LLenny's Podcast
← All episodes
Jake Knapp & John Zeratsky (Character Capital)13 July 2025

Rapidly test and validate any startup idea with the 2-day Foundation Sprint (from the creators of the Design Sprint)

6Frameworks
12Insights

Episode overview

Jake Knapp and John Zeratsky explain the two-day Foundation Sprint, a structured process for aligning a core team on its customer, problem, competition, differentiation, and implementation approach. The sprint produces an explicit founding hypothesis, which teams then test through consecutive Design Sprints using prototypes, customer interviews, and a scorecard. Examples from Latchet, Melo, and Axion Orbital illustrate how teams can revise weak assumptions quickly, while the guests argue that AI-assisted building makes deliberate differentiation and upfront thinking even more important.

Key ideas

  • A Foundation Sprint brings the core leadership team together for roughly ten focused hours, usually split across two days.
  • The process has three phases: establish the basics, identify meaningful differentiation, and choose an implementation approach.
  • Teams work “alone together” using silent note-and-vote exercises and a designated decider to gather independent thinking without prolonged group debate.
  • A useful differentiation promise must matter to customers and be something the team can actually deliver, ideally using its distinctive insights, motivations, or capabilities.
  • The output is a one-sentence founding hypothesis covering the customer, problem, approach, competitors, and differentiators, plus a backup approach.
  • Consecutive Design Sprints test the hypothesis with realistic prototypes, real target customers, and a scorecard that exposes which assumptions are supported.
  • Fast AI-assisted prototyping can create generic products; teams should use AI to accelerate execution without outsourcing customer understanding, positioning, or product judgment.

Transcript available · source text is retained privately and is not published

Frameworks in this episode

People & resources mentioned

Attributed to the moment in the episode. Timestamps are approximate.

People · 17

  • Jake KnappMentionsHe developed the Design Sprint after prototyping what became Google Meet, later refined it at Google Ventures, and co-founded Character Capital.

    Today, my guests are Jake Knap and John Zeratsky.

  • John ZeratskyMentionsAuthor of Sprint and Make Time; formerly a leader on Google Ads and YouTube; co-founder of Character.

    Today, my guests are Jake Knap and John Zeratsky.

  • Brandon FooMentionsI'm very excited to chat with Brandon Fu, founder and CEO of Paragon.

    I'm very excited to chat with Brandon Fu, founder and CEO of Paragon.

  • Lenny RachitskyMentionstoday we've got another very special compilation episode something I've been pulling on more and more with the podcast and the newsletter

    Hey Lenny, thanks for having us.

  • Larry PageMentionswhen I was at Google the thing I took away most from any review and Lariat survey

    We couldn't convince Larry and Sergey and Eric that they should fund this thing.

  • Eric SchmidtMentionsmentioned

    We couldn't convince Larry and Sergey and Eric that they should fund this thing.

  • Sergey BrinMentionswhen I was at Google the thing I took away most from any review and Lariat survey

    We couldn't convince Larry and Sergey and Eric that they should fund this thing.

  • Eli GoldmanMentionsCo-founder of Character Capital with Jake Knapp and John Zeratsky.

    And then after a few years at Google Ventures, John and I left and eventually together with our co-founder Eli Goldman

  • JamesMentionsLatchet co-founder and former Substack engineer who helped lead its growth team.

    Latchet, a couple co-founders, Chris and James, who were they had left Substack

  • ChrisMentionsLatchet co-founder and former Substack engineer who helped lead its growth team.

    Latchet, a couple co-founders, Chris and James, who were they had left Substack

  • David PlacekMentionsDavid is the founder of Lexicon branding which pioneered the field of brand naming

    David Plac who is this naming expert that we just had on the podcast.

  • April DunfordMentionsApril Dunford was on this podcast and she has this book that she put out recently called the sales pitch

    also April Dunford who's like the I don't know the the the god of of of positioning.

  • Rahul VohraMentionsRahul from superhuman is kind of well known for his approach to finding product Market fit

    the founder of superhuman Rahul he actually spent there's a lot of cool context on how he differentiated superhuman

  • Bob BaxleyMentionsBob is a designer, executive, and adviser who's built and led design teams at Apple, Pinterest, Yahoo, and most recently THSpot.

