A case study that sells is built from three parts only: the problem the client actually had, what you actually did about it, and one real number that proves it worked. Everything else is padding, and padding is why most case studies never get read past the first paragraph.
Why most case studies are unreadable
Open ten case study pages from ten agencies and count how many follow the same dead formula: a vague "challenge" paragraph, a list of buzzwords under "solution," and a "results" section that says something improved without saying by how much. Nobody reads past line three of a case study like that, because it reads like it was written to satisfy a template, not to convince a specific buyer of anything.
This matters more than most marketing teams assume. Case studies and customer stories rank as the second most effective B2B content type, with 53 percent of B2B marketers rating them as producing their best results, just behind video at 58 percent, and 75 percent of B2B marketers used case studies or customer stories in their content mix over the past year, according to the Content Marketing Institute's 2025 B2B benchmark research Content Marketing Institute, 2025. Case studies are one of the most-used and most-trusted formats in B2B marketing, and most of them still waste that trust with vague writing.
The three-part structure that actually works
1. The problem, stated in the client's words
Do not open with your company. Open with what was actually broken or missing for the client, described specifically enough that a reader in the same situation recognizes it immediately. "Their checkout abandonment rate was costing them roughly 40 sales a month" beats "they needed to improve their online presence" every time. Specificity signals honesty. Vagueness signals a template.
2. The approach, without the buzzwords
Describe what you actually did, in the order you did it, using words a non-marketer would understand. Skip "leveraged synergies" and "holistic strategy." Say what changed: the pages you rebuilt, the automation you set up, the messaging you rewrote, the timeline it took. If a competitor reading this could not tell what specifically you did differently from anyone else, rewrite it.
3. The result, with one real number
This is the part most case studies get wrong by hedging. "Significant improvement in engagement" is not a result, it is an excuse for not having a real one. A case study needs at least one concrete, specific figure: a percentage, a euro amount, a time saved, a count. If the client will not release exact figures, use a real range or an honest proxy metric rather than inventing precision that does not exist.
The client quote: how to get one that is not generic
Most client quotes are useless because they are collected with the wrong question. Asking "what did you think of working with us" gets you "great experience, highly recommend," which is a testimonial, not proof.
Instead, ask the client three specific questions during the interview:
- "What was the actual moment you knew this was working?"
- "What would you tell someone in your position who is hesitant to start?"
- "What is different in your business now, in numbers if you have them?"
The answers to these questions produce quotes with texture, hesitation, and specificity, the exact qualities that make a quote sound like a real person said it rather than a marketing department wrote it and asked for a signature.
Weak case study versus strong case study
| Weak case study | Strong case study |
|---|---|
| "We helped improve their brand presence" | "We rebuilt their site and cut load time from 6.2s to 1.4s" |
| "Client was very happy with results" | "Client quote: 'I stopped worrying about the site crashing during launch week'" |
| No dates, no numbers, no names | Named client, real timeframe, one specific figure |
| Generic stock photography | Real project screenshots or product photos |
| Buzzword-heavy solution section | Plain description of what was actually built or changed |
Case study writing checklist
- Does the opening line describe a specific problem, not a vague need?
- Is there exactly one real, checkable number in the results section?
- Does the client quote include a specific moment or detail, not just praise?
- Would the client sign off on every sentence as accurate if they read it today?
- Could a competitor tell what you specifically did, not just that you "helped"?
- Is it short enough that a busy prospect reads it in under three minutes?
FAQ
What if the client will not share exact numbers? Ask for a range, a rough order of magnitude, or an honest proxy ("fewer support tickets" rather than an exact percentage). Never invent a number to fill the gap. An honest range beats a fabricated figure every time a client could later contradict it.
How long should a case study be? Long enough to cover the three parts properly, usually 400 to 700 words. Length is not the goal, proof is. A short case study with one real number outperforms a long one with none.
Should we publish case studies for every client? Only for results you can back up with something concrete. A case study without a real number is closer to a testimonial, and should be labeled and used that way rather than dressed up as a case study.
Creative Hero writes and designs case studies built around real numbers and real client quotes, the kind sales teams actually use in a pitch. If your current case studies would not survive a client fact-check, we should look at them together.
