Separate evidence of a problem, willingness to take a next step and your ability to deliver. One encouraging conversation does not establish all three.
Name the assumptions that could stop the service
Start with a plain description of the service: for whom, in what situation, with what deliverable. Underline the parts you do not yet know. Perhaps the problem is infrequent, the buyer is someone else, the necessary information is unavailable, or the work takes longer than your proposed process allows.
The SBA’s planning material distinguishes research about customers from analysis of available alternatives. It provides a starting point for investigation, not proof that your particular service has demand. Read the original planning guidance.
Keep the first version small enough to investigate. “A monthly product-information review for independent shops” is easier to discuss than “a complete growth solution for every small business.” Our positioning worksheet can help make the customer, situation and boundaries explicit.
Keep three kinds of evidence apart
| Assumption | Useful observation | What remains unknown |
|---|---|---|
| The problem matters | A person describes a recent occurrence, its consequence and what they did. | How common it is beyond the people you spoke with. |
| A next step is worth taking | A relevant buyer agrees to a defined review, trial or proposal discussion. | Whether the arrangement will become a purchase. |
| You can deliver the service | A bounded trial reveals the work, inputs, time and review requirements. | Whether you can repeat it under different conditions or at a larger scale. |
A compliment is evidence that someone expressed approval. It is not automatically a commitment, a purchase or a delivery test. Likewise, a completed trial demonstrates that trial under its conditions; it does not establish unlimited capacity or a profitable business model.
Ask about the current situation before pitching the future
Invite a relevant person to describe the last time they faced the problem. What triggered it? What did they try? Who participated in the decision? What remained unresolved? Use follow-up questions to understand the sequence rather than steering the conversation toward your preferred service.
Keep a short record of observed statements, your interpretation and the missing information. With permission, capture the detail needed for the work; an anonymized summary may be enough. Do not present invented interview quotes as research or assume that a small convenience sample represents an entire market.
If people describe different problems, preserve those differences. One shop may need photography, another inventory information and another a reliable review process. Combining them into “everyone needs better content” would hide the decision your offer needs to make. The existing customer-question guide helps turn recurring questions into clearer explanations after you understand them.
Design a bounded next test
Choose one assumption and write what observation would make you continue, change direction or stop. Set the scope, required inputs, responsibilities and endpoint before beginning. A test should not quietly become an open-ended promise to provide the full service.
Stanford d.school’s Design Thinking Bootleg describes modes including prototyping and testing. It is a useful methodological reference; our evidence log and example below are original editorial aids, not a Stanford-certified validation process. Explore the original resource.
For a service, a bounded test might be reviewing a small supplied sample with a customer and checking whether the proposed deliverable answers the agreed question. If money or a customer project is involved, agree the actual terms separately. A landing page or mockup should not claim capabilities you have not established.
A fictional test and a decision it could support
Fictional scenario: a founder proposes a product-information review for independent shops. The uncertain assumption is that a checklist and short review will help the shop prepare usable information without repeated clarification.
| Part | Proposed record |
|---|---|
| Input | A small product sample the shop is permitted to share. |
| Task | Use the checklist to prepare the information, then discuss the unclear fields. |
| Observation | Which fields were completed, what clarification was needed and what work the review required. |
| Decision | Revise confusing fields; reconsider the offer if its real need is outside your scope. |
| Unknown | Wider demand, repeatability, willingness to buy and the eventual commercial outcome. |
This plan contains no findings. Before running it, the founder would decide what makes the trial useful and whether the proposed work is feasible. Afterwards, the record should report what actually happened, including unexpected results and an unfinished test. Do not replace an inconvenient observation with a testimonial.
Download the log and decide what to do next
Download the service-validation log
For each assumption, record the planned observation, evidence collected, alternative explanation, limitation and next decision. Leave unavailable fields marked unknown. The log is a working document, not a percentage score for how “validated” your business is.
Use the Content Brief Builder when turning the offer into a page: distinguish the proposed answer from verified evidence and unanswered questions. If the next step is a proposal conversation, use our follow-up guide to keep the decision and contact plan clear.
You may decide to narrow the offer, test a different assumption, postpone the launch or proceed with a clearly bounded service. Any of those can be a useful outcome if it follows from what you learned rather than from pressure to declare the idea a success.
Follow the evidence.
- U.S. Small Business Administration · Plan your business ↗
- Stanford d.school · Design Thinking Bootleg ↗
References checked Oct 16, 2026. Examples and workflow recommendations are editorial illustrations, not measured results. Provider guidance may change. Our editorial standards.
