Which way this is heading
In & Out
Down: giant headlines, from 25% of 2025's winners to 15% of 2026's. Up: true depth, 33% to 41%. Every movement is measured two independent ways and marked with which ones agree — plus the 3 traits that have not moved at all, which is the half nobody publishes.
On the way out
- giant headlines 25→15%
- thumbnail grids 14→5%
- many colours at once 28→21%
- flat poster layouts 10→3%
On the way in
- true depth 33→41%
- a guided sequence 22→29%
- a real footer 6→11%
- edge to edge 5→9%
Share of award-winning sites doing each, 2025 to 2026.
The short answer
Posters Are Going Out. Pages Are Coming Back.
A year ago the confident thing to do was set the headline enormous, take almost everything else off the screen, and let the words and the colour do the work. That is going away fast. Sites using giant headlines as the main event dropped from 25 in every 100 winners to 15 in twelve months — the biggest single change on the board. Going with it: grids of thumbnails to browse, flat poster-style layouts, and pages using several strong colours at once.
What is arriving is much more ordinary. The fastest riser is the footer — the strip at the very bottom of a page, now actually designed instead of dumped there. Next to it: pictures running right to the edges of the screen with no border, and pages that walk you through something in order.
Put simply: a site used to be something you stood in front of and admired. It is going back to being something you walk through — a way in at the top, something to follow, a way onward at the bottom.
On the way out: winners carrying typography, down from 25% to 15%.
On the way in: winners carrying 3d, up from 33% to 41%.
2025 against 2026
What Moved
The hollow dot is where each thing stood last year, the solid dot where it stands now, and the line between them is the size of the change. Everything sits on one scale, so a long line is a big move wherever it appears.
Going out
Coming in
2025· 2026, falling· 2026, rising·share of winners carrying each
The half that usually gets left out
What Did Not Change
Winners carrying interaction design, one of the decisions that did not move at all this year.
Read a list of changes on its own and it looks like everything is changing. It is not. These things appear on more than one winner in ten and moved less than four points all year. This is the settled part — and the reason most work does not need pulling apart every season.
Half of all winners have things moving on the page, and that has been true for two years running. If something is on this list, it is not a risk and it is not an edge. It is just the floor — do it, and expect no credit for it.
Safe to act on
The Ones That Are Really Going
A single year can flatter a wobble into a trend. These six survived a second, independent check over a much shorter and more recent stretch of winners — so they are the ones worth briefing against rather than watching:
- Outheadlines set enormous, so the words are the picture25% to 15% over the year, and going the same way in the newest winners too
- Outa grid of thumbnails to browse14% to 5% over the year, and going the same way in the newest winners too
- Inthe page walks you through something, in order22% to 29% over the year, and going the same way in the newest winners too
- Outflat, poster-like layouts — print, put on a screen10% to 3% over the year, and going the same way in the newest winners too
- Inthe strip at the very bottom, actually designed6% to 11% over the year, and going the same way in the newest winners too
- Inimages running right to the edges, no border5% to 9% over the year, and going the same way in the newest winners too
Retiring any of those costs nothing and buys a page that does not look like last season. None of them is load-bearing: a gallery wall, a sliding background and a menu you have to solve are all decoration that a visitor has to work around.
3 have stopped mid-move: objects with real depth, that you can turn or move around, several strong colours at once and things move. It climbed 8 points over the year, 33% to 41%; It dropped 7 points over the year, 28% to 21%; It dropped 5 points over the year, 53% to 48%, and then went flat in recent weeks. Pausing or peaked — nobody can tell yet. Build with them if the work wants them, but do not chase them as a rising thing.
What to do about it
If You Are Starting Something
Kode Sports Club8.30
Lando Norris8.18
The New Mobile Workforce8.17
Nike Reactor8.17
Into the Storm8.15
Winners carrying 3d, the decision rising fastest — and the one worth spending on next.
The one line worth taking away: if a job is being briefed off last year's examples — giant headlines, flat colour, a grid of thumbnails — it will already look a season old by the time it goes live.
Drop the fake things before the real ones. A background that slides slower than the text to fake depth, a wall of thumbnails, a menu you have to figure out — all three are on the way out, and none of them was doing much work. Put what you save into the dull parts that are on the way up: a plain menu that says where things are, a footer that is actually useful, pictures that go right to the edge, and one strong colour against a plain background rather than five colours competing.
And leave alone anything on the "did not change" list. Movement, empty space, depth and long scrolling pages are all staying exactly where they are.
The shape of it
One Layer Changed
Read the two lists together and they are not a season's fashions turning over. They are all the same kind of decision: the frame around the work. How you get in, how you move through, where the edges are, what happens at the bottom.
Nothing on the falling list was doing structural work. A gallery grid, a sliding background, large type as the main event — those are ways of presenting a page, and the field is presenting pages differently. What the pages contain is untouched: the same amount of movement, the same depth, the same appetite for telling a story.
Which is why this is cheaper to act on than it looks. Almost none of it reaches the part of a project that costs money to change.
Award winners only, 171 from 2025 against 157 from 2026, with a second reading over the 120 most recent against the 592 before them. Figures for frameworks and libraries are deliberately absent — studios nearly tripled how fully they list them over this span, so those move with the entry form rather than the web. One award, leaning towards agency, campaign and launch work; the labels are the studios' own; and a year is two points, enough to say what moved and not that it keeps moving.