A customer story is a case study rebuilt as narrative: one named person, the problem they actually had, the moment something changed, and the result they can attest to. The difference is structural, not stylistic. A case study is organised by chronology and outcome; a customer story is organised around unresolved tension, which is what holds attention and what makes the outcome mean anything at all. Five beats do the work — before-state, breaking point, decision, change, recommendation — and four of the five have to come out of the customer's mouth rather than your marketing team's, because those four are the parts a prospect can check.
- Most case studies fail because the before-state gets one sentence and the solution gets six paragraphs. The persuasion is all in the before-state.
- Tension is missing by default, not by resistance. Customers will keep the hard part in — if you ask for it.
- A leading question does not produce a weak answer. It manufactures a false one that everybody involved will happily sign off.
- Specificity is the credibility test. "It saved us time" is a claim; "we stopped spending Fridays reconciling invoices" is testimony.
- The most persuasive sentence in a customer story is usually one the customer would never have volunteered — and was entirely willing to say.
Why most case studies aren't stories
Open almost any customer case study and you will find the same three headings: challenge, solution, results. It reads like a résumé — a record of competence with every difficulty removed. Nothing in it is untrue. Nothing in it is memorable either, because there is no point at which the outcome is in doubt.
The failure pattern is measurable in paragraph counts. The before-state gets one sentence: "the team was managing renewals manually." The solution gets six paragraphs, because that part is about you and you know it best. But the reader's interest is not in your process. It is in whether the person in the story had their problem — the same problem — and how bad it was before it went away. Cut the before-state to a sentence and you have removed the only part of the story a prospect can recognise themselves in.
Green and Brock's work on narrative transportation (2000) explains why the tension matters mechanically: absorption in a narrative is what moves belief, and there is nothing to be absorbed into when the ending is visible from the first line. A case study announces its ending in the headline. A story withholds it for ninety seconds, which is the entire trick.
The five beats of a customer story
In order. Each beat does one job, and each has a recognisable sound when it is missing.
Beats one, two, four and five must be spoken by the customer. Beat three is the only one where your own account legitimately adds something, because you were in the room for the sales conversation too.
The questions that surface each beat
These are close to what we actually ask on a shoot day. They are deliberately unhelpful — they do not suggest an answer, and they leave the subject somewhere they have to think.
- Before-state: "Walk me through what a Thursday looked like back then." Then wait. The second half of the answer is the usable half.
- Breaking point: "Was there a point where you thought, we can't keep doing this?"
- Decision: "What made you hesitate?" and "What else were you looking at?"
- Change: "What do you not do any more?" — a better question than "what improved", because it asks for a behaviour rather than a judgement.
- Recommendation: "Who do you know who has this problem right now?"
Set those against the version most teams ask: "How much did we improve your process?" That question contains its own answer. It tells the subject the shape of the reply you want, and — this is the part worth being careful about — they will give it to you. Willingly, warmly, and better phrased on the second take. Our nine-question set for filmed interviews is in the interview questions guide; the principle underneath all nine is that the question must not know where it is going.
Across 31 customer-story films we have delivered, exactly one subject has asked us to change the edit. Subjects do not police the story. They defer to the crew, they assume we know what good footage sounds like, and they will offer a tidier, more quotable version of any answer if we imply the first one was close. Set that beside Green and Brock's finding that narrative moves belief whether it is labelled fact or fiction and you get an uncomfortable conclusion: neither end of the chain is a check. So a leading question does not produce a weak answer — it manufactures a false one that everyone in the process will approve. The count, and its limits, are set out here.
Written, filmed, or both?
They carry different things, and the choice is usually decided by where the story has to work rather than by budget.
| Format | What it carries | Effort | Best placement |
|---|---|---|---|
| Written | Detail, numbers, skimmable structure, search indexing | Low | Site, proposals, sales follow-up |
| Filmed | Tone, hesitation, the pause before an honest sentence | High | Landing pages, pitch meetings, sales calls |
| Both | Film carries belief; the page carries the specifics and the indexable words | High | Anywhere a deal is decided by a group |
The case for film is not production value. It is that the parts of testimony you cannot fake are non-verbal — the half-second of thinking before someone admits what was hard, the change in register when they stop reciting and start remembering. Wyzowl's State of Video Marketing 2026 reports that 89% of people say video quality affects their trust in a brand, from a survey of 266 respondents fielded in late 2025, which is worth exactly as much as its sample. The stronger argument is structural: a written quote can be composed by anyone, and a filmed one is attached to a face that can be looked up. More on the trade in B2B customer testimonial videos.
