Write a falsifiable claim
Name the customer, problem, current alternative and why they would switch. Add a behaviour you can observe, such as booking a call, paying a deposit or completing a purchase.
Interview past behaviour
Ask when the problem last occurred, what it cost and how the person handled it. Avoid pitching until you understand the sequence. Specific stories are more reliable than answers about an imagined future.
Test the riskiest assumption
If demand is uncertain, use a landing page and direct outreach. If delivery is uncertain, manually fulfil a pilot. If price is uncertain, ask for payment. Do not build the easiest part while the fatal question remains untouched.
Set thresholds in advance
Decide what result means continue, revise or stop before seeing the data. Include traffic quality, conversion, willingness to pay, delivery cost and repeat behaviour. This reduces the temptation to reinterpret weak evidence.
Keep the learning asset
Record objections, customer language, acquisition sources and actual costs. Even a failed idea can produce a valuable map of the market. A successful test should lead to the next smallest commitment, not an immediate leap to scale.