Treat the plan as a set of claims
After creating a business plan, identify its riskiest assumptions, check the requirements for operating, and run a small test that can produce useful evidence. Then update the plan with what actually happened. Funding, hiring, and leasing are possible steps, rather than automatic next steps for every business.
A plan may say customers need your product, will pay your price, and can be reached affordably. It may also assume that you can deliver on time. These claims need different tests. Start with the uncertainty that would most change whether the idea is worth pursuing.
The SBA distinguishes traditional and lean planning formats. The level of documentation should fit its use. An exploratory service pilot does not need the same document as a substantial funding request.
Extract an assumption register
| Claim | What could contradict it? | First test |
|---|---|---|
| Cafés struggle to update images | Owners already update them easily | Ask about the most recent menu change |
| A limited package is useful | Owners need video or full branding instead | Discuss a specific scope |
| $100 is acceptable | No suitable prospect accepts an honest offer | Offer a small paid pilot |
| Delivery takes two hours | Editing and revisions double the time | Track one full delivery |
Distinguish “we have not tested this” from “we tested this and it failed.” Leave the result unknown until observation. A founder’s confidence is not a substitute for an evidence field.
The customer-interview guide develops the first research step. When delivery is feasible, define a plain-language service offer and use a bounded paid pilot to test a specific delivery question.
Run a week with a question
- Monday: pick one segment. Name who has the problem, when it occurs, and how you can reach them appropriately.
- Tuesday: learn about recent behavior. Ask what happened last time, what they tried, how long it took, and what they paid. Avoid leading people toward your preferred answer.
- Wednesday: define the pilot. Set scope, price, schedule, maximum spending, and what the customer receives.
- Thursday: check delivery and obligations. Confirm permissions, operating requirements, insurance needs, data handling, and any relevant professional boundaries.
- Friday: make an honest offer. Do not imply that an unfinished product is ready. If selling is premature, test understanding or feasibility instead.
- At the review: record outcomes. Separate interest, booking, payment, delivery, and repeat use.
This is an example sequence, not a universal seven-day launch promise. Regulated activities and physical operations can require a much longer preparation period.
Make the numbers operational
List one-time startup spending separately from recurring fixed expenses and variable costs per sale. Include payment charges, consumables, delivery, refunds, and labor where relevant. Use the same time period for fixed costs and volume.
A $100 package with $40 variable cost contributes $60. With $1,200 monthly fixed costs, 20 packages cover those fixed costs. A 30-package month produces $600 before excluded items. But if available hours allow only ten packages, the initial setup does not cover its costs.
The next action could be a lower fixed-cost pilot, a different scope, a higher price customers accept, or stopping. Buying more equipment before resolving the capacity and demand questions adds exposure.
Operating result is different from cash available. A customer paying next month does not fund a supplier bill due today. Maintain a separate cash calendar before making commitments.
Choose evidence before collecting it
Define what observation would justify the next small step. For an exploratory conversation, the goal may be understanding a repeated problem. For an offer test, the goal may be acceptance from suitable customers under a clear scope. Do not use the same standard for both.
Strategyzer’s Experiment Library illustrates that tests differ in time, cost, and evidence. Our canvas is a separate, original worksheet for recording one test. It does not reproduce those templates or assign a success probability.
A tiny convenience sample can expose problems but does not represent an entire market. Three friends saying “I would buy” is weak evidence of future sales. A paid pilot is more direct evidence for that particular offer and customer, while still leaving repeat demand and scale unresolved.
Finish with a decision, not more formatting
- Continue: run another bounded test when the current evidence supports it.
- Change: revise one important assumption, such as the audience or scope, and explain why.
- Pause: stop spending while a permission, capacity, or personal constraint is unresolved.
- Stop: close the experiment when it crosses a loss limit or the case no longer makes sense.
Update the plan with actual costs, observations, remaining unknowns, and the next review date. Keep the original assumptions so you can compare them with reality. You are creating a decision record, not rewriting the past to make the plan look correct.
Use the canvas to capture the next experiment, then use the Business Numbers Lab to see whether its proposed economics fit your capacity.
Your next useful step.
Open the Idea Validation Canvas, work through your own assumptions, and save a private copy as a JSON file.
Open the Idea Validation Canvas ↗Sources & scope
Consulted November 5, 2026. Sources support the definitions and methods identified in the text. Examples, exercises, and calculations are our own; they are not case studies or claims of customer outcomes.
General education, with US official resources where noted. This guide does not determine legal obligations, eligibility, or an individual financial decision. Read our editorial policy.