Symptom: Problem block reads like a feature list.
Cause: You wrote your solution into the customer's problem.
Fix: Describe the customer's current situation and pain. Keep the product idea in Solution, linked to the problem.
Lean Canvas
Ash Maurya's nine-block Lean Canvas, built as a working model instead of a form. Every cell is a claim with an evidence state, so you can see what is supported, what is contradicted, and which unknown deserves your next test. This is the Lean Canvas, not the Business Model Canvas. Free, no login, runs in your browser.
- Nine blocks plus Existing Alternatives, Early Adopters and High-Level Concept
- Evidence states: Unknown, Assumption, Observed, Supported, Contradicted, Decision
- An explainable riskiest-assumption queue and a learning contract per test
- Autosave, share link, Markdown/CSV/JSON export, JSON import, PNG and print
Mode:
Blocks with current hypotheses: 0/12, this counts hypotheses, not quality or readiness. Claims: 0. Evidence items: 0.
Not reviewed yet.
Share privacy: a share link encodes the whole canvas in the URL. Compression is not encryption, anyone with the link can inspect your segment, pricing, evidence and advantage notes. Remove confidential content before sharing, and use an export for oversized canvases.
Riskiest assumption decision queue
Ordered, not scored. Every item shows why it is here.No unresolved claims. Every entry is supported and settled, or has a recorded decision.
Problem
The top 1–3 problems your early adopters have today.
Focus limit: 3
No current hypothesis here yet.
Existing Alternatives
How customers solve this today. Maurya calls this the most important field: customers judge you against what they already do.
No current hypothesis here yet.
Customer Segments
Who you create value for, and who pays.
Focus limit: 3
No current hypothesis here yet.
Early Adopters
The small, findable group that will try you first.
No current hypothesis here yet.
Unique Value Proposition
One clear message: why you are different and worth attention.
No current hypothesis here yet.
High-Level Concept
An X-for-Y analogy that makes the idea easy to repeat.
No current hypothesis here yet.
Solution
The smallest thing per problem, top 3 features, one per problem.
Focus limit: 3
No current hypothesis here yet.
Channels
Free and paid paths to your early adopters.
No current hypothesis here yet.
Revenue Streams
Revenue model, price hypothesis, and who actually pays.
No current hypothesis here yet.
Cost Structure
What it costs to build and run over the next 3–6 months.
No current hypothesis here yet.
Key Metrics
A small set of numbers that tell you the model is working, each tied to a decision.
Focus limit: 5
No current hypothesis here yet.
Unfair Advantage
Something hard to copy or buy. It is normal to leave this unresolved early.
No current hypothesis here yet.
How it works
Step 1
Use Wizard mode to follow Maurya's numbered order for a first pass, or Canvas mode to edit all nine blocks and the three sub-fields. Each cell is a claim, not a finished answer.
Step 2
For every claim, set an evidence state and record what you actually observed, source, direction and strength, instead of clicking “Validated”. The audit shows where the state and the evidence disagree.
Step 3
Open the decision queue to see the unresolved claim with the highest consequence and uncertainty, and why. Create one learning step with a decision criterion, then record the result and update the canvas.
What the Lean Canvas is, and who created it
The Lean Canvas is a one-page business-model tool created by Ash Maurya in 2010 and set out in Running Lean. It adapts Alexander Osterwalder's Business Model Canvas for startups operating under extreme uncertainty. Where the Business Model Canvas describes an operating business, the Lean Canvas concentrates attention on the startup-stage unknowns: the problem, the early adopters, the numbers that would prove the model works, and whatever would stop a copycat.
It is deliberately a hypothesis map, not a plan. Ash Maurya's guidance is that a first draft should take well under an hour and change as you learn. A finished, untouched canvas is usually a warning sign, not an achievement.
Attribution: Lean Canvas by Ash Maurya, adapted from the Business Model Canvas by Strategyzer, used under a Creative Commons Attribution-ShareAlike licence. This is an independent CraftUp implementation and is not affiliated with or endorsed by LEANSTACK.
The nine blocks explained
Problem
The top 1–3 problems your early adopters have today.
Customer Segments
Who you create value for, and who pays.
Unique Value Proposition
One clear message: why you are different and worth attention.
Solution
The smallest thing per problem, top 3 features, one per problem.
Channels
Free and paid paths to your early adopters.
Revenue Streams
Revenue model, price hypothesis, and who actually pays.
Cost Structure
What it costs to build and run over the next 3–6 months.
Key Metrics
A small set of numbers that tell you the model is working, each tied to a decision.
Unfair Advantage
Something hard to copy or buy. It is normal to leave this unresolved early.
The three sub-fields
The printed template has always carried three fields inside other blocks. LEANSTACK's current site also counts them as blocks, which is why you sometimes see "twelve blocks". They are the same canvas.
Existing Alternatives
How customers solve this today. Maurya calls this the most important field: customers judge you against what they already do.
Early Adopters
The small, findable group that will try you first.
High-Level Concept
An X-for-Y analogy that makes the idea easy to repeat.