    I had this guest recently, Bob Baxley.

  • Maor ShlomoMentionsMayor started a company called Base 44, which is essentially a more advanced vibe coding tool.

    I recently had a founder on the podcast uh uh Mayor who built this company called base 44.

  • Maruthi JohnMentionsFounder in Character Labs who previously worked in sales at Rippling.

    I'm thinking of Maruthie John who's uh comes from doing sales at Rippling

  • DenjMentionsSolo founder of Axion Orbital who ran the sprint process and built its prototypes.

    Axion Orbital it's called and the founder Denj is actually running this whole process solo

Resources · 46

  • ClickCoinedbook · Jake Knapp and John Zeratsky

    They published a book with this framework at the beginning of this year called Click

  • Lenny's NewsletterCoinednewsletter · Lenny Rachitsky

    if you become an annual subscriber of my newsletter, you get a year free of a bunch of amazing products

  • Foundation SprintRecommendsother · Jake Knapp and John Zeratsky

    And I highly encourage you to do this if you're in the process of baking an idea.

  • GranolaRecommendssoftware · Granola

    including Bolt, Linear, Superhum, Notion, Perplexity, Granola, and more.

  • PerplexityRecommendssoftware · Perplexity AI

    including Bolt, Linear, Superhum, Notion, Perplexity, Granola, and more.

  • NotionRecommendssoftware · Notion Labs

    including Bolt, Linear, Superhum, Notion, Perplexity, Granola, and more.

  • LinearRecommendssoftware · Linear

    including Bolt, Linear, Superhum, Notion, Perplexity, Granola, and more.

  • BoltRecommendssoftware · StackBlitz

    you get a year free of a bunch of amazing products, including Bolt, Linear, Superhum, Notion, Perplexity, Granola, and more.

  • BrexMentionscompany

    This episode is brought to you by Brex, the financial stack used by one in every three US venture-backed startups.

  • ParagonCoinedsoftware · Brandon Foo

    I'm very excited to chat with Brandon Fu, founder and CEO of Paragon.

  • MicrosoftMentionscompany · Bill Gates and Paul Allen

    I had been working at the beginning of my career at Microsoft for a few years.

  • GmailMentionssoftware · Google

    Went to Google. This is like 2007, 2008, 2009 and was working on the Gmail team.

  • GoogleMentionscompany · Google

    Went to Google. This is like 2007, 2008, 2009 and was working on the Gmail team.

  • StockholmMentionsplace

    They were in Stockholm. I went there for a week.

  • Google MeetCoinedsoftware · Google

    And that prototype was what became Google Meet.

  • Design SprintCoinedother · Jake Knapp and John Zeratsky

    the notion that maybe you could repeat that and recreate that is what led me to create the design sprint.

  • Google VenturesMentionscompany · Google

    and then I went to go work at Google ventures.

  • Character CapitalCoinedcompany · Jake Knapp, John Zeratsky, and Eli Goldman

    together with our co-founder Eli Goldman, we started our own venture firm, Character Capital.

  • Foundation SprintCoinedother · Jake Knapp and John Zeratsky

    that led us to say, hey, we need to create a new sprint method

  • LatchetMentionscompany · Chris and James

    We're gonna talk about this company called Latchet first.

  • Character LabsCoinedcourse · Character Capital

    this is a kind of an accelerator-l like program that we run at Character Capital

  • SubstackMentionswebsite · Substack Inc.

    they were engineers leading up the growth team at Substack

  • EtsyMentionscompany · Etsy

    You can put your products on Etsy, but Etsy has become quite commoditized.

  • ShopifyMentionscompany · Shopify

    you can build a site on Shopify, but you still have to market it if you do that.

  • AnthropicMentionscompany

    we're uh uh invited in to speak at openai and anthropic about this

  • OpenAIMentionscompany

    we're uh uh invited in to speak at openai and anthropic about this

  • MeloMentionssoftware

    So, this is a startup called Melo.

  • BindwellMentionscompany

    we have a company that we invested in called bindwell and uh they use AI to design precision pesticides

  • CursorMentionssoftware · Anysphere

    like cursor you basically is competing on uh you know you're saving engineering time

  • SuperhumanMentionssoftware · Superhuman

    the founder of superhuman Rahul he actually spent there's a lot of cool context on how he differentiated superhuman

  • CodaUsessoftware · Coda

    I personally use Koda every single day to manage my podcast and also to manage my community.

