Choose evidence for a service case study by starting with the claim you want a reader to assess. Then show the relevant work, the measurement method and the limits of the result. A polished screenshot can demonstrate a design change. It cannot, on its own, demonstrate additional sales or explain why customer behavior changed.

If you have no revenue evidence yet, say so. You can still describe the delivered work and report an available measure such as views. Keep that measure clearly named. A useful case study gives a potential customer a reason to trust the evidence without asking them to assume a business result that has not been observed.

Match case study evidence to a specific claim

Write the proposed headline before collecting screenshots. If the headline claims that a task became easier, identify what would support that statement: an observed completion, a documented customer comment or a comparable task measurement. If the only evidence is that a new page was delivered, narrow the headline to describe that delivery.

Make a simple claim register with a source beside each statement. Give unsupported statements their own unresolved status. This makes editing easier because you can remove or revise an overstatement without rebuilding the entire story. Keep a distinction between what your team did and what the client subsequently experienced.

Keep the case study baseline and period visible

For a before and after comparison, record the date range, metric definition and collection method on both sides. If those changed, explain the change rather than treating the results as directly equivalent. Preserve the original export or report so the comparison can be checked later.

The GOV.UK Service Manual discusses measuring service benefits against a baseline and using case studies to communicate value. For your own service story, apply the narrower practical lesson: make the basis of comparison visible. Do not imply that a government service measurement framework certifies your commercial result or establishes causation for a particular project.

Separate measured results from possible explanations

A result can be real while its cause remains uncertain. Record other relevant changes during the observation period, such as a promotion or a different posting schedule, when you know about them. Do not invent missing context. Explain whether your evidence establishes a comparison, a customer opinion or a causal conclusion.

For an early project, a clear statement might report views over a named period and separately state that revenue has not yet been measured. Avoid substituting impressions, followers or page visits for one another. Our guide to measuring rework uses the same principle: name the outcome you actually counted.

Publish evidence the customer can understand

Choose a small number of examples that support the central claim. Label screenshots with the relevant state and date, remove private material and obtain the necessary permission before identifying a client or quoting their feedback. Keep internal source records even when the public example needs redaction.

End with the remaining question or next measurement, rather than a promise that another customer will receive the same result. A potential client should be able to explain what changed, what was observed and what is still unknown after reading the case study. Our process mapping guide can help describe the work clearly before you choose a result to highlight.

Sources

Related

Read next: Describe the work before automating it.