Lean Canvas vs Business Model Canvas
Both are one-page, nine-box canvases. Five blocks survive from the Business Model Canvas; four are replaced. Neither is better in the abstract, they answer different questions at different stages.
| Business Model Canvas | Lean Canvas | Why Maurya swapped it |
|---|---|---|
| Key Partners | Problem | Chasing partners as an unknown startup is usually premature; most startups fail by solving the wrong problem. |
| Key Activities | Solution | Activities describe what you do once you know the solution; the small Solution box is a guard against falling in love with the first idea. |
| Key Resources | Key Metrics | Listing resources rarely registers as risk; measuring the wrong thing does, and leads to premature optimisation. |
| Customer Relationships | Unfair Advantage | Early on the relationship is direct, through interviews; what matters is whether anything stops a copycat later. |
Shared by both: Customer Segments, Value Proposition (Lean Canvas adds "Unique"), Channels, Revenue Streams and Cost Structure.
Use the Lean Canvas when
Your biggest unknown is whether anyone has the problem you think they do. It concentrates attention on Problem, Solution, Customer Segments and the UVP, the places most pre-product companies die.
Use the Business Model Canvas when
The problem is settled and your harder question is operations, partnerships, resources and scale economics. Many teams use the Lean Canvas first and graduate to the Business Model Canvas.
What order should you fill it in?
Maurya's numbered order starts with the customer and the problem and moves toward what you are guessing. He has also written that there is no single right order, where you start says something about what you assume matters.
- Customer Segments and Early Adopters
- Problem and Existing Alternatives
- Unique Value Proposition and High-Level Concept
- Solution
- Channels
- Revenue Streams and Cost Structure
- Key Metrics
- Unfair Advantage (often honestly unresolved at the start)
Assumptions vs evidence: why there is no “Validated” button
A canvas is a wall of guesses. Marking one "validated" hides the thing that matters: what you observed, what you infer, and what you decided. This tool uses a small, honest set of states instead.
- Unknown, Not framed yet, or we genuinely do not know.
- Assumption, A stated belief with no evidence recorded yet.
- Evidence observed, We have raw evidence, but not yet a pattern we trust.
- Supported, Evidence currently supports this claim. Still revisable, not timeless truth.
- Contradicted, Evidence weakens or contradicts this claim. Decide what to do; do not delete it.
- Decision made, The team made a decision. It can still be revisited when evidence changes.
A claim moves to Supported only when the evidence you recorded actually supports it. Contradicting evidence is kept rather than deleted, because a weakened assumption is still a decision input. Evidence can support a claim without turning it into timeless truth.
Which assumption should you test first?
Not the first box on the page, and not whichever claim has the longest text. The decision queue ranks unresolved claims by three things you can inspect: the consequence if the claim is wrong, how much the next decision depends on it, and the uncertainty in the evidence. Open contradictions rise to the top. The tool shows its reasoning on every item instead of a hidden score.
Each learning step is tied to one claim, with a question, a method matched to the risk type, the evidence to collect, a success criterion, and what you will do if the result contradicts you. Record the result, add it as evidence, and update the claim, that loop is the point.
Common anti-patterns
Solutions in the Problem block: you wrote your product idea where the customer's pain belongs. Confirmation bias: contradictions get deleted instead of recorded. Fake certainty: everything is marked supported on one anecdote. Metric theatre: a number with no decision attached. Invented advantage: a brag in the Unfair Advantage box instead of an honest “not established yet”. Dead tests: a learning step with no decision criterion, so nothing changes afterwards.
Pro tips
- Start from the customer and the problem together. A problem with no named customer cannot be researched or prioritised.
- Record the Existing Alternatives field properly. Customers compare you against what they already do, not against an empty market.
- Keep the Problem block about the customer's situation. If it names your app, dashboard or platform, move that idea to Solution.
- Write the UVP as one message for one segment. If it needs three sentences, it is not yet a UVP.
- For every claim, record where the evidence came from. Source type and direction matter more than the amount of text.
- Do not mark something Supported on a single anecdote. One story is useful; it is not a pattern.
- Record contradicting evidence instead of deleting the claim. A contradicted assumption is still a decision input.
- Size the risk explicitly: how much does the next decision depend on this, and what happens if it is wrong?
- Give each learning step a decision criterion before you run it, so the result changes a decision rather than just producing data.
- Treat the canvas as a living model. Save a change note when evidence moves a claim, and re-read the canvas before big bets.
Common mistakes and how to fix them
Symptom: Everything is marked Supported.
Cause: Evidence is being counted by feel rather than recorded.
Fix: Attach evidence items with a source, direction and strength. A claim moves to Supported only when the recorded evidence supports it.
Symptom: The riskiest claim is not the one you are testing.
Cause: Priority is coming from block order instead of consequence and uncertainty.
Fix: Use the decision queue: size the consequence if wrong, note the decision weight, and look at the evidence. Fix the highest-consequence high-uncertainty claim first.
Symptom: A test ran but nothing changed.
