Notes — Jake Knapp and John Zeratsky on the Foundation Sprint, Differentiation, and Finding What Clicks

Four questions [Adler frame]

Q1 — What is it about? A structured 10-hour, 2-day methodology (the Foundation Sprint) for startup and product teams to produce a Founding Hypothesis before writing any code. Three sequential phases: Basics (who is the customer, what is the problem, who are the competitors, what are the workarounds, what are the team’s advantages), Differentiation (a 2x2 “Loserville” diagram identifying two defensible differentiators), and Approach (magic lenses comparing implementation paths). The output is a single Mad Libs sentence encoding the whole hypothesis. The Foundation Sprint precedes the Design Sprint, which tests the hypothesis via weekly prototype-and-customer-interview cycles.

Q2 — How is it argued? Practitioner inference from hundreds of sprints — first at Google Ventures, now at Character Capital’s Character Labs programme. Jake Knapp traces the methodology’s origin to a Stockholm week in 2009 when he cleared his calendar to build the prototype that became Google Meet. JZ argues the Foundation Sprint was created because Design Sprints run with inception-stage startups kept surfacing the same unfilled gap: no alignment on basics. Case studies (Latchit/Lyric, Mellow, Axion Orbital) show scorecards evolving from all-red → mostly-yellow → all-green over three consecutive weekly sprints. The AI-generic-prototype observation is new and compelling: models trained on existing products generate undifferentiated output, so AI-accelerated building makes the Foundation Sprint more valuable, not less.

Q3 — Is it true? High plausibility. The “three co-founders, three answers” observation matches widespread experience. The Loserville 2x2 argument — that most competitive analysis decks fail because they’re investor-audience, not customer-audience, and unproven — is well-reasoned. The “all-green scorecard” case studies are real but self-selected; Character Labs is a controlled environment with intensive support. The 3–4x acceleration claim (“three to four months of work in three to four weeks”) is founder-reported, not measured. The AI-generic-prototype trap [?] is plausible but not yet systematically quantified.

Q4 — What of it? The Foundation Sprint is actionable today. Miro template at character.vc. The core question it answers — “Does your whole team agree, in a single explicit sentence, on what hypothesis you are testing?” — is worth asking of any product or startup effort regardless of whether you run the full process. The Loserville 2x2 + project principles (“mini manifesto”) is the most portable artefact for PM teams.

Glossary

Foundation Sprint: Jake and JZ’s 10-hour, 2-day structured session producing a Founding Hypothesis and a mini manifesto (Basics page + Loserville 2x2 + project principles); precedes Design Sprints
Founding Hypothesis: Mad Libs sentence — “If we solve [problem] for [customer] with [approach], we think they’ll choose it over [competitors] because of [differentiator 1] and [differentiator 2]”
Design Sprint: Jake Knapp’s original 5-day sprint from Google (map–sketch–decide–prototype–test); now run weekly after the Foundation Sprint to iteratively test the Founding Hypothesis
Basics (phase 1): Who is the customer? What is the problem? Who are the competitors (including workarounds and alternatives)? What are the team’s advantages?
Differentiation (phase 2): Work-alone-together activity producing classic differentiator scores → custom differentiator scores → Loserville 2x2 → project principles (mini manifesto)
Loserville: The three non-winning quadrants on the 2x2 chart; all competitors should land here
Magic Lenses (phase 3 activity): Plotting implementation paths against multiple lenses — customer experience, pragmatic (cost/speed), growth (reach), money (LTV/TAM), differentiation, and founder conviction — to reveal which approach wins consistently or which lens matters most when there is no clear winner
Work alone together: Silent individual note-writing followed by voting and a designated-decider decision; avoids groupthink, speeds through decisions, surfaces divergence before alignment
Click: The moment a product concept resonates with a customer during a prototype interview; also the title of Jake and JZ’s book; evidence signal for product-market fit

§Basics phase

[§ Basics phase] The Basics phase is designed to be fast (2–3 hours) and is primarily a context-loading exercise before the main event of Differentiation. Four questions with work-alone-together answers and decider decisions:

Who is the most important customer? (Single, specific, not multiple segments)
What is the problem? (Often less clear than teams assume; problem discovery can itself be surprising)
Who is the competition? (Extends beyond direct competitors to workarounds and alternatives — “if this is a real problem, the customer is already solving it somehow”)
What are our advantages? (Unique insights, motivations, special capabilities; the raw material for custom differentiation later)

The value is not the answers themselves but the act of making them explicit and jointly visible. Three co-founders commonly have three different answers; the Basics phase surfaces and resolves that divergence before it infects product decisions.

