Updated 20 October 2026.
Before building a feature, identify the assumption that would make the work unnecessary if it were false. Choose a small, honest way to examine that assumption, decide what evidence would change your mind, and review the result before committing to a larger build.
The Lean Startup by Eric Ries emphasizes validated learning and experimentation. The official methodology describes testing ideas and using feedback to steer development. The experiment below is an original application of that theme, not a claim that a small test proves product-market fit.
Separate the idea from the assumption
“We should build automated weekly reports” is a proposed solution. It hides several assumptions: customers need a weekly overview, someone acts on it, the necessary data is available, and automation is worth the maintenance.
List the assumptions individually. Then ask which uncertain one would most strongly affect the decision to build. Testing a color preference will not help if you still do not know whether anyone needs the report.
Use what you already know. Existing support conversations may reveal a recurring need, but they do not automatically establish how many customers share it or whether the proposed solution fits.
Define what you need to observe
A useful test has a decision behind it. Write the current belief, what you will observe, and what would make you revise the plan.
“Customers like the concept” is difficult to interpret. A more informative observation could be whether a customer uses a sample report in an actual review and which decision it helps them make.
Set a practical limit on time and effort. The limit is yours to choose for the situation; it is not a universal threshold from Ries or proof that a given sample size is sufficient.
Try a manual version of the learning task
Imagine a team considering that automated report. For a fictional trial, it offers a small number of willing customers a manually prepared sample using only data they have authorized the team to use.
The team explains the manual process honestly and asks when the report would be used. Afterward, it asks what action, if any, the customer took and what was missing.
The test could reveal that customers need an exception alert rather than a weekly document. It could show that the report is useful but too expensive to prepare. It could also produce little usable evidence. These are possible findings, not reported results.
Because the trial is manual, it can explore the value of the information before testing a full automation system. It does not establish technical feasibility or delivery at scale.
Keep observations separate from conclusions
Record what happened before writing what it means. “Two participants opened the sample” differs from “customers need a reporting product.” Note who participated, what they were asked to do, and which parts of the situation may have influenced their response.
Look for alternative explanations. A participant might use the report because a team member is personally helping them. A positive comment may describe politeness rather than a recurring need.
Make the next decision smaller than the original ambition
Choose whether to repeat the test with a changed assumption, investigate technical feasibility, narrow the feature, or stop. Preserve the evidence that led to the choice so a future discussion does not restart from enthusiasm alone.
The startup-validation learning path offers a broader sequence. Here, the deliverable is one assumption, one small test, and one explicit decision.
For recall, explain why a manually produced report can test usefulness while leaving feasibility unresolved. Then write the assumption behind a feature you are considering and the observation that would cause you to reconsider it.
Continue with ReadSprint
Open the related book summary to revisit the book's broader ideas. Use the free Book-to-Action Plan to organize your own takeaways into practice and review phases; the tool arranges the entries you supply rather than validating the method or predicting results.
Source checked 20 October 2026: Eric Ries’ Lean Startup methodology. The source supports the attributed book concept. The scenarios, questions, and practice steps in this article are original ReadSprint suggestions; example outcomes are hypothetical.