Cause: The learning step had no decision criterion or no canvas update.
Fix: Write the success and contradicting criteria first, then record the result, the learning and the claim you changed.
Symptom: Metrics that look good but inform nothing.
Cause: The metric was chosen for comfort, not for a decision.
Fix: For each metric, state what you would do differently depending on what it does. Awareness metrics are legitimate when they inform a channel decision.
Symptom: Unfair Advantage is invented to fill the box.
Cause: Treating an empty box as failure.
Fix: “Not established yet” is a valid state. Record aspirational assets as aspirational and do not present them as defensible.
Symptom: Economics is one-sided or a single number.
Cause: Revenue is recorded without a payer or price basis; costs are recorded without a driver.
Fix: Record the payer, the unit or price basis, and the main cost driver as explicit hypotheses you can test.
Symptom: The canvas is rebuilt from scratch every review.
Cause: No change ledger and no learning loop.
Fix: Save a short change note when evidence moves a claim, and link each learning step back to the claim it updates.
When the Lean Canvas is the wrong tool
Skip it when the problem is already settled and your hardest question is operations, partnerships or scale economics, the Business Model Canvas fits better. For a complete financial model, use a financial-model tool; for a delivery plan, use a roadmap; for deep customer-value fit, use the Value Proposition Canvas. The Lean Canvas is a hypothesis map for early, uncertain businesses.
FAQ
Who created the Lean Canvas, and is this the official one?
The Lean Canvas was created by Ash Maurya in 2010, adapted from Alexander Osterwalder's Business Model Canvas. This is an independent CraftUp tool: it implements the nine-block canvas and its three standard sub-fields with attribution, but it is not affiliated with or endorsed by LEANSTACK.
Is this the Lean Canvas or the Business Model Canvas?
This is the Lean Canvas. It replaces four Business Model Canvas blocks, Key Partners, Key Activities, Key Resources and Customer Relationships, with Problem, Solution, Key Metrics and Unfair Advantage. Use the Lean Canvas when your biggest unknown is whether anyone has the problem you think they do.
Does Lean Canvas have 9 blocks or 12?
The printed template has nine blocks. LEANSTACK's current site also counts Existing Alternatives, Early Adopters and High-Level Concept as blocks, making twelve. They are the same canvas: CraftUp treats them as sub-fields of Problem, Customer Segments and Unique Value Proposition respectively.
What order should I fill in a Lean Canvas?
Maurya's numbered order starts with the customer and the problem, then the UVP, solution, channels, revenue streams, cost structure, key metrics, and unfair advantage last. He has also said there is no single right order. CraftUp's wizard follows the numbered order as a first draft and lets you edit in any order afterwards.
How is this different from a static template or an AI generator?
A blank template records what you already believe; many AI generators fill nine boxes with invented detail. This tool keeps every cell as a claim you can attach evidence to, shows what supports or contradicts it, and tells you which unresolved claim most deserves your next test, deterministically, with no AI and no account.
What do the evidence states mean?
Unknown, Assumption, Evidence observed, Supported, Contradicted and Decision made. Supported means the evidence you recorded currently supports the claim, not that it is proven forever. Contradicted claims are kept, not deleted, because a weakened assumption is still a decision input.
How do you decide the riskiest assumption?
By an explicit rule you can inspect: consequence if wrong, how much the next decision depends on it, and how uncertain the evidence is. Open contradictions rise to the top. The queue shows its reasoning rather than a hidden numeric score, and it never claims mathematical precision.
Does the tool work without a login or an account?
Yes. It runs entirely in your browser, autosaves to local storage, and needs no account. You can export Markdown, CSV, JSON, PNG or PDF, share a URL, and import your data back. Nothing about your canvas is sent to CraftUp.
Can I import and export my canvas?
Yes. JSON is the lossless round-trip format and now has an import path. CSV is for bulk spreadsheet entry and is not lossless, it cannot represent multiple evidence items or entity links perfectly. Markdown is a readable share format, and PNG or PDF is for printing.
When should I not use the Lean Canvas?
When the problem is already settled and your hardest question is operations, partnerships or scale economics, the Business Model Canvas fits better. For a full financial model, a business plan, or a delivery roadmap, use those tools instead, the Lean Canvas is a hypothesis map, not a plan.
Learn more with CraftUp
Turn your canvas into clear next steps
Use CraftUp to connect problem discovery, evidence and execution without rebuilding the work.
Freshness
Last updated: 2026-09-24
- Rebuilt as an evidence-backed business-model assumption system: every cell is a ledger entry with an explicit epistemic state, structured evidence and provenance.
- Added the Business Model Assumption Ledger, the explainable Riskiest Assumption Decision Queue, and the Experiment-to-Canvas learning contract.
- Replaced binary Assumption/Validated with Unknown/Observed/Supported/Contradicted/Decision states, added v1 migration, lossless JSON import, RFC-4180 CSV, and share privacy controls.
- Launched the original Lean Canvas Helper with Wizard and Canvas modes, quality checks and history snapshots.