Photo by Lauren Mancke on Unsplash. Illustrative workspace photograph.
A client case study helps a prospective buyer understand a real project: the problem, the decisions, the work and the outcome. The difficult part often arrives when the work is finished but the evidence is incomplete. A client remembers that things improved, a dashboard has changed, and nobody recorded the starting point.
You can still write a useful account. Start by identifying what the records actually support, then build the story around that evidence. The process below is a practical editorial framework, not a report of a Nadouri Digital client project.
1. Choose a question the case study can answer
Give the piece one central question. For example: How did the team simplify an enquiry process? How did it replace inconsistent product copy? What changed when it introduced a shared editorial brief?
Choose a question that fits the available records. If you have approved drafts, revision notes and a completed handover, you can explain a delivery process. You cannot infer a revenue increase from those materials. A narrower, well-supported story is more useful than a broad success claim with gaps.
A process-focused case study can draw on an approved brief, a marked-up draft and the final deliverable. Compare those materials to identify decisions worth explaining: a confusing instruction removed, a missing answer added or a review step clarified. Choose only changes visible in the records. This gives readers something practical to evaluate even when the project has no reliable sales or conversion data.
2. Build a small evidence record before interviewing
Create a working list with five fields for each proposed claim:
- Claim: the exact sentence you want readers to believe.
- Source: a dated report, approved document, project record or attributable interview answer.
- Definition: what the measure includes and excludes.
- Period: when the observation was made and which comparison period applies.
- Publication permission: whether this information can appear publicly.
Keep the underlying material in an approved workspace. A source can substantiate a statement without being suitable to publish. Before using a screenshot, check for customer details, confidential figures and information outside the case study’s scope.
Give each source a short identifier and keep a copy of the version used. Link it to the relevant sentence in your working draft, so a reviewer can trace a claim without searching an entire project folder. When records disagree, resolve the discrepancy with the person responsible for the data. If it remains unresolved, omit the figure or explain the uncertainty; selecting the most flattering version weakens the account.
The US Federal Trade Commission’s advertising guidance states that advertising claims must be truthful and evidence-based. That provides a useful reason to check the claims in a marketing case study before release. Requirements vary by market and subject; sensitive or regulated claims may need specialist review.
3. Separate delivery, observation and explanation
Sort your notes into three groups. Delivery describes what was completed, such as a revised navigation or an agreed set of pages. Observation describes what the evidence shows afterward. Explanation describes why someone believes a change occurred.
Keep those distinctions visible in the article. A completed website redesign is a deliverable. A change in recorded enquiries is an observation. Claiming that the redesign caused the entire change requires stronger support, especially if pricing, advertising, tracking or seasonality also changed.
A practical claim check
Consider the proposed claim “The new enquiry page made the sales team more efficient.” An approved page can establish that the content changed, but it does not establish time saved. To assess efficiency, look for a defined measure such as time spent clarifying incomplete enquiries, collected consistently before and after the change.
Without that evidence, describe the verifiable work instead: which questions the page answers, how the form changed and who approved it. A client may also offer a genuine observation about the new process; attribute it clearly and get permission to use it. These are examples of how to evaluate wording, not results from a client engagement.
When causation is uncertain, say what happened during the measured period and name the other relevant changes. For a broader measurement workflow, see our guide to measuring content marketing ROI.
4. Ask interview questions that invite specifics
A short interview can explain decisions that a spreadsheet cannot. Ask:
- What was difficult before the project, and can you show an example?
- Which constraint most affected the work?
- What did the team try, change or decide against?
- What can you confirm has improved, and what is still unknown?
- Which statement would you be comfortable approving for publication?
Retain the speaker’s meaning when editing. Send the proposed quotation back for approval, including its attribution and surrounding context. Do not turn a general compliment into a specific performance claim or combine different people’s answers into a single quotation.
5. Use numbers with enough context to be understood
For each result, name the metric, source, dates and relevant baseline. Check that both periods use the same definition and tracking method. If one period excludes spam enquiries while another includes them, the comparison needs correction or a clear limitation.
Use this sentence scaffold only after filling it with verified information: “According to [source], [defined metric] changed from [baseline] during [first period] to [result] during [second period]. During that time, [relevant additional changes] also occurred.” Remove every placeholder before publication.
If no reliable baseline exists, describe the current measured state without calculating improvement. If no outcome data exists, explain the completed work, the documented decisions and what the team plans to measure next. Label an early result as early; do not present a target as an achievement.
Keep rates and counts distinct. More enquiries can coincide with more traffic, so the enquiry count alone does not establish a better conversion rate. When reporting a rate, identify its denominator, such as eligible sessions or completed applications. Distinguish a percentage-point change from a relative percentage change, and have someone check the calculation against the source. Avoid rounding a small change into a dramatic headline.
6. Write a structure readers can check
- Context: who the project served, within the permission granted.
- Challenge: the specific problem and constraints.
- Approach: the decisions and actions that addressed it.
- Evidence: verified results or completed deliverables, with dates.
- Limits: missing data, other contributing changes and unanswered questions.
- Takeaway: what a reader can reasonably learn from this project.
Google’s people-first content guidance encourages original information, clear sourcing and descriptive titles without exaggeration. Apply that as an editorial check, rather than a promise that a particular case-study format will rank.
7. Approve the evidence and the wording together
Ask the client’s designated reviewer to check names, quotations, figures, images, confidentiality and the final interpretation. Record what was approved and which version it applies to. If permission is limited to an anonymized account, make sure the remaining details do not identify the client indirectly.
Keep an approval record that identifies the article version, the reviewer and any restrictions on reuse. Approval for a case-study page may not cover a paid advertisement or a different quotation in a sales deck. When adapting the story, check the permitted use and preserve the qualifying language. Set a review date if the article includes information likely to change, such as the client’s service offering or a still-developing measurement period.
Finish with a separate editing pass. Check that the title makes no stronger claim than the body, every number has context, and the conclusion stays within the evidence. A useful case study lets readers understand the work and its limits without asking them to take invented results on trust.