§Differentiation phase

[§ Differentiation phase] The heart of the Foundation Sprint and its most distinctive contribution. Two layers:

Classic differentiators — a pre-built starting list to warm up the team and cover universal dimensions: fast–slow, smart–not-smart, easy–hard, free–expensive, focused–one-size-fits-all, simple–complicated, integrated–siloed. Teams score their product against competitors on each dimension. The exercise normalises the feeling of not winning on everything.

Custom differentiators — the team writes their own continuums (e.g., “networked vs. siloed”, “painless business growth vs. labour-intensive business growth”, “purposeful vs. generic”). Custom differentiators encode the team’s specific insight about what their product can do that others cannot. After writing and voting on ~20–25, the team selects the 2 that are (a) deliverable and (b) compelling to customers. [?] Both conditions must be true.

Loserville 2x2 — the selected two differentiators become the axes. All competitors fall into the three losing quadrants. The startup should occupy the top-right. This becomes the guiding decision-making artefact throughout development.

Project principles / mini manifesto — 2–3 principles derived from the differentiators to guide day-to-day product decisions (e.g., “help sellers help each other”; “do the thing that makes sellers more money”). Analogous to Google’s “fast is better than slow” being a tiebreaker in product meetings.

§Approach phase

[§ Approach phase] The second day’s main activity. Most teams either have 3–4 possible implementation paths or think they have only one. Both states benefit from explicit comparison:

Step 1 — List approaches: Name each distinct path (e.g., Latchit considered: a standalone app, a newsletter platform, a Shopify plugin, or a full-stack build).

Step 2 — Magic Lenses: Plot approaches on 5–6 2x2 charts, each representing a different stakeholder/priority lens:

Customer lens (easy to use × perfect solution)
Pragmatic lens (cheap/fast × high quality)
Growth lens (easy to adopt × wide reach)
Money lens (large audience × high LTV)
Differentiation lens (aligned with the Loserville differentiators)
Conviction lens (founder excitement — e.g., “F*** yeah” vs. “meh”)

If one option wins consistently across lenses, the decision is clear. If no option wins, the magic lenses reveal which lens is the most important tiebreaker for this team.

Step 3 — Commit + backup: One first-choice approach and one backup plan. Having an explicit backup reduces the psychological cost of being wrong and allows fast pivoting.

§Founding Hypothesis and scorecard

[§ Founding Hypothesis and scorecard] After all three phases, the Founding Hypothesis collapses into a single sentence:

“If we solve [problem] for [customer] with [approach], we think they’ll choose it over [competitors] because of [differentiator 1] and [differentiator 2].”

This sentence encodes the team’s full strategy. Every Design Sprint thereafter starts by reviewing the Founding Hypothesis and identifying: which assumption within it is the biggest risk right now?

The scorecard (a new addition not in the original Design Sprint book) maps each customer interview back to the Founding Hypothesis components: Was this the right customer? Did they have the problem? Did they choose the product over the competition? Was differentiation actually compelling? Red/yellow/green per interview per dimension. A sprint producing all-green is a strong signal to start building.

§AI and the generic prototype trap

[§ AI and the generic prototype trap] Key observation with implications for product development in 2024–25: AI-assisted prototyping generates convincing-looking outputs quickly, but because LLMs are trained on existing products, AI-generated prototypes tend to be generic. Teams that skip the Foundation Sprint and go directly to vibe-coding end up testing generic products against customers — and either fail to get useful signal, or ship something undifferentiated.

The recommended sequence: Foundation Sprint (slow thinking about differentiation) → Design Sprint prototyping (fast building, grounded in explicit differentiation). Sketching the prototype on paper before prompting the LLM is a form of prompt engineering that yields an opinionated, differentiated result. Don’t outsource the thinking while outsourcing the building.

§Character Labs

[§ Character Labs] Jake Knapp and John Zeratsky’s VC programme at Character Capital (character.vc). They invest in a cohort of pre-seed, sometimes inception-stage companies simultaneously and run them through one Foundation Sprint + three consecutive weekly Design Sprints (3.5–4 weeks total). Founders report 3–4x acceleration (3–4 months of early-stage work compressed into the programme). The Miro template used in Character Labs is publicly available at character.vc.