Section 1
Result first, story second
Open Stripe's customer stories and notice what does the talking: named companies, and outcomes stated up front, often in the title itself. The reader knows the destination before the journey begins, which is exactly how skeptical buyers want it, they are not reading for suspense, they are scanning for relevance and proof. Compare that with the standard service-firm format, 'How We Helped a Leading Company Transform,' which withholds the client, the industry, and the number, and therefore communicates nothing scannable at all. The result-first rule has a second benefit: it forces internal honesty. If a project produced no stateable outcome, it becomes obvious at the headline stage, and the piece either gets reframed around a concrete qualitative result or does not get published. A case study archive curated this way stays short and devastating instead of long and forgettable, and buyers read it the way they read evidence. The thinking here builds on [Case Study: How Category-Leading Brands Built a Movement with Storytelling](/blog/case-study-how-built-a-movement-with-storytelling).
Section 2
Five patterns from the best public examples
The table below assembles the case-study patterns visible across three very different proof cultures, Stripe's metric-driven stories, IDEO's design narratives, and Pentagram's portfolio-as-interface. The point is not to pick one culture but to notice they agree on structure: specificity up front, the work made visible, and a real human voice attached. For a service business, the minimum viable case study uses the first, second, and fourth rows, a result headline, a short situation-constraint-intervention-outcome arc, and two or three numbers pulled out as large callouts. The grid and the client quote can come later as the archive grows. What can never come later is specificity; a vague case study published today is a credibility cost, not an asset, every day it stays up.
Section 3
Write for the scan, and disclose the hard parts
Nielsen Norman Group's reading research (1997) found that 79% of web users scan rather than read, which means a case study's headline, subheads, callout numbers, and pull quote must tell the complete story alone, with prose serving only the minority who commit. Structure every case study so a ten-second scan delivers client, problem, intervention, and result. Then do the counterintuitive thing: include what was hard. NN/g's four credibility factors put up-front disclosure near the center of trustworthiness, and a case study that admits a constraint or a misstep reads as testimony rather than advertising. Sophisticated buyers have read hundreds of sanitized success stories; the one that says 'the first approach failed, here is what we changed' earns more belief with one sentence than a page of superlatives. Friction is what makes a story credible, and credibility, not polish, is the entire job of this page. For the step that usually comes next, see [Post-Launch Website Iteration: What to Do in the First 90 Days](/blog/post-launch-website-iteration-first-90-days).
Section 4
When you cannot name the client
Confidentiality is real in services, but anonymity does not have to mean vagueness. The pattern used across professional research applies here: McKinsey's Business Value of Design study names no individual company, yet it persuades because the specifics are rigorous, 300 companies tracked over five years, with top-quartile design performers posting 32 percentage points higher revenue growth than peers. An anonymized case study earns belief the same way: precise industry, company size, timeline, baseline numbers, and method, with only the name withheld. 'A 14-person HVAC company in the Southeast' is checkable in spirit; 'a leading services firm' is not. This evidentiary standard is also why this series analyzes public examples rather than inventing composite wins, and it is the standard ConvertOS applies when building proof systems inside LeverageOS. If your archive currently proves nothing a buyer could test, that gap is worth a strategy call.