Ímpetus Host
Validation of ideas

Test demand before building: manually deliver the result

A handmade service can reveal whether the problem matters, which steps generate value and what only then deserves automation.

Luiz Philipe (LP)

Luiz Philipe (LP)

Updated 3 October 2026

Test demand before building: manually deliver the result

It's tempting to build an entire platform to prove that there is interest. But when you don't yet know the flow, each screen incorporates assumptions. A manual test reduces risk: you deliver a simple version of the result and observe the end-to-end experience.

Delimit the test promise

Choose a small task, a participant profile and a result that the person can evaluate. Describe what will be done, what will not be done, what information will be needed, and how to end the test.

Don't pretend automation exists. Explain that you are a manual pilot. Transparency helps to obtain honest observations and avoids creating expectations that the process does not yet support.

Do the work and record the steps

Note what the person sends, what needs to be reordered, where there is a wait, and which decisions require human context. Mark repeated tasks and exceptions. This sequence is the first map of the actual process.

Then ask what was most helpful, what raised doubts, and what the person would do next without your help. Compare the response with the observed behavior: completing a task usually says more than declaring satisfaction.

Decide what deserves to become a product

Automate only the section that is repeated, has understood rules and can be checked. If each case requires different judgment, learn more before coding. If no one finishes the flow, investigate the proposal, don't add features by reflex.

Manual testing is not a cheap version in disguise. It is a learning tool to decide if, how and for whom it is worth building.

Frequently asked questions

Do I need a product ready to test?

No. You need an honest and safe test that allows you to observe the essential task. A guided conversation or manual delivery may suffice.

How many people should participate?

Use a group small enough to follow each step and diverse enough to notice relevant differences in the chosen audience.

When should I automate?

When the process is repeated, the exceptions are mapped and you know how to check the quality of the result.

To receive more practical ideas to test and improve your business, subscribe to the Ímpetus newsletter for free.

Subscribe to the Ímpetus newsletter for free

Get news, insights, tips and tools to help grow your business.