Why Your Case Studies Do Not Work (And What to Put in Them)

The genre problem

Most case studies read like this. The client faced challenges. We deployed our proven methodology. Results were transformative. The client is delighted.

Nobody believes it, because nothing in it could have been otherwise. A document that could only ever have had one ending carries no information, and buyers discount it accordingly.

What a buyer is actually doing when they read one

Three things, in order.

Looking for themselves. Same sector, same size, same problem. If they cannot place themselves in the story within a few lines, they stop.

Looking for the hard part. Every real project has one. A case study with no friction reads as marketing; one that names the obstacle reads as memory.

Looking for something forwardable. Given that most of a B2B decision happens in rooms you are not in, the real question is whether this document survives being sent to a sceptical colleague with no context.

The structure that works

1. Who they are, concretely. Sector, size, and situation in two lines. Name them if you have permission, because an anonymous case study is a claim with the evidence removed.
2. The situation before, with a number. "Enquiries had been flat for three quarters" beats "they wanted growth". The starting number makes the ending number mean something.
3. What you actually did, specifically enough that a competent reader could argue with it. Vagueness here reads as either a secret or an absence.
4. What went wrong or took longer. The most persuasive paragraph in the document. It converts the whole piece from advertising into testimony.
5. The result, in their metric. Revenue, qualified leads, cost per acquisition. Not impressions.
6. What you would do differently. Confidence signals competence more reliably than a flawless narrative.

The numbers question

Two failure modes. Refusing to publish any number, which makes the whole thing unfalsifiable. Or publishing a percentage with no base, so "300% increase" might mean one enquiry became four.

Give the base and the period. If a client will not let you publish absolute figures, publish the percentage plus the timeframe plus the starting scale in words. And never round up in the direction that flatters you, because the buyer will meet the real number eventually.

If you have no results worth publishing yet, say so and publish process instead. That is honest, and it is far better than inventing outcomes, which is the fastest way to lose a deal at the reference call.

Where they belong

Not in a gated library. A gated case study mostly prevents the buyer from becoming convinced, and it stops the sceptical colleague reading it at all.

One page per story, on its own URL, plain enough to read on a phone. Link to them from the service pages, quote the strongest numbers on the page where buyers decide, and reference them in proposals. Public proof is also what makes you quotable to AI assistants answering questions about your category.

Frequently asked

What if clients will not be named? Anonymise the company but keep the sector, size, and numbers. "A 40 person UK logistics firm" is credible. "A leading enterprise" is not.

How many do we need? Three good ones covering your main buyer types beats a dozen thin ones.

How does Fixora handle this? We write them from a real interview with the client, include the difficult part, and publish nothing we cannot evidence. Talk to us and get a plan within 48 hours.