  • Google MapsMentionssoftware · Google

    It's like pulling up directions on Google Maps.

  • The Lord of the RingsMentionsbook · J. R. R. Tolkien

    you're Gandalf you're Frodo you're starting off you're trying to get to Mordor

  • Base44Coinedsoftware · Maor Shlomo

    I recently had a founder on the podcast uh uh Mayor who built this company called base 44.

  • Y CombinatorMentionscompany · Paul Graham and co-founders

    versus just I'm just going to do what I hear on YC, just build it, launch it, iterate

  • Google PhotosMentionssoftware · Google

    the early days of what became Google Photos or Google Trips

  • Google TripsMentionssoftware · Google

    the early days of what became Google Photos or Google Trips

  • SlackMentionssoftware · Slack Technologies

    here's this new messaging software called Slack.

  • ReclaimMentionssoftware · Reclaim.ai

    So 2019, this company called Reclaim going after calendar management.

  • DropboxMentionscompany · Drew Houston and Arash Ferdowsi

    Dropbox acquired them last year.

  • RipplingMentionscompany · Parker Conrad

    I'm thinking of Maruthie John who's uh comes from doing sales at Rippling

  • Axion OrbitalCoinedcompany · Denj

    Axion Orbital it's called and the founder Denj is actually running this whole process solo

  • FigmaUsessoftware · Figma

    which parts of this can just be a static Figma mockup.

  • Miro Foundation Sprint templateRecommendstool · Character Capital

    we actually have a template, a Miro template that folks can use if they want to run their own foundation sprint.

  • MiroUsessoftware · Miro

    We we do our work in Miro.

  • Lenny's PodcastCoinedpodcast · Lenny Rachitsky

    You can find all past episodes or learn more about the show at lennispodcast.com.

Spot an error or want something removed? Request a correction or removal.

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Myth Buster· 1

Myth Buster40:30

Price Is Rarely Your Best Differentiator (Except With AI)

Founders often reach for price as a differentiator, but the guests argue it's rarely the most important one and isn't durable — bigger competitors with more resources can always undercut you. The exception is AI companies solving problems that were previously impossible with software, where costs collapse; they cite a portfolio company, Bindwell, using AI to design precision pesticides far more cheaply.

  • Price is rarely the most important differentiator and is not a durable advantage
  • It's hard to beat giants like Google or OpenAI on price because they have far more resources and can run loss leaders
  • AI creates real pricing advantages when it solves problems previously unsolvable with software
  • Portfolio company Bindwell uses AI to design precision pesticides without a massive team of R&D chemists
  • Rule of thumb heard: AI needs to be roughly 10x cheaper than the legacy manual approach to win

I think price is rarely the the most important differentiator.

John Zeratsky · 40:30

it's very difficult to to compete on price in general and we think we find that it is not as it's not as durable of…

John Zeratsky · 42:00
#pricing#differentiation#ai#startups

Hot Take· 2

Hot Take27:00

Big Group Brainstorms Are Dead

The conversation turns to a recurring theme: large-group brainstorms don't work. Instead of everyone shouting ideas, people think deeply and quietly on their own first, then compare. Even IDEO's famous post-it brainstorming has been abandoned by many because it doesn't actually produce the best ideas.

  • Working alone in silence and then comparing beats a big open brainstorm
  • A naming expert who named Pentium, PowerBook and Sonos abandoned brainstorms for small teams thinking quietly
  • Even IDEO's post-it brainstorm approach is widely being dropped because it wasn't working
  • The value comes from thinking deeply yourself, then seeing what everyone else independently came up with

not doing brainstorms in large groups with a lot of people like it feels like that's just dead in every way

Lenny · 27:00

And I feel like IDO just like created this whole we need to brainstorm post-its and then everyone's like that's not actually working. Let's stop…

27:30
#brainstorming#decision-making#teams#ideation
Hot Take35:00

Why Most 2x2 Diagrams Are Consultant Baloney

The guests admit they used to dismiss 2x2 differentiation diagrams as BS. The reason most of them are worthless: one person makes them quickly for a slide deck, they describe technology or market opportunity rather than the customer's perspective, and there's no evidence behind whether the claims are actually true.

  • Most 2x2 diagrams are made by one person quickly for a slide deck
  • They usually describe technology or market opportunity, not the customer's perspective
  • They aren't proven — there's no evidence the differentiators are real or that the team even agrees on them
  • A 2x2 only becomes valuable when built around the customer, agreed by the whole team, and validated through testing

these 2 by two diagrams feel like this is just consultant baloney

35:00

They don't talk about the customer's perspective. And the other thing is that they're they're not proven.

35:30
#differentiation#positioning#strategy#customer

Explainer· 2

Explainer1:01:00

The Faster You Build With AI, the More Generic It Gets

A pattern the team keeps seeing: the more AI-generated or AI-assisted a product is, the more generic it turns out — which makes sense because LLMs are pre-trained on the same data. Rushing to ship means you launch something less differentiated than what already exists, and you only get data from the few who found it, never from those who bounced off it as 'another generic AI thing.'

  • More AI-generated or assisted products tend to turn out more generic
  • This follows from LLMs being pre-trained on the same underlying data
  • Rushing produces something less differentiated than what already exists in the world
  • You get data from people who used it, but none from those who never engaged because it looked generic
  • Counterintuitively, going fast can slow you down in the long run

the more AI generated or assisted they are the more generic they tend to turn out

1:01:00

it's actually kind of counterintuitive that going fast can actually slow you down in the long run.

1:02:30
#ai#differentiation#product#speed
Explainer1:12:30

Differentiation Is the One Thing Every Winner Shared

Looking back across the hundreds of teams they've worked with, the common element in the biggest successes was always a clear, customer-facing differentiation — a promise the whole team drove toward. They trace it through Gmail (massive storage, great search), Slack (fun and teamwork), ChatGPT (zero clicks, just the answer), and portfolio company Reclaim (AI-prioritized calendars).

  • Across the greatest-hits successes, the common element was always differentiation
  • Gmail's promise was massive storage and great search, repeated relentlessly
  • Slack framed a new worldview around fun and boosting teamwork to justify a painful switch
  • ChatGPT's differentiation: a new way of looking at the world with zero clicks
  • Reclaim used AI to auto-prioritize calendars and was acquired by Dropbox

we realized there is a common element and that common element in all of them is differentiation

Jake Knapp · 1:12:30

it's a new way of looking at the world. There's zero clicks, just tells you the answer.

Jake Knapp · 1:15:00
#differentiation#case-studies#positioning#product-market-fit

Story· 2

Story06:00

The Stockholm Week That Became Google Meet

Jake Knapp recounts how a struggling three-year side project at Google finally broke through when the team cleared their calendars for a week in Stockholm and built a prototype instead of chasing the perfect pitch or PRD. That prototype became Google Meet, and the experience of total focus on a real prototype is what led him to invent the design sprint.

  • A three-year Gmail-era side project was going nowhere and the office was about to be shut down after the financial crisis
  • They stopped trying to make the perfect pitch, PRD, or design and instead cleared the calendar for a week to build a prototype
  • The prototype became Google Meet, a browser-based video conferencing tool
  • The power of clearing the calendar and focusing every person on a real prototype is what seeded the design sprint method

And that prototype was what became Google Meet.

Jake Knapp · 07:00

it was so different to clear the calendar for a week and have the entire focal point of every person on the team be how…

Jake Knapp · 07:30
#design-sprint#prototyping#google#origin-story
Story1:22:00

From a Sea of Red to All Green in Three Sprints

The guests walk through Latchet's three back-to-back design sprints. The first scorecard was almost all red — right customer and problem, but the approach and differentiation weren't landing. By the second, red started flipping to yellow, and by the third, every single cell was green. It's an extreme but real example of how fast learning compounds week over week.

  • Latchet's first scorecard was mostly red: right customer and problem, but approach and differentiation failed
  • Second sprint showed 'sun breaks' — red flipping to yellow as differentiation started to dial in
  • Third sprint's scorecard was entirely green, an extreme but real outcome
  • In one week they built three prototypes and tested them with four real customers

if you look across Latchett's first scorecard, there's a ton of red. It's it's a lot of things that didn't work.

Jake Knapp · 1:22:00

Everything is green. Everything's green. If you can't see, every single one is green.

1:26:30
#design-sprint#validation#scorecard#case-study

Q&A· 1

Q&A1:07:00

When You Should Skip the Sprint (The Base44 Question)

Lenny asks about a founder who built Base44, sold it for $80M in six months, and never did any of this process. The guests point to selection bias: we only hear about the brilliant founders who succeeded, and the process is about improving your odds. If you already have product-market fit and conviction, skip it and run — but most people aren't there yet.

  • Base44 was built and sold for $80 million in six months with no formal process
  • There's a selection-bias trap: we hear from the brilliant founders who happened to succeed
  • The sprint doesn't guarantee success — it increases your odds of a good opportunity that clicks with customers
  • If you already have product-market fit and conviction, stop experimenting and run with it

this just increases your odds this is a chance to to get more clear on is this a good opportunity

1:09:00

If you've got product market fit, if you've got evidence that this thing is clicking and you're you have conviction, run with it.

1:09:30
#product-market-fit#founders#validation#selection-bias

Tool· 1

Tool1:37:00

Free Foundation Sprint Miro Template

The guests share the exact Miro template they use with founders to run a foundation sprint, and make it freely available. They do their sprint work in Miro even when in person (or on paper, then captured in Miro), because it's a great shared canvas. It can be found at character.vc.

  • The same Miro template they use with founders is available for free
  • They run sprints in Miro even in person, or work on paper and capture it in Miro
  • The template is a plug-and-play way to run your own foundation sprint at home
  • Available at character.vc

we actually have a template, a Miro template that folks can use if they want to run their own foundation sprint.

John Zeratsky · 1:37:00

You should find it at character.vc

1:38:00
#tools#miro#foundation-sprint#template

Takeaway· 3

Takeaway1:02:30

The Primal Mark: Wait As Long As Possible to Sketch

Lenny relays an insight from guest Bob Baxley he calls the 'primal mark': wait as long as possible before sketching or prototyping, because the moment you draw what you're building, everything after becomes a response to that first idea. The more time spent thinking before concepting the solution, the more likely you are to land on something that works.

  • Delay the first sketch or prototype as long as possible
  • Once you draw the solution, everything you do afterward is a reaction to that first mark
  • More thinking before concepting the solution increases the odds of landing on the right thing
  • This is the reasoning behind doing a foundation sprint before jumping into design/prototyping

wait as long as possible to start any sort of sketch or prototype because as soon as you start drawing what you're building it now…

Lenny · 1:02:30
#prototyping#design#product#ideation
Takeaway1:24:00

One Conversation With a Hypothesis Beats 50 Without

Customer conversations become dramatically more useful once you have a clear hypothesis and prototypes to react to. A Character Labs founder from Rippling had done 50 customer conversations over a month, yet learned more from the very first conversation once he had a hypothesis and prototypes to show — 'night and day.'

  • Customer conversations are more fruitful and pointed when anchored to a specific hypothesis and prototype
  • A founder had 50 conversations in a month but learned more from the first one once he had a hypothesis
  • Prototypes give customers something concrete to react to instead of an abstract concept
  • The difference was described as 'night and day'

had 50 conversations over the past month but I learned so much more from even the first conversation when I had a hypothesis and I…

John Zeratsky · 1:24:00
#customer-research#validation#prototyping#hypothesis
Takeaway1:33:00

Outsource the Prototyping, Not the Thinking

AI is best used in sprints to make prototypes look realistic faster — like having a prototyping team on standby. But the guests warn against outsourcing the thinking: you still have to decide the copy, how you describe the product, and how it's differentiated. Teams that jumped straight to vibe-coding produced believable-but-generic prototypes and had to step back.

  • AI's biggest use in sprints so far is making prototypes look realistic faster
  • It's like having an entire prototyping team on standby
  • You must not outsource the thinking — copy, product description, and differentiation are yours to define
  • Teams that vibe-coded prototypes immediately got believable but generic output and had to step back
  • Detailed sketches act as prompt engineering for an opinionated, well-defined prototype

while you're outsourcing that prototyping work, you don't outsource the thinking

John Zeratsky · 1:33:00

while you're outsourcing prototype prototyping don't outsource the thinking

1:35:00
#ai#prototyping#vibe-coding#product