How We Write Case Studies That Win Work

We rewrote every case study on our site this year, and the exercise taught us more about our own weaknesses than any offsite ever has. The old versions followed the industry template: challenge, approach, four paragraphs about workshops, a gallery of the prettiest deliverables, and a closing stat with suspicious precision. They were fine. Fine is invisible. Prospective clients told us, politely, that our case studies read like everyone else's, which was fatal, since distinctiveness is literally what we sell.

So we studied how clients actually read case studies, asked recent clients what had convinced them, and rebuilt around three ingredients: stakes, decisions, and proof. New business conversations changed within a quarter. Here is the playbook.

Stakes: what would have happened otherwise

The standard case study opens with a client description and a brief. Nobody cares. What makes a story a story is what was at risk. The rewrite rule: the first paragraph must contain the consequence of failure.

Not "Møller wanted to refresh its brand for a new generation," but the truth with stakes attached: a beloved heritage brand watching its average customer age up a year every year, facing the specific arithmetic of when admiration turns into irrelevance. Now the reader knows why any of the following work mattered, and, more important, they recognize their own 2 a.m. version of the same fear. Case studies persuade by recognition, not admiration.

Decisions: show the fork in the road

Process sections are where case studies go to die. Workshops, sprints, alignment, iteration: every agency claims the same verbs, so the verbs carry no information. What carries information is decisions, especially the ones that had a real alternative.

We now structure the middle of every case study around two or three forks: here is what most brands in this position would have done, here is what we did instead, and here is why. The courage is in the specifics. Admitting we killed the first creative direction after testing. Explaining why we advised spending less on the launch film and more on the retail environment. Clients are not hiring a portfolio; they are hiring judgment, and judgment is only visible at forks in the road.

Proof: numbers with their clothes on

Every agency site has a wall of suspiciously round percentages. Uplift claims without a baseline, a timeframe, or a mechanism are decoration, and sophisticated buyers discount them to zero. Our proof rules now:

• Every number gets context: the baseline, the period, and what else was happening that could explain it.
• We name what we cannot claim. If the client also cut prices that quarter, we say so.
• Qualitative proof is quoted, attributed, and specific. "They pushed us harder than we pushed them" from a named CMO outperforms any anonymous superlative.
• One honest limitation per case study, minimum. Nothing builds credibility faster than an agency admitting what did not work.

Write for the second reader

The person who finds your case study is rarely the person who signs. They forward it, with a one-line summary they compose themselves. So we now write the forwardable version into the piece: a headline that states the outcome, a first paragraph with the stakes, subheads that carry the argument for skimmers. If a busy executive reads only the bold text, they should still get the story. Design the reading experience the way you would design anything else: for the reality of attention, not the fantasy of it.

The meta-lesson sits nicely with our clients too. Everything above is just brand strategy applied to ourselves: differentiate or disappear, claim only what you can prove, respect the audience's time. Agencies give this advice daily and take it rarely. The rewrite was us taking our own medicine, and we can report the medicine works.