What a winning award entry looks like from our side
The difference between entries that score well and entries that don't is rarely the product. It's the evidence. What editors open first, what makes them raise a score, and the five mistakes that cost the most points.
Two entries arrive for products of roughly equal quality. One scores 7.2 and is Featured. The other scores 8.6 and takes Editors' Choice. Nine times out of ten the difference isn't the product — it's that the second entry made it possible to see the product properly, and the first one didn't.
This is what that difference looks like from the inside of a scorecard.
What an editor does first
Not read your case study. An editor opens the product. They spend fifteen to thirty minutes in it, cold, in the way a new user would — and the first three minutes set the tone for every score that follows. The written material is read afterwards, mostly to answer questions the product raised.
- 01
Open the URL
Cold, no context, often on a phone first.
- 02
Try the main job
Sign up, find the thing, complete a task, hit an edge.
- 03
Probe the edges
Keyboard only. Resize. Break something on purpose.
- 04
Read the entry
Which now either explains what they saw — or contradicts it.
That order matters for a practical reason: nothing you write can rescue a product an editor couldn't get into. Broken sign-up, a region lock, a demo that requires a sales call — these don't produce low scores, they produce entries that can't be scored fairly at all.
The five things that raise scores most
- A task that finishes. The most reliable score-lifter in the programme is a first-time user completing the main job without help. Everything else is an argument; this is a demonstration.
- Edge states that are designed, not defaulted. An empty state that teaches, an error that says what to do, a loading state that sets an expectation. These are where craft is visible and where most products give up.
- Keyboard access that actually works. Tab through the product. If focus is visible, ordered and never trapped, you've quietly earned points across two criteria.
- Honest, specific numbers. "Support tickets about billing fell 41% in the quarter after launch" tells us you measured something. "Users love it" tells them nothing.
- Disclosed gaps. Saying "the date picker is not yet screen-reader complete, here's the plan" raises trust in everything else you claim. Editors find these anyway; the difference is who tells them.
The five that cost the most
- Product unreachable5No access, no login, region-locked. The entry can't be scored.
- Screens only4Eight hero shots, no states, no real data.
- Unverifiable claims3Percentages with no method attached.
- Marketing case study3About the company, not the problem.
- Overclaimed accessibility2"Full WCAG AA" that fails a keyboard check in the first minute.
The case study, in three honest paragraphs
Challenge, approach, outcomes. The best ones are short, specific, and written as if explaining the work to a peer who is about to be critical.
Renewal was a phone call. Eighty per cent of customers renewed, but forty per cent of those calls were people who simply couldn't find the button. We made the renewal state the first thing on the account page, with the date in plain language rather than a timestamp.
Forty words, one decision, one reason. We can go and check it in the product — which is exactly what they'll do next. Compare that with "we reimagined the customer lifecycle through a holistic, user-centred lens," which can't be checked at all, and therefore can't earn a point.
Evidence, and what counts as it
Different criteria are persuaded by different things. You rarely have all of this — send what you have, and don't manufacture the rest.
| Compared on | Strongest evidence | Also useful |
|---|---|---|
| Usability | Task completion or error rates, with n and method | Session recordings, support-ticket categories |
| Information architecture | Tree test or card sort results | Search logs, navigation depth data |
| Interaction | A short screen recording of the key flow | State inventory, motion specs |
| Performance | Field Core Web Vitals, not lab scores | Cold-start times, payload budgets |
| Accessibility | Third-party audit with findings and fixes | Statement with disclosed gaps, AT test notes |
| Learnability | Time-to-first-value, measured | Onboarding completion, what you changed after watching |
A test you can run this afternoon
Give your entry to someone in the company who has never used the product. Ask them to do three things: open the URL and try the main task from the screens you've chosen, read the case study, then tell you what the product is for and who it's for.
If they can't answer, an editor won't be able to either — and unlike your colleague, an editor is scoring seven criteria and moving on to the next entry. Twenty minutes of that test is worth more than another week of polish.
The twelve-minute pass an editor makes
Different editors work differently, but most settle into a rhythm like this. Knowing it tells you what to put in front of them.
0–2 min
First impression
Landing, sign-up or first screen. Is it obvious what this is and what to do?
2–8 min
The main task
Attempted the way a new user would, without reading anything.
8–11 min
The edges
Keyboard, resize, an error forced on purpose, a look at the settings.
11–12 min
Sanity checks
Does the case study match what just happened?
Screens: a set that works, in order
Eight screens is the limit and most entries waste them on variations of the same view. A set that earns its space looks like this.
- Cover — the product at its most characteristic, at real density. Not a device mockup floating in gradient.
- The main task, mid-flow — the screen where the work actually happens.
- An empty state — a genuinely new account.
- An error, handled well — validation, a failed save, a dropped connection.
- A dense view — real data volume, long strings, edge-case names.
- A narrow viewport — if anyone uses the product on one.
- A settings or permissions screen — where complexity collects.
- A before-and-after — if you rebuilt something, show what it replaced.
Words that raise scores, and words that don't
| Compared on | Earns a point | Earns nothing |
|---|---|---|
| About the problem | Account managers exported to a spreadsheet every morning | Users wanted a better experience |
| About the decision | We rejected filters; the default view was the problem | We took a user-centred approach |
| About the outcome | Time to first invoice: 11 min → 4 min | Engagement improved dramatically |
| About the gaps | The date picker is not yet screen-reader complete; fix due in Q1 | Fully accessible |
Questions we get asked
Does a longer case study score better?
No. Two or three tight paragraphs per section beat a page of context. Editors read the case study after using the product, mostly to answer questions the product raised.
What if we have no quantitative data?
Say so and send qualitative evidence instead — what people said in testing, what stopped being escalated, what you changed after watching someone struggle. An honest absence costs almost nothing; an invented percentage costs the entry.
Should we hide the parts we're not proud of?
No, and you can't — editors find them in the first ten minutes. Disclosing a known gap with a plan raises trust in every other claim you make.
Is a video worth including?
A short screen recording of the key flow is one of the most useful things you can send, particularly for interaction-heavy work. A produced brand film is not.
How much difference does the entry itself make?
Enough to change a level. The product sets the ceiling; the entry decides how much of that ceiling we can see.
- entry craft
- evidence
- case study
Written by
The UXE editors
The team that selects the work, uses it, scores the seven criteria and writes every verdict on this site.
Keep reading
Judging & criteria8 min read
How to submit a UX design award entry (and what happens after you do)
A step-by-step guide to submitting a UX design award entry: what to prepare, which categories to choose, what we actually look at, and what happens between sending it and hearing back.
The UXE editors
Craft & practice5 min read
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 editors
Judging & criteria6 min read
Digital product design awards: a guide for SaaS and product teams
Why most digital product design awards are built for marketing sites, what a product team should look for instead, and how to enter a SaaS product so we see the work that actually took the year.
The UXE editors