The Case Study Page Structure That Wins Work
Case study page structure that converts: situation, constraint, decision, work, outcome — plus how to interview a client for one in twenty minutes.
Open the "Our Work" page of almost any Australian service business and you'll find the same thing: a grid of twelve tiles, each one a client logo over a photo, each one labelled with an industry and a year. Click one and you get a full-bleed image, three sentences of copy, and a link to the live site. The prospect who clicked arrives back at the grid knowing exactly as much as they did before.
That page sits at the end of the consideration path, gets read by people who are close to enquiring, and is built as a scrapbook. The work is real, the clients are real, and none of it is arranged to answer a question the buyer actually has.
A case study is a sales document, not a portfolio entry. It exists to let a stranger recognise their own situation in someone else's, then watch how it was handled. Everything about the format follows from that — including the uncomfortable part, which is that the interesting content is the client's problem, not your solution.
The claim that gets pushback: one properly written case study beats nine tiles. Volume feels like proof and reads like filler. A buyer with a $40,000 problem does not need to know you've done forty projects. They need one story close enough to theirs that they can stop worrying.
What the prospect is actually doing on that page
Nobody lands on a case study page cold. They've read the homepage, skimmed a service page, decided you might be plausible, and now they're looking for the thing that stops them from second-guessing the enquiry.
They're running three checks, roughly in this order. Has this studio handled a situation shaped like mine — not the same industry necessarily, but the same constraint, the same size, the same mess. What is working with these people actually like, week to week. And is there any evidence they did something other than make it look nicer.
The third one is the killer for design-led businesses. A beautiful screenshot proves taste. It doesn't prove judgement, and judgement is what the buyer is nervous about, because they've usually been burned by someone whose taste was fine. The homepage that closes large briefs works on the same principle: at higher price points, the buying decision moves away from aesthetics and toward evidence of thinking.
A case study page is the only page on your site where you can answer all three checks at once. Which is why it's worth more attention than the grid of tiles suggests it's getting.
The five-part structure that works
The order below is not arbitrary. Each part sets up the anxiety that the next part resolves.
Most agency case studies start at "work" and stay there. The page opens with a hero shot of the new site, moves into a list of features, and closes with a link. Everything the buyer needed — the situation they might share, the constraint they're worried about, the reasoning they want to borrow — has been skipped in favour of the part the agency is proudest of.
Write it in the order above and the page reads as an account of a problem being solved. Write it work-first and it reads as a receipt.
The decision section nobody writes
Here's the part that separates a case study from a brochure: show the reasoning, including what you argued against.
A concrete version. A client comes in wanting a full rebuild with a headless CMS because a competitor has one. You look at the site, find that the content changes four times a year and the actual problem is a booking flow that drops people on mobile. You recommend fixing the flow and leaving the CMS alone, which is a smaller invoice than the one they walked in expecting.
Write that down. The prospect reading it learns something no screenshot can tell them — that you'll say no to work, and that your recommendation tracks the problem rather than the budget. That single paragraph does more for a $30,000 enquiry than an entire gallery.
The same applies to trade-offs you lost. The deadline that forced a phased launch. The photography budget that didn't exist, so the design was built to work with what the client had. Buyers know projects have compromises. A case study with no compromises in it reads as marketing; a case study with two reads as a report.
The most persuasive sentence in a case study is usually the one where you explain what you talked the client out of.
There's a limit. This isn't the place to relitigate a difficult client relationship or imply the client was the problem. The tone that works is matter-of-fact.
Outcomes: what you can claim, and what you can't
The outcome section is where most case studies quietly become fiction. "Conversions up 340%." Up from what, measured how, over what period, compared against which season? Nobody says, because nobody checked.
The rule we hold ourselves to, and the one worth adopting: a number goes on the page only if the client gave it to you from their own reporting, agreed in writing to it being published, and it comes with a baseline and a timeframe. A sentence shaped like "direct bookings went from 18% of total to 31% over the first six months post-launch, per the client's channel report" is a claim you can defend. "Bookings tripled" is not.
This is not only an ethics question. Under Australian Consumer Law, businesses can't make false or misleading representations about their services, and testimonials or performance claims that can't be substantiated fall squarely inside that — the ACCC has been consistently active on fabricated and unsubstantiated claims. A made-up percentage on a case study page is a small lie with a real legal edge.
If you can't get numbers — and often you can't, because the client doesn't measure, or won't share, or the data is genuinely too noisy to attribute — write the honest qualitative outcome instead. What stopped happening. What the team no longer does manually. What the sales conversation sounds like now. A line like "the office manager was spending roughly half a day a week re-typing enquiry emails into the practice system; that's gone" is more persuasive than an invented statistic, because it's the kind of detail nobody bothers to fabricate.
One industry-specific trap worth flagging. If your client is a regulated health provider — physio, psychology, dental, cosmetic — AHPRA's advertising rules prohibit testimonials about regulated health services in advertising. That applies to the practice's site, and it constrains what you can publish about them too. You can write a case study about a clinic's booking system, their operational problem and the build. You cannot include a patient saying the treatment was wonderful, and you shouldn't be advising those clients to collect them either. We've covered what social proof is allowed to look like in more detail, because the rules differ sharply by sector.
One deep case study beats nine tiles
The logo grid persists because it's easy to maintain and feels like scale. It converts badly for a simple reason: it gives every visitor the same shallow signal and asks them to imagine themselves into it.
A single 1,200-word case study with a real situation, a named constraint and a defensible outcome gives one segment of your market a complete answer. Nine tiles give everyone a partial one. If you only have capacity to produce work-page content once a quarter, produce one deep page, not nine shallow ones.
There's a search argument too. A case study written properly is a page about a specific problem in a specific industry, which means it can rank for the long-tail queries buyers actually type when they're mid-problem. A tile in a grid has no text, no query it could plausibly match, and no reason for anyone to link to it.
The other half of this is craft. Case study pages are image-heavy by nature, and an image-heavy page built carelessly is a slow page — which matters most on mobile, where the majority of this traffic sits. Properly sized responsive images, modern formats, lazy loading below the fold: web.dev documents the whole set. We've written separately about project galleries that read like a phone dump, which is the same failure showing up as a visual problem rather than a structural one.
The twenty-minute client interview
You cannot write a good case study from your own project notes. Your notes record what you did. The case study needs what the client felt, which only they have.
Book twenty minutes, record the call with permission, and ask these:
- What was happening in the business that made you go looking? The trigger event, not the general dissatisfaction.
- What had you already tried? Tells you what your prospects have already tried too.
- What were you worried about when you signed? The objection your next prospect is having right now, in the client's own words.
- What was the internal argument? Who had to be convinced, and on what grounds.
- What surprised you during the project? Usually where the honest detail lives.
- What's different now, day to day? Push past "it looks better" — ask what they do differently on a Tuesday.
- Do you have any numbers you'd be comfortable with us publishing? Ask plainly. The worst answer is no.
- What would you say to someone in your position who's hesitating? This becomes the pull quote.
Then send a draft before publishing, ask for written approval of the quote and any figures, and confirm whether they're happy to be named. Twenty minutes of theirs, an afternoon of yours, and you have a page that will keep working for years.
When the client won't be named
Plenty of good projects can't be published under the client's name — confidential briefs, competitive markets, or a client who doesn't want their web budget public.
Anonymous case studies still work, provided they stay specific. "A Melbourne accounting firm, fourteen staff, mid-tier commercial clients" is anonymous and useful. "A leading professional services organisation" is anonymous and worthless — it reads like you invented it, and half the time readers assume you did.
Keep the situation, the constraint, the decision and the honest outcome. Drop the logo, the screenshots that identify them, and any figure they haven't cleared. Say plainly at the top of the page that the client asked not to be named. Buyers read that as discretion, which is a signal in its own right.
Start with the project you just finished
The case study page nobody writes is the one for the project that wrapped last month, while the details are still recoverable and the client still likes you. Six months later, the numbers are buried, the contact has moved on, and the enthusiasm has gone flat.
The practical sequence: pick your most recent completed project, book the twenty-minute call this week, write it to the five-part structure, get the quote and any figures approved in writing, and publish it as its own page with its own URL. Then repeat quarterly. Four solid case studies a year will out-perform a grid of thirty tiles, and you'll never again have to write "trusted by leading brands" on a homepage.
If you'd rather have this built properly into the site rather than bolted onto a page template that was never designed for long-form work, that's part of what a custom build covers — a case study system with proper structure, indexable individual pages, and images that don't cost you three seconds on mobile. Custom projects start at $8,000 AUD.
And if you're not sure whether the work page you've got is helping or just sitting there, book twenty minutes and we'll go through it with you. Bring one recent project you're proud of. Most of the time the story is already there and nobody has written it down.