Skip to content
UXE Award

How to read a UX case study from an award winner

Award-winning UX case studies follow a recognisable shape. What to look for when you read one, what to copy, what to ignore, and how to write yours so anyone can verify every claim.

The UXE editors5 min read

UX case study award winners get read for the wrong reasons. People open them looking for a layout to imitate, when the useful part is the reasoning — the decision that was taken, the alternative that was rejected, and the evidence that the choice was right.

Here's the structure that recurs in case studies that score well, and how to use it in both directions: reading other people's, and writing your own.

The shape that keeps appearing

Four moves, in order
  1. 01

    A specific failure

    One task, one group of people, one thing going wrong.

  2. 02

    A decision

    What was changed — and what was considered and rejected.

  3. 03

    A measurement

    What moved, how it was measured, over what period.

  4. 04

    A disclosed gap

    What is still not right, and what the plan is.

The fourth move is the one most teams leave out, and it's the one that makes the other three believable. A case study with no unresolved problems reads like marketing, because real products always have some.

Reading one: five questions to ask

  • Who was failing, and at what? If the problem is described as "users found it confusing," the team hadn't finished diagnosing it when they started designing.
  • What was the alternative? Every real decision had one. A case study that presents the chosen design as inevitable is hiding the interesting part.
  • How was the outcome measured? Look for n, method and time frame. "Conversion up 24%" over one week of a seasonal product is noise wearing a suit.
  • Does the live product still match? Open it. The gap between a case study and a shipped product tells you how much of it survived contact with the roadmap.
  • What did it cost? Rebuilds that took eighteen months and three teams are not transferable lessons for a team of two.

What's worth copying, and what isn't

Two halves of a case study

Copy the method

  • How they framed the problem before designing
  • The research that changed their mind
  • How they measured the outcome
  • How they handled the edge cases

Don't copy the surface

  • The layout, which solved their constraints
  • The tone of voice, which belongs to their brand
  • The navigation pattern, which fits their content
  • Their metric, which may not be your goal
Interface patterns are cheap to copy and rarely transfer. Reasoning transfers to every project you'll ever work on.

Writing yours: the three fields, and what they're for

Our entry form asks for Challenge, Approach and Outcomes. The same three questions sit behind most award entries, whatever they're called.

Challenge — the problem, not the brief

A brief is what the business asked for. A challenge is what was actually wrong. "Redesign the dashboard" is a brief. "Account managers were exporting to a spreadsheet every morning because the dashboard couldn't answer their first question" is a challenge — and it tells us exactly what to test.

Approach — one decision, with its alternative

Don't list what you shipped. Name the decision that mattered, the option you rejected, and why. "We considered adding filters; we found the real problem was that the default view answered nobody's question, so we made the default configurable and left filters for later." That is a sentence an editor can go and verify in ninety seconds.

Outcomes — a number with a method behind it

The same claim, twice
Compared onVerifiableNot verifiable
Completion71% → 89%, unmoderated, n=600"Much easier to use"
SupportBilling tickets −41% over one quarter"Fewer complaints"
SpeedLCP 3.8s → 1.9s, field data, p75"Significantly faster"
AdoptionTime to first invoice 11m → 4m"Engagement up 300%"

If you don't have numbers, say so and give qualitative evidence instead: what people said in testing, what the support queue looked like, what stopped being escalated. An honest absence of data costs almost nothing. An invented percentage costs the entry.

The published ones are the best reference

Every recognized project here has its case study published next to all seven of its scores — which means you can read the work and then see how it was actually rated, criterion by criterion. That combination is rare, and it's the fastest way to calibrate your own writing.

Browse the showcase, pick a project in your field, and read its case study against its scores. Then write yours the way you wish theirs had been written, and enter it.

The five case studies worth your time

Rather than reading twenty, read five properly — and read them against the live product, which is where most of them stop being true.

  • One in your industry, to calibrate what counts as hard in your field.
  • One in a regulated field — health, finance, public services — where constraints are visible and the design has to work anyway.
  • One about a rebuild, where you can compare the before and after.
  • One about accessibility, because that's where reasoning is usually most explicit.
  • One that scored badly on a criterion you care about, which teaches more than another winner.

Reading a scorecard next to a case study

This is the part that is unusual here and the reason it's worth doing. Open a published profile and read the case study, then look at the seven criterion scores underneath it. The gap between the story and the scores is where the learning is.

Four patterns, and what to take from each
Compared onWhat it means
Strong story, strong scoresThe claims are real. Copy the method.
Strong story, weak usabilityThe team designed a vision and shipped something else.
Weak story, strong scoresGood product, badly presented. Very common.
High on six, low on oneUsually accessibility or performance — the two nobody owns.

Questions we get asked

  • How long should a UX case study be?

    For an award entry, two to three short paragraphs per section. For a portfolio, as long as it takes to make one decision legible — usually a lot less than people write.

  • What should a case study include if the project failed?

    The same four moves: the problem, the decision, what actually happened, and what you'd do differently. A well-reasoned failure is more readable than a vague success, and most senior interviewers prefer it.

  • Can I write a case study about work under NDA?

    Often, with the client's permission and with numbers expressed as relative change rather than absolutes. For an award entry you need permission to publish, so ask before you enter.

  • Do editors read the whole case study?

    They read it after using the product, and they read it looking for answers to specific questions. Front-load the specifics; nothing is gained by building to a conclusion.

  • case study
  • writing
  • evidence

Written by

The UXE editors

The team that selects the work, uses it, scores the seven criteria and writes every verdict on this site.