Structuring for 60 seconds, two minutes, and long form
Runtime is a decision about which beats survive.
Sixty seconds holds three: before-state, change, recommendation. The breaking point is the first casualty, which is precisely why most 60-second testimonials are pleasant and forgettable — you have kept the problem and the result but removed the reason anyone acted. If you only have a minute, cut the recommendation instead and keep the breaking point. A story with stakes and no endorsement outperforms an endorsement with no stakes.
Two minutes holds all five, at roughly 20 seconds each, and this is the format we recommend for anything a buying group will watch together. There is room for one specific number and one moment of admitted doubt.
Long form — four minutes or a 900-word page — earns its length only by adding a second voice: a colleague of the customer, or the person who inherited the process. A second perspective on the same breaking point is the only thing that makes long form feel fuller rather than slower. Repeating one voice for four minutes just dilutes it.
The specificity rule
One rule covers most of the editing decisions in a customer story: a named person, a named number, a named moment. Miss any of the three and the story reads as composed rather than reported.
Compare "it saved us a lot of time" with "we stopped spending Fridays reconciling invoices." Both describe the same outcome. The first is a category of benefit and could be said about any product by anyone, which is why a reader discounts it automatically. The second could only be said by someone who had actually lived that Friday. Vagueness reads as invention; specificity reads as testimony — and the reason is simple, if slightly bleak: a specific claim is easier to check, so making one is a risk, and taking that risk is what signals it is true.
This is also the test to run on your own draft. Take any sentence in it and ask whether a competitor could publish that sentence unchanged about a different customer. If they could, it is doing no work.
Five ways teams ruin a good customer story
- Over-editing the grammar. The false starts and the plain vocabulary are the evidence of a real person. Fix the audio, not the sentences. Fix: edit for length only.
- Cutting the doubt. Removing the part where they nearly chose someone else removes the only moment your prospect is in the story too. Fix: keep one sentence of hesitation, always.
- Leading the witness. Not a soft error. Because subjects comply, you get a sentence that is quotable, approved and untrue. Fix: ask questions that do not contain their answer.
- Composite customers. "One of our clients in financial services" is a claim with nobody behind it. Fix: one named person, or don't publish it as proof.
- Approvals that sand off every specific. Six rounds of review turn a number into "significant improvement". Fix: negotiate detail, not claims — a range, a ratio, or a changed routine instead of a metric.
Frequently asked questions
How long should a customer story be?
Written, 600 to 900 words carries all five beats comfortably. Filmed, two minutes holds the full arc and 60 seconds holds three beats. Set length by how many beats you have real material for, not by a content calendar.
Can we write it for them?
You can draft the structure and edit for length. You should not write the sentences they are quoted as saying. Their approval is not evidence the words are theirs — subjects approve almost anything put in front of them, which is the finding this whole post is built on.
What if the customer won't give a number?
Ask for a changed routine instead of a metric: what they stopped doing, what a Friday looks like now, how long something used to take. A concrete behavioural change is more credible than a percentage nobody can source.
What if legal won't approve the specifics?
Negotiate at the level of detail, not the level of the claim. Use a ratio, a range, or move the specific from the metric to the routine. A story stripped of every specific is not safer — it is just unpersuasive, which has its own cost.
How many stories do we need?
Three well-built stories covering your main buyer types beat a dozen thin ones. Buyers look for the story that resembles their own situation, so coverage of situations matters more than volume.
Who owns the final edit?
You do — and the subject should have an unconditional right to ask for changes. They will almost never use it. One in 31, in our case. That is an argument for more restraint on your side, not less.
The five beats with the exact question for each, the leading-question swaps, and the specificity check to run before you publish. Print one per subject.
Sources
- Green, M. C., & Brock, T. C. (2000). The Role of Transportation in the Persuasiveness of Public Narratives. Journal of Personality and Social Psychology, 79(5), 701–721. Four experiments, N=97, 69, 274, 258.
- Wyzowl, State of Video Marketing 2026. 266 respondents, surveyed late 2025; 12th edition. 89% say video quality affects brand trust.
- SeventhSlate first-hand count: 31 customer-story films delivered to date; one subject requested a change to the edit after review. Own internal record, not an audited sample; method and limits stated in full on the storytelling statistics page.
- Keep the before-state and the breaking point. Everything persuasive in a customer story lives there.
- Ask questions that don't know their own answer, because your subject will answer the question you implied.
- One named person, one named number, one named moment. That is the whole credibility test.
