Key Takeaways
- A useful case study can explain the problem, constraints, decisions and verified changes without publishing revenue or conversion figures.
- Separate what you delivered, what the client observed and what you can prove about business outcomes.
- Get approval for the story, quotations, names and visuals. Removing the client's name does not automatically remove confidentiality concerns.
- Do not replace missing data with an invented percentage, an unsupported success claim or a testimonial that implies more than it proves.
- Give the reader enough detail to judge whether your approach is relevant to their situation.
You can write a strong client case study without sharing the numbers. The answer is to make the work understandable, not to disguise the absence of financial data with bigger adjectives.
A prospective client often wants to know how you think. Did you understand the problem? What did you prioritise? How did you handle awkward constraints? What was actually delivered? A clear account of those decisions can be useful even when sales, traffic or conversion figures must remain private.
The limitation matters too. If you cannot verify a commercial outcome, do not imply that the project caused it. A case study should help someone assess your suitability, not leave them with an impression you cannot support.
Decide What the Reader Needs to Learn
Start with the question the case study should answer. “Can this agency simplify a complicated service offer?” is more useful than “Can we make this project sound impressive?”
Choose one main challenge. It might be a confusing enquiry journey, inconsistent brand materials or a website that the internal team could not update confidently. A focused story is easier to support than a sweeping account claiming to have transformed the entire business.
Connect that challenge to the type of work you want to attract. If you provide brand strategy, explain how the positioning decisions informed the output. If the project was mainly production work, describe that honestly rather than retrofitting a strategy narrative.
Our guide to reviews, testimonials and case studies explains their different roles. A case study has room to show context and reasoning that a short review cannot provide.
Make an Evidence Inventory Before Writing
Gather the material that already exists. Useful sources include the approved brief, workshop notes, signed-off designs, delivery records, documented user feedback and a client's own description of the previous process.
For each item, record whether it is accurate, current and permitted for publication. An internal screenshot may be useful evidence for you but unsuitable for a public page. A draft design may not represent the final delivered work.
| Evidence | What it can support | What it cannot prove on its own |
|---|---|---|
| Approved before-and-after designs | A visible design or content change | A sales or conversion improvement |
| Delivery checklist | What was completed and accepted | Every future business outcome |
| Client quotation | That person's stated experience | An objective claim beyond the evidence |
| Documented workflow change | A specific operational difference | A quantified saving without measurement |
| Usability observation | What happened in the stated test | How every customer will behave |
This distinction keeps the story credible. It also prevents a writer from turning “the team found it easier to edit” into “productivity doubled” because the stronger sentence sounded better.
Separate Delivery, Observation and Outcome
Use three simple labels while drafting. Delivery is what you made or changed. Observation is what someone saw or experienced. Outcome is a business result supported by suitable evidence.
For example, “We introduced a reusable service-page template” describes delivery. “The marketing team said the new structure made routine edits clearer” reports an attributed observation. “The redesign increased qualified enquiries” is an outcome claim and needs evidence appropriate to that claim.
The ASA's substantiation guidance explains that objective advertising claims require objective support. A genuine testimonial does not automatically substantiate an effectiveness claim.
You do not need to publish every piece of supporting evidence, but you should know what supports each assertion. Where a result is confidential, agree whether a narrower, non-numerical statement can be published and whether it remains accurate without the missing context. If not, leave it out.
Ask the Client Better Interview Questions
A short interview often reveals more useful material than a request for “a nice testimonial”. Ask about the situation before the project and the choices that mattered.
Try questions such as: What was difficult for your team? What did customers misunderstand? Which part of the brief could not change? What made you approve the final direction? What can your team do now that it could not do before?
Ask for specific examples rather than praise. “We used to ask the designer for every event graphic; now the team starts from approved templates” tells a reader more than “the service was amazing”. It still needs to be accurate and approved, but it explains a practical change.
Record quotations faithfully. You can suggest edits for clarity, but the client should approve the final wording and attribution. Do not combine several remarks into a polished quotation that appears verbatim unless they have explicitly accepted it as their statement.
Use a Structure That Makes the Work Visible
A simple case-study structure is usually enough: context, problem, constraints, approach, delivered work, verified change and next steps. The interest comes from the decisions inside it, not from unusual headings.
Explain why you rejected alternatives. Perhaps the business needed a clearer message before new visual design. Perhaps the content team needed fewer page types, not more animation. These details show judgement without requiring a confidential revenue chart.
Keep the technical detail proportionate to the reader. A business owner may care that staff can update a location page without breaking the layout. They may not need a long account of every component or plugin involved.
For brand identity work, show how the system was intended to behave across real applications. A logo displayed alone says less than a coherent set of approved examples with a clear explanation of the design choices.
A Worked Example Without Invented Results
Consider a hypothetical regional maintenance company. This is an illustrative writing example, not a MattDarm client claim.
The brief says its services are spread across several overlapping pages, and the office team repeatedly explains which enquiry form customers should use. The agreed work is to clarify the service structure, rewrite key introductions and create a consistent enquiry route.
A defensible case-study paragraph might say: “The new structure gives each core service a dedicated page with a clear description of the work covered. Enquiry links now lead to one form that asks visitors to select the relevant service. The office team reviewed the wording and routing before launch.”
That paragraph describes verifiable work. It does not claim fewer calls, higher conversion or more revenue. If those results are later measured and approved for publication, the case study can be updated. Until then, the story remains useful because a similar business can recognise the problem and evaluate the approach.
Use Visuals as Evidence, Not Decoration
Choose images that demonstrate the change. A before-and-after navigation example, an approved template set or a short annotated process diagram can explain more than a generic meeting photograph.
Make comparisons fair. Do not choose the worst possible old screenshot and an unusually polished new mock-up without explaining the context. Label concepts, prototypes and live work accurately. If the visual has been simplified to protect information, say so.
Check rights as well as confidentiality. GOV.UK's copyright ownership guidance explains that commissioning work does not necessarily make the commissioner the first copyright owner. Contracts and licences matter. The guidance on using somebody else's copyright material is a useful starting point when deciding what you may reproduce.
Avoid including personal information, account details or unpublished commercial plans in screenshots. Blurring a small section is not a substitute for checking the whole image carefully.
Make Approval Specific
Send the client the complete proposed page, not just the quotation. The title, captions and social copy can introduce claims that were not present in the main text.
Ask them to check facts, permissions, names, roles, dates and the scope of the published work. Make any anonymisation agreement explicit. A distinctive project description can identify a business even when its name is removed.
Keep a record of the approved version and who authorised it. If the page changes substantially later, obtain any further approval required by the agreement. Do not treat permission to show one design as permission to publish every internal document associated with the project.
This is a practical editorial process, not legal advice. Where confidentiality or licensing terms are unclear, have the relevant agreement reviewed before publication.
Write Like a Person Who Understands the Work
Replace broad claims with concrete explanations. “We transformed the customer experience” gives the reader little to assess. “We moved the eligibility information above the enquiry form so visitors could check the service was suitable” describes a decision.
Use the client's vocabulary where it helps understanding. Keep your own tone clear and natural. Our article on human brand voice in the age of AI covers why recognisable language matters more than polished generic enthusiasm.
End with a relevant next step rather than a claim that the same result is guaranteed for everyone. A reader with a similar problem can review the related service or discuss their own constraints.
If you are building a content marketing plan, start with one project whose decisions you can explain and whose evidence you can publish. A specific, approved case study is more persuasive than an impressive-sounding story nobody can verify.
Frequently Asked Questions
Can a case study work without statistics?
Yes. It can explain the original problem, constraints, decisions, delivered work and approved observations. Be clear about the evidence and avoid implying commercial outcomes you cannot substantiate.
Can we publish a case study anonymously?
Only within the relevant permissions and confidentiality terms. Removing a name does not guarantee anonymity, particularly when the project details are distinctive. Agree the complete published version with the appropriate client contact.
Can a testimonial replace missing performance data?
A testimonial can report the client's genuine experience, but it does not automatically prove an objective effectiveness claim. Keep the quotation accurate and separate it from unsupported claims about sales, savings or conversion.
What should we show instead of a revenue chart?
Use approved deliverables, fair before-and-after examples, process changes and specific design or content decisions. Explain what each item demonstrates and label prototypes or illustrative examples accurately.
Who should approve the case study?
The person with authority under the client agreement should review the full proposed page, including quotations, images, captions and associated claims. Keep the approved version and revisit permission when material changes are made.
