I once watched a prospect read a case study during a call. He skipped the first four paragraphs, scrolled to the results, read them, scrolled back up, searched for something, did not find it, and asked me the question the document should have answered: "How similar was their situation to ours?"
That is the entire job of a case study and almost nobody writes for it. The reader is not evaluating whether you are good. They are checking whether you have solved their specific problem before.
What buyers are actually doing with it
Three uses, and only one of them involves being impressed.
Pattern matching. Is this company like mine, in size, industry, and constraint? If they cannot establish that in the first thirty seconds, they leave.
Risk assessment. What went wrong, how long did it take, what did it require from the client? These are the questions a buying committee asks and the ones case studies systematically omit.
Internal selling. Someone forwards it to a person who controls budget. That person will read the first paragraph and the numbers. Nothing else.
Write for those three uses and the structure follows.
The structure I use
A headline that is a result, with the segment attached. "How a 40 person accounting firm cut cost per lead by 61 percent in five months." Not "Client success story." Not the client's name alone. The segment lets the reader pattern match before they commit to reading.
A summary block at the top, not the bottom. Client type, problem, what we did, results, timeframe. Five lines. This is what gets forwarded and it needs to survive being read alone.
The situation, with numbers. Where they started, quantified. Traffic, revenue, close rate, whatever the relevant baseline was. A case study without a starting point is a story about a company that is doing well, which proves nothing.
Why previous attempts failed. The most persuasive section and the one most often missing. If the client had tried two other approaches, say so and say why they did not work. The reader has probably tried the same things. This is where they recognize themselves.
The approach, with the reasoning visible. Not a list of deliverables. The decisions and the tradeoffs. "We chose to rebuild the three category templates before touching content, because the template was suppressing four hundred pages at once." That sentence teaches something and demonstrates judgment, which is what they are actually buying.
What was hard. The constraint, the setback, the thing that took longer than expected. Including this makes every other claim in the document more believable. A case study where everything went smoothly reads as fiction because every buyer knows projects do not go smoothly.
Results, with a timeframe and a method. The number, the period it covers, and how it was measured. "Organic sessions up 61 percent" is weak. "Organic sessions to product pages up 61 percent between March and August, measured in Search Console, against a flat previous year" is defensible.
What the client had to contribute. Hours, access, decisions. This sets expectations and it filters out prospects who will not do their part, which saves you a bad engagement later.
One quote that is not praise. The best case study quote is not "they were great to work with." It is the client explaining a decision or a moment of doubt. "I was not convinced we should rebuild the templates first and I was wrong" is worth ten compliments.
On honesty and numbers
I am strict about this because the industry is not. Report only what you measured, state the period, state the method, and state the limits.
If the client cannot approve a specific number, use a range or a relative figure and say why the exact figure is withheld. Do not invent a percentage because a round number looks better in a header. Anyone sophisticated enough to be worth selling to can smell an unsourced statistic, and one fabricated number destroys the credibility of every other claim you have ever published.
The same discipline applies to attribution. If you cannot cleanly separate your contribution from a simultaneous product launch, say so. That admission costs you nothing with serious buyers and buys enormous trust, for the reasons in trust signals and conversion.
Getting clients to say yes to the case study
The obstacles are almost always legal review, competitive sensitivity, and the client's own marketing team wanting control.
What works for me: ask at the moment of peak result, not at the end of the engagement. Offer to write it entirely and let them edit. Offer an anonymized version as the fallback, which is worth roughly 70 percent of a named one and gets approved far more often. And give them something back, usually distribution, a co-promoted piece, or a quote from me they can use.
Build it into the contract at signing, with an opt out. A clause saying you may publish results with client approval, requested at a defined milestone, converts a favor into a process.
Formats, in order of what gets used
The written page on your own site is the core asset, because it is indexable, linkable, and forwardable.
A one page PDF version for sales to attach. Different format, same content, designed to survive being printed.
A short video with the client speaking, if they will do it. Highest trust, lowest production reliability, and rarely worth the effort unless the client is genuinely enthusiastic.
Three to five social posts pulling out individual moments, particularly the "what was hard" section, which performs far better than the results section. That is the receipt post format I described in the LinkedIn founder brand system.
Check the promotional versions against platform limits with a social post length checker before you write them rather than after, because trimming a good specific detail to fit a cap is how these turn generic.
Build a matrix, not a pile
The strategic layer: you do not need many case studies, you need coverage.
Map your buyer segments against the problems you solve. Aim for one case study per meaningful cell. Three studies covering three distinct situations sell better than fifteen covering the same one, because the reader is pattern matching and coverage is what determines whether they find themselves.
When a prospect fits a cell you have not covered, that is the next case study to chase, and it should influence which engagement you volunteer to write up.
The distribution question matters too. Case studies buried in a resources tab get read by nobody. They belong linked from the service page relevant to that problem, referenced in sales sequences, and cited in proposals, which is the same internal distribution logic as internal linking architecture.
Add the section you have been avoiding
If you have existing case studies, do not rewrite them. Add one section: what was hard.
Two paragraphs on the constraint, the setback, or the thing that took longer than planned. It is the section every client will hesitate to approve and the one that changes how the document reads.
I have made this single edit to case studies that had been live for a year and watched sales teams start using them, because the document finally sounded like it was describing real work rather than advertising it. Everything else on this list is refinement. That section is the difference between a case study and a brochure.