Who this is for: Technical managers and businesses assessing a prototype before funding another iteration or a pilot batch.
Prototype testing is useful when its result supports a decision. Before starting, establish what you need to assess, how it will be observed and which outcome will be acceptable. Choosing the threshold after seeing the result makes it difficult to distinguish successful verification from an undeclared compromise.
The programme should fit the sample’s purpose and the significance of failure. Tests do not all need to be elaborate, but they should be understandable and repeatable within the agreed context. Tests involving particular hazards need suitable expertise, equipment and conditions, established before execution rather than improvised when the sample is already on the bench.
Connect the requirement, method and acceptance criterion
Start with the required behaviour and the conditions under which it matters. Then define the evidence-gathering method and the criterion used to interpret the result. NASA’s distinction between requirement verification and intended-use validation is useful here: a favourable bench result should not silently become a conclusion about applications that were never examined.
Thresholds should come from the intended application and agreed specification, not generic online values. If a measurable requirement is missing, an exploratory test may be needed first. Its purpose is to inform the specification. Distinguish it from an acceptance test, which evaluates a requirement already established before the result is known.
Scroll horizontally to see every column.
| Record field | Information to capture |
|---|---|
| Question | The decision the result must support |
| Configuration | Sample, revision, materials and relevant components |
| Conditions | Agreed environment, preparation and operating method |
| Observation | The measurement or behaviour and how it is recorded |
| Criterion | Acceptable result, limitations and handling of uncertainty |
References: NASA — Systems Engineering Handbook: product verification and validation
Know exactly which specimen was tested
Identify the sample before the test and maintain the link to its model, materials and modifications. Any adjustment made during the test belongs in the record. If two specimens have different configurations, their comparison should not be described as repetition under identical conditions. That distinction matters when deciding whether a change caused an improvement.
The initial condition also matters: previous use, assembly and preparation may affect interpretation. Photographs and a short configuration sheet help reconstruct the setup without relying on someone’s memory. Focus the detail on variables that could change the answer, keeping the record useful enough that another person can understand what was assessed and what remains unknown.
Record anomalies and incomplete results
Distinguish completed, interrupted and unperformed tests. Missing data is not a positive result. An interruption needs its reason recorded and a decision on whether the partial observation is usable. Unexpected behaviour also deserves a record, even if it was outside the main test objective. It may influence what should be examined next.
For measurements, identify the instrument and method and assess whether they suit the characteristic being observed. Do not present more precision than the method supports. If a result is close to a limit, account for measurement reliability and the agreed treatment of uncertain cases. The conclusion should follow that method rather than the team’s preference for a favourable answer.
Turn the report into a development decision
The report should connect each finding to a requirement, note deviations and state the conclusion supported by the evidence. The next action may be a design change, repetition, a different experiment or progression to the next stage. Assign an owner and a closure condition so that unresolved findings do not disappear when the project moves on.
A sample accepted for one purpose does not become suitable for every possible use. Likewise, an internal test report is not a certification or a substitute for any applicable regulatory documentation. Before a pilot batch, review which conclusions remain valid if materials, processes, components or operating conditions change. This preserves the value of earlier work without claiming more than it established.
Before starting prototype tests
- Link each test to a requirement or an explicitly exploratory question.
- Identify specimens, revisions and differences from the intended product.
- Define conditions, instruments, responsibilities and stopping criteria.
- Agree acceptable outcomes and how uncertain results will be handled.
- Prepare a record for modifications, anomalies and follow-up actions.
Common questions
Can one prototype validate a product?
It depends on the question and the variability being assessed. One specimen may clarify a principle or interface but cannot by itself establish production repeatability. Sample numbers and configurations should follow the assessment objective.
What happens when a test fails its criterion?
Record the result and conditions, distinguish possible causes and decide which evidence is needed before changing the design. Repeating a test without changing or clarifying anything may consume resources without resolving the uncertainty.
Sources and further reading
Apply this to your project.
Tell us which functions need assessment, what evidence you already have and the decision that follows. We can help define a proportionate prototype testing scope.
Discuss your prototype