaries rather than a peer-reviewed instrument, offering the clearest real, checkable equivalent to this course's "audience mismatch" diagnostic. Cited in The Actual Conversation and Diagnosing a "No".
• Steli Efti, co-founder and CEO of Close, and his own most-documented story — 48 emails sent to a single investor before winning the investment — plus his repeatedly published rule, "cold emails make sense when you have something specific to offer and something specific to ask for." Narrative and heuristic value, not statistical; Efti has not published raw response-rate data with disclosed methodology. Cited in Founder-Led Outbound, The Actual Conversation, and From One Customer to a Flywheel.
• Lenny Rachitsky's newsletter (lennysnewsletter.com), sourced from direct, named, attributed founder interviews — Figma's Dylan Field and Sho Kuwamoto, the B2B security company Zip, Vanta, and Gong's founder-connection-sourced first dozen customers. Journalism built on real, checkable companies and quotes, stronger than an anonymous forum post, but not a peer-reviewed study. Cited in Finding Your First Buyer, Founder-Led Outbound, and From One Customer to a Flywheel.
• Indie Hackers' own community threads on finding a first customer — a real, long-running, Stripe-acquired platform, with real, checkable named or pseudonymous handles behind each account. Explicitly flagged as self-reported and subject to survivorship bias: founders who tried the same tactic and failed quietly don't write "how I got my first customer" posts. Cited in Finding Your First Buyer as illustrative of a category of tactic, not as proof any specific number generalizes.
• Multiple independent 2026 subreddit-marketing guides, converging consistently on the "9:1" or "90/10" non-promotional-to-promotional posting ratio as Reddit's de facto community norm — a pattern repeated across sources rather than traceable to one named, written policy. Cited in Finding Your First Buyer.
• The Sean Ellis "40% test," popularized through Rahul Vohra's own account of building Superhuman (reported via a TechCrunch interview and corroborated across multiple retrospectives) — a widely used practitioner heuristic, not a controlled study, for when a pattern of responses starts counting as signal. Cited in Pricing Your First Sale, From One Customer to a Flywheel, and Diagnosing a "No".
• Owen Van Syckle's own essay naming "pre-emptive discounting" as the specific, common failure mode of a first-time founder lowering price before any real objection has been raised — a named practitioner's account from direct coaching experience, not a controlled study. Cited in Pricing Your First Sale and The Actual Conversation.
• Xaver Lehmann, writing as "The Honest Founder" on Substack, disclosing a first-customer underpricing mistake he estimates cost roughly €54,000 in a single year — a single, self-published founder account with no independent verification of the exact figures, used as an illustrative disclosed anecdote. Cited in Pricing Your First Sale.
• Patrick McKenzie (patio11), writing on his Kalzumeus blog — a primary, dated account of a real price change (Bingo Card Creator, $24.95 to $29.95, no reported drop in conversion). The commonly attributed "double the highest number you're considering" phrasing could not be traced to a verbatim quote and is flagged in the lesson itself as [Speculative] even while his general underpricing position is well documented. Cited in Pricing Your First Sale.
• Jason Lemkin, writing on SaaStr — a named investor's own stated, segment-dependent position (overcharge a large buyer, undercharge a small one), drawn from direct portfolio experience rather than a controlled study. Cited in Pricing Your First Sale.
• Bonsai, Wise, and PandaDoc, three independent template-agreement vendors that converge on the same basic structure for a freelance or solo-service agreement — a consistent pattern across multiple vendors, not a single primary-sourced legal standard, so the structure is treated as reliable and the specific wording as adaptable. Cited in The Actual Conversation.

#4. Tier 3 — treat with active skepticism

A version of the same problem this platform's Marketing course names for "CAC benchmark 2026" content exists here for "how founders get their first customer": a search surfaces round, confident, unsourced numbers on dozens of near-identical sites, and this course treats noticing that pattern as part of its own honesty obligation rather than an excuse to repeat the numbers anyway.

• "63% of B2B revenue comes from referrals" and its close variants — appear across dozens of near-identical sites with no named study and no disclosed sample behind any of them. Named directly in How to Use This Course as the example of a statistic this course defaults to [Speculative] and refuses to launder into fact, with the real, sourced Schmitt/Skiera/Van den Bulte figure (§2 above) offered in its place.
• "Founder-led sales converts 2–3x higher than delegated outreach" — recurs across sales-tooling blogs and practitioner writing with a different multiplier every time, no disclosed methodology behind any version of it. Named directly in How to Use This Course as a claim this course won't state as fact, even though the underlying direction — a founder's own outreach reads as more credible than a templated one — is real and supported by the Hovland/Weiss research in §2.
• Cold-email response-rate benchmark data from vendors selling cold-email or sales-engagement software (Instantly, Apollo, Woodpecker, Belkins, and similar) — real, large-sample datasets, but methodology is rarely disclosed, the publishers have a direct incentive to make personalization (and their own tool) look effective, and different vendor reports disagree with each other by 2–3x on what should be the same metric. Used in Founder-Led Outbound only as a hedged, directional claim — personalization plausibly multiplies reply rate rather than nudging it — never as a specific, quotable percentage.
• "Feel, felt, found" as an objection-handling structure — old enough that its origin is genuinely disputed (mid-century sales training, sometimes attributed to Zig Ziglar or Xerox, with no verifiable original source found). No disclosed experimental backing for its effectiveness exists. Taught in The Actual Conversation explicitly as a scripting convenience that doesn't require improvising live, not as an evidence-based technique.

#5. What this course does not cover, and why

This course deliberately stops short of two adjacent bodies of material this platform already owns elsewhere, and the boundary matters enough to state plainly rather than let a reader discover it mid-lesson. It does not build out the Sales course's full objection-handling spine — SPIN's situation/problem/implication/need-payoff sequencing, the Challenger sale's insight-led reframe, MEDDPICC's qualification discipline, each mechanism tagged by replication status. The Actual Conversation's four objections (price, "I already have something," timing, trust) are a strict, narrow subset built for a single owner-operator who can say yes on the spot with no procurement process at all — the moment a prospect is negotiating in multiple rounds, or a buying committee is involved, this course's own material says so and hands off explicitly.

It also does not build the Marketing course's CAC/LTV and paid-channel-auction machinery, for the reason argued at length in Why the First Customer Is Different: an auction requires a stable, measurable relationship between spend and outcome, and that relationship doesn't exist yet for a founder with zero proven sales. Diagnosing a "No" names the specific, checkable handoff point directly — once a channel has been run more than once on purpose, a price has held across more than one real deal, and a second win has landed inside the correctly diagnosed audience, the evidence has crossed from Blank's Discovery into Validation, and that is exactly where this platform's Marketing course's tools stop being premature.

Beyond those two explicit handoffs, this course does not cover raising capital, formal business-entity structure, or tax treatment — The Actual Conversation's note on invoicing says plainly that it isn't legal or tax advice, and a reader should check their own local rules before treating anything here as a substitute for either. And it deliberately stays general-purpose rather than building a channel-specific playbook for any one business type: Marlowe's bookkeeping practice carries the numbers through every lesson, but each lesson also names one different real business type — a SaaS founder, an e-commerce brand, an agency — briefly enough to keep the frameworks honest about applying past bookkeeping specifically, without turning this course into nine separate vertical playbooks it was never built to be.