How to Write a Case Study That Persuades Without Exaggerating

Short answer: structure a case study around five parts: the client’s situation in specific terms, the constraint or problem that made it hard, what was actually done and why, what happened — with real numbers where the client agrees — and what did not work or would be done differently. Get the client’s approval, attribute results carefully rather than claiming everything that improved, state the period and context of any numbers, and avoid presenting one result as typical. Specific, honest case studies persuade more than polished ones.

Case studies are among the most persuasive content a business can publish, because they show rather than claim. They are also among the most frequently exaggerated, which is why readers treat many of them sceptically.

The ones that persuade are specific, honest about limits, and written so a reader can recognise their own situation.

The five parts

A reliable structure:

The constraint and the reflection are what make a case study credible. Without them it reads as an advertisement.

Situation: specific enough to recognise

Readers read case studies looking for themselves. Give them enough to do so: sector, size, the situation before, what they had already tried.

A mid-sized manufacturer publishing one article a month and getting no enquiries from it is recognisable. A leading company looking to improve its content is not.

Action: what was actually done

Describe the specific steps and the reasoning. This is where a reader judges whether your approach would work for them, and where your expertise shows.

Include decisions and trade-offs: we considered X, chose Y because of the client’s constraint. Generic lists of activities — strategy, execution, optimisation — say nothing.

Results: honest and attributed

This is where most case studies go wrong.

• State the period and the baseline: from what, to what, over what time.
• Attribute carefully. If other things changed at the same time — a new product, a price change, a season — say so.
• Use the client’s numbers, with their agreement, not estimates.
• Do not imply it is typical. One result is one result.

In many jurisdictions, advertising rules also apply to results claims. fact-checking before you publish covers checking every number before publication.

Reflection: what did not work

Including something that did not work, or that you would do differently, is counter-intuitive and powerful. It shows honesty, it shows learning, and it makes the rest of the story more believable.

It also helps readers: the mistake you describe is one they can now avoid, which is exactly the kind of first-hand material described in adding something original.

Client approval

Always get the client’s approval before publishing a named case study, including the numbers. Many clients are happy to be featured; some want to review wording; some prefer anonymity.

An anonymised case study is still valuable if the specifics are real. Describe the client by sector and size, change identifying details, and say that it is anonymised. Never publish an invented or composite case as if it were real.

Quotes

A quote from the client adds credibility when it says something specific. We saw a clear change in the kind of enquiries we received, and our sales team started using the articles in calls is better than They were great to work with.

Ask a specific question to get a specific answer: what changed for you, what surprised you, what would you tell someone considering this.

Gathering the material

Case studies are easier to write when the material is gathered as the work happens rather than reconstructed months later.

At the start of an engagement, note the baseline figures and the client’s own description of the problem. During it, keep brief notes of decisions and surprises. At the end, ask the client three questions: what changed, what surprised you, and what would you tell someone in your position. With that file, writing the case study takes an hour, and the numbers are ones you can actually stand behind.

Ask about featuring the client early, too. It is an easier conversation while the work is going well than after it has finished.

Where to publish it

Publish case studies on your own site, in a section of their own as well as in the blog, so prospects looking specifically for evidence can find them all in one place. Then link each from the relevant service pages and related articles.

Length and format

A case study for a blog works well at 800 to 1,500 words, with a short summary box at the top: client, challenge, approach, result. Readers who want detail read on; those who want the outline get it immediately.

Link it from the service or product pages it relates to, and from articles on the same problem. Case studies are often most effective when a prospect meets them at the moment they are considering the service — calls to action that fit the article covers placing them.

Related reading

If this was useful, these cover the questions that usually come next.

• Adding something original — case studies are the purest form
• Fact-checking before you publish — checking every number in the story
• A blog for a software business — where use cases fit

The bottom line

Write the situation specifically, include the constraint, describe the real actions and reasoning, state results with their period, baseline and caveats, and add what did not work. Get the client’s approval, anonymise honestly where needed, and use specific quotes. Specific and honest persuades more than polished and inflated.

DUK

What should a case study include?

The client’s situation, the constraint that made it hard, what was done and why, what happened with real numbers and context, and what did not work or would be done differently.

How long should a case study be?

Around 800 to 1,500 words for a blog, with a short summary box at the top for readers who want only the outline.

Do I need client permission for a case study?

Yes for a named case study, including the numbers. An anonymised one with real specifics is a good alternative when a client prefers not to be named.

How do I present results honestly?

Give the period and baseline, note anything else that changed at the same time, use the client’s own numbers, and do not imply one result is typical.

Should a case study mention what went wrong?

Yes. It makes the rest of the story credible and gives readers something useful to avoid.

Can I combine several clients into one case study?

Only if you say clearly that it is a composite. Presenting a composite or invented case as a real client is misleading.