Validation Evidence
Validation evidence is the information a product team collects to judge whether an assumption or hypothesis about customers, the market or the product holds up. Its value depends on its strength. Observed behavior and real commitments, such as payments or repeated use, count for more than opinions, predictions or polite interest captured in interviews and surveys.
How Validation Evidence Works
Evidence always relates to a specific hypothesis and is judged against a success criterion set before the test. In Testing Business Ideas, David Bland and Alex Osterwalder describe evidence getting stronger along four axes:
- From opinions to facts. "I would use this" is weak. "Last month I spent four hours rebuilding that report by hand" is a fact about past behavior.
- From what people say to what they do. Stated intent is weaker than observed action.
- From lab settings to the real world. Behavior in a test session is weaker than behavior in normal use, where people do not know they are being studied.
- From small to large commitments. An email address is a small commitment. A pre-order, a signed pilot or time spent integrating the product is a large one.
Strength is not the only question. Teams also check relevance (did it come from the target segment?), quantity (one enthusiastic customer is a lead, not a pattern), recency, and whether the evidence was gathered in a way that could have produced the opposite result.
Why Validation Evidence Matters
Product decisions such as investing, pivoting or stopping should rest on evidence in proportion to their cost. A cheap reversible change needs little. A quarter of engineering work needs much more. Being explicit about evidence strength protects teams from confirmation bias, where weak positive signals are counted and inconvenient ones ignored.
Recording evidence with its source also keeps decisions traceable. When someone later asks why a feature was built, the answer can point to specific interviews, usage data or payments, not to a memory of a meeting. For a view of applying consistent evidence standards across many products, see standardizing validation evidence across a startup portfolio.
Validation Evidence Example
A team is deciding whether to build an automated audit-report feature for compliance managers. It has three pieces of evidence:
- Twenty survey respondents say they would "definitely use" it. This is weak: opinion, stated intent, no commitment.
- In six customer interviews, managers describe how they assembled their last audit report, how long it took and what it cost them. This is stronger: facts about past behavior.
- Of thirty trial accounts shown a "request early access" option with a stated price, four sign a paid pilot agreement. This is the strongest: real-world behavior and a meaningful commitment.
The team treats the third item as decisive and the first as supporting context only.
Validation Evidence vs. Data
Not all data is validation evidence. Page views, sign-ups or survey scores become evidence only when they are tied to a specific hypothesis and a criterion that could have proven it wrong. A dashboard full of metrics with no hypothesis behind it describes what is happening but does not validate anything. The step from evidence to a confident decision on a problem is what problem-solution fit assessments rely on.