The Foundation Sprint
A 10-hour, 3-phase team sprint that turns a vague startup idea into one testable founding hypothesis
- Difficulty
- Moderate
- Time to result
- ~days to results
- Steps
- 5
- Confidence
- 95%
The Foundation Sprint is a highly-scripted two-day (roughly 10-hour) working session for the core founding/leadership team to make every key early decision together before writing code. It moves through three phases — Basics, Differentiation, and Approach — that combine into a single 'founding hypothesis' you then test with design sprints. It exists to give founders clarity about strategy that teams otherwise spend months failing to nail down.
Origin
Created by Jake Knapp and John Zeratsky at their venture firm Character Capital (2021-2022), building on the Design Sprint that Knapp originated at Google (the process that produced Google Meet) and refined with John Zeratsky at Google Ventures across a few hundred teams. Published in their 2025 book 'Click.'
Core principles
- 01Clear the calendar and put the whole core team in one room — the co-founders and whoever leads product, engineering, design, marketing, sales
- 02Make the founding hypothesis explicit; every product already has one, it is just usually hidden and different teammates hold different versions
- 03Going fast can slow you down in the long run — a deliberate pause up front prevents building the wrong thing quickly
- 04The output is a hypothesis to test, not a plan to execute — the word 'hypothesis' keeps the team honest
How to run it
- 1
Clear calendars and assemble the core team
Get the co-founders (or the product/eng/design/marketing/sales leads inside a larger org) together for roughly 10 hours, spread over two days as two 4-6 hour blocks. Turn off Slack and email.
Pro tip Time-box to ~10 hours; a fast small team may finish in 8, a larger team with more discussion may take 12.
- 2
Phase 1 — Nail the Basics
Rapidly answer the embarrassingly simple questions: who is the most important customer, what problem are you solving, who/what is the competition and how do people solve it today (including workarounds and alternatives), and what advantages do you have. Move fast — the goal is to load context, not perfect it.
Pro tip Don't linger on the basics; their real job is to 'boot up' the right context in everyone's head for the differentiation work that matters most.
Watch out Three co-founders will often give three different answers to 'who is your customer' — that divergence is exactly what this phase surfaces and resolves.
- 3
Phase 2 — Choose your Differentiation
Score yourself against competitors on classic differentiators, then write and score custom ones, and land on the two strongest to build a 2x2 differentiation chart. This is the heart of the sprint.
Pro tip Be brutally honest about where you actually stack up; it is normal and healthy to not be best on every axis.
- 4
Phase 3 — Pick your Approach
On day two, lay out the different implementation paths that all solve the same problem for the same customer with the same differentiation, evaluate them, commit to a first choice, and name a backup plan.
Pro tip Even if you think there's only one option, force yourself to invent alternatives so you can pivot fast if the first approach fails.
- 5
Assemble the founding hypothesis and go test it
Combine the three phases into a single Mad Libs-style sentence, then run back-to-back one-week design sprints (clear 2-4 weeks) to build prototypes and test the hypothesis with real customers.
Pro tip Give yourself the chance to be wrong — schedule the design sprints so the hypothesis can actually fail early.
In the wild
Two ex-Substack growth engineers, Chris and James, joined Character Labs wanting to build a community-driven sales tool for artisans. As builders their instinct was to write code immediately, but they were pushed to run a Foundation Sprint first — clarifying customer (artisans who find tech and marketing hard), problem (sales growth), competition (Shopify, Etsy, art fairs), and differentiation before building anything.
→ They produced a founding hypothesis and, over three consecutive design-sprint weeks, went from a scorecard full of red to promising yellows to (in a later real example shown) an all-green scorecard — compressing months of learning into weeks.
Common mistakes
Skipping straight to building because AI makes it fast
Knapp and Zeratsky observed that the more AI-generated a fast-built product is, the more generic it turns out, because LLMs are trained on the same existing data. Building momentum is hard to stop, so heading fast in the wrong direction actively hurts you.
Treating the basics as too obvious to bother with
Teams assume they're aligned on customer, problem, and competition, but when everyone answers independently the specifics differ. The failure mode is a team that never agrees on the basics and one that never tests them.
Is it for you?
Best for
Pre-seed / inception-stage founders and product teams at the very beginning of a new product, before writing code or committing to an approach
Not ideal for
A product that already has evidence of product-market fit and traction — if it's clicking with customers, run with it instead of pausing to sprint
From the transcript
“there are three phases to the foundation sprint. First phase is the basics. Second phase is differentiation and the third phase is the approach to…”
“We recommend this kind of crazy idea that you clear your calendar”
“it's going to be 10 hours roughly give or take”
“going fast can actually slow you down in the long run”
From the episode
Rapidly test and validate any startup idea with the 2-day Foundation Sprint (from the creators of the Design Sprint)
Jake Knapp & John Zeratsky (Character Capital)