Key takeaways
- Write a claim that can be proved wrong.
- Ask about past behaviour, not imagined intentions.
- Make test offers clear and non-deceptive.
- Define a decision threshold before seeing results.
Test a business idea by isolating its riskiest assumption, collecting evidence from real customer behaviour, and deciding in advance what result means continue, change or stop. A useful test reduces uncertainty; it does not merely create encouraging activity.
Turn the idea into assumptions
Write one sentence for each uncertainty: customer, problem, urgency, current alternative, access, willingness to pay and ability to deliver. Then make each falsifiable.
| Vague belief | Testable assumption |
|---|---|
| People need help with admin. | Five of ten UK freelance designers will show a repeated invoicing problem and two will book a £40 set-up session. |
| This market is growing. | Three named competitors have added the same service in the last year and customers can describe why they switched. |
Run evidence-seeking interviews
- Ask for the last time the problem happened.
- Ask what they did, paid and disliked.
- Do not pitch until the evidence questions are complete.
- Record exact themes without naming people unless they consent.
- End with a proportionate next step, not pressure.
Questions such as “Would you buy this?” invite politeness. Ask “What happened the last time?”, “What did that cost?” and “What have you already tried?” Past behaviour is more useful than a hypothetical yes.
Choose the smallest honest test
A manual pilot, paid workshop, clickable prototype, sample deliverable or focused landing page can test different risks. A smoke test must not imply that a finished product exists when it does not. If you accept pre-orders, state the delivery date, refund terms, product status and trader details clearly.
A seven-day validation plan
- Day 1: list assumptions and choose the riskiest one.
- Day 2: identify ten plausible customers and write neutral questions.
- Days 3 and 4: run five interviews; update the problem language.
- Day 5: create the smallest honest offer and delivery outline.
- Day 6: invite a relevant group individually; follow direct-marketing rules.
- Day 7: compare results with the pre-set threshold and decide.
One-page experiment sheet
- We believe
- Which customer will take what action, and why?
- Riskiest assumption
- What must be true for the idea to work?
- Test
- What is the smallest ethical action that produces evidence?
- Pass signal
- What number and quality of responses justify another test?
- Stop signal
- What result means pause, narrow or abandon?
- Learning
- What changed in the offer, audience or delivery?
Sources
Sources were checked on 7 September 2026. External information can change.
Spot something that needs correcting? See the corrections process.