Start Canvas

A start canvas is a working decision document for a business that is still learning. It is not a miniature business plan and should not make uncertain assumptions look settled.

Use it to answer: What do we believe, what evidence supports it, and what should we test next?

Who has the problem?

Write a specific customer situation, not a broad audience.

• Customer type:
• Situation in which the problem appears:
• Who experiences it:
• Who decides or pays:
• Who else influences the decision:

If you cannot name a concrete situation, the customer definition is probably too broad.

What is happening today?

Describe the customer's current behaviour.

• Trigger:
• Current workaround:
• Cost of the workaround:
• Frequency:
• Consequence if nothing changes:

Prefer a recent real event over a general statement such as "customers struggle with X."

What do they do instead?

Record the real alternative, including doing nothing.

• Current product or service:
• Internal workaround:
• Competitor:
• Manual process:
• Delay / acceptance of the problem:

Your offer is competing with current behaviour, not only with companies that look like competitors.

What would be measurably better?

Describe the change from the customer's point of view.

Examples:

• fewer hours spent;
• faster turnaround;
• fewer errors;
• more revenue;
• lower risk;
• simpler compliance;
• better access or convenience.

Avoid "a great experience" unless you can explain what changes in practice.

What could you provide?

Write the smallest credible offer that could produce the outcome.

• What is delivered?
• For whom?
• In what form?
• Over what period?
• What is deliberately excluded?
• What would make the customer say yes?

Do not design the full product here. Design the smallest exchange that can test the important assumption.

How might the customer find or buy it?

Name one or two plausible routes.

• Where does the customer already look?
• Who can introduce you?
• What buying process exists?
• What must happen before purchase?
• What is the cost or effort of reaching the buyer?

Treat distribution as a hypothesis until real behaviour supports it.

Can the exchange work financially?

Record:

• expected price or price range;
• direct delivery cost;
• acquisition or sales cost;
• payment timing;
• implementation cost;
• recurring versus one-time revenue;
• major economic assumption.

A customer can value an offer while the business still has bad economics.

List the assumptions that could make the plan fail.

For each:

Assumption → evidence we have → evidence missing → cheapest credible test → decision if weakened

Prioritise assumptions with both high uncertainty and high consequence.

Separate:

Known: observed behaviour, records, customer commitments, verified external facts.

Inferred: your interpretation of what the evidence means.

Unknown: information you have not tested.

This prevents the canvas becoming a collection of founder beliefs written as facts.

Write one test that can change what you believe.

Hypothesis: Customer: Action: Evidence to collect: Time / resource limit: Signal that increases confidence: Signal that reduces confidence: Decision: continue / change / stop / test again.

A useful test is small enough to run and strong enough to change a decision.

Suppose a founder believes independent consultants will pay for a service that turns messy client notes into a usable project brief.

A weak canvas says: "Consultants need better documentation."

A stronger canvas says:

• Customer: independent consultants with repeated client discovery calls.
• Problem: after calls, they spend 1–2 hours converting notes into a structured brief.
• Existing alternative: write the brief manually or leave notes unstructured.
• Outcome: produce a client-ready brief in less time without losing important context.
• Offer hypothesis: a fixed-price manual brief service delivered within 24 hours.
• Distribution hypothesis: test with consultants reached through existing professional relationships.
• Economics: record price, delivery time and direct labour before automating.
• Key assumption: the time saved is valuable enough to justify payment.
• Test: offer the manual service to relevant buyers and observe payment, repeat use and objections.
• Decision rule: if buyers use it but reject the price, test scope or economics before building software.

The example is not proof that the business will work. It shows how a vague idea becomes a set of testable claims.

Update the canvas when evidence changes.

Ask:

1. Which assumption did we test?
2. What did customers actually do?
3. Which belief became weaker?
4. Which belief became stronger?
5. What is now the largest unresolved risk?
6. What decision is blocked by that risk?
7. What is the smallest useful next test?

Archive old assumptions rather than silently rewriting them. That preserves the learning history.

The Start Canvas is SOE editorial synthesis informed by lean experimentation and business-planning principles. It is a reasoning and testing tool, not a substitute for customer research, financial modelling, accounting, legal advice or market-specific diligence.