Design trend lists have a built-in problem: by the time a trend is common enough to be listed, adopting it makes you late rather than current. And most of them are fashion — they will look dated in three years exactly the way skeuomorphism and then flat design did.
So this splits the current list into two piles: the changes that are solving a real problem and will stay, and the ones that are decoration and will pass.
Likely to stay
- Dark mode as a first-class design, not an inverted one. Teams have stopped flipping the palette and started selecting dark values deliberately. It stays because people genuinely use it and because the tooling now makes it cheap.
- Type-led layouts. Big, well-set text carrying the page instead of stock photography. Fast, accessible, and it survives translation into a language where words are 40% longer.
- Larger touch targets everywhere. Desktop interfaces adopting phone-sized hit areas. Nobody misses the 11px close button.
- Reduced-motion respect. Animation that checks the system setting before running. This is now an expectation rather than a nicety.
- Content-first spacing systems. Consistent scales rather than eyeballed margins — the thing that keeps 200 screens looking like one product.
Likely to pass
| Trend | Why it will date |
|---|---|
| Heavy glassmorphism | Contrast is fragile; it fails at the edges of every background |
| Oversized cursor effects | Desktop-only, and it interferes with the cursor’s actual job |
| Scroll-jacking narratives | Takes control from the reader; consistently disliked in testing |
| Brutalist deliberate ugliness | Fashion cycle, already turning |
| AI-generated hero illustration | Already reads as “we had no photo” |
| Micro-interactions on everything | Delight on the tenth use is friction on the hundredth |
The test that sorts the two piles
Ask: does this make the interface easier to use on the hundredth visit?
Almost everything in the first list does. Almost nothing in the second does — micro-interactions are delightful the first time and slow the tenth; scroll-jacking is impressive once and infuriating thereafter.
Designers evaluate on first impression because that is how portfolios are judged. Users live on the hundredth visit. The gap between those two vantage points explains most bad trends.
The one worth adopting this year
If you take one thing: design your dark mode deliberately. Not an inversion — a selected set of values, checked against a dark surface.
Inverted palettes fail predictably: pure white text on pure black vibrates, shadows disappear so hierarchy collapses, and brand colours that passed contrast on white often fail on dark. Each one needs its own step chosen for the dark surface.
It is a day of work for a small product and it removes an entire category of complaint.
What never trends
- A visible focus indicator
- Error messages that say what to do next
- Forms that keep your input when validation fails
- Buttons labelled with the action they perform
- Loading states on anything slower than a second
None of these will appear on a trend list, because they were true a decade ago and will be true a decade from now. They are also, reliably, the things users actually notice. The full set is in six design principles.
Common questions
Should we redesign to look current?
Only if the current design is causing measurable problems. Regular users have built a map of your interface; a redesign discards it, and the cost of that is real and immediate while the benefit is speculative.
How do I keep a design from dating?
Lean on typography, spacing and hierarchy rather than effects. Effects date; proportion does not. A design built on a good type scale still looks reasonable in ten years.
Do trends affect performance?
Substantially, and the second pile is the expensive one. We put actual byte costs on the popular ones in trends that make sites faster.
Where should I look for what is next?
Browser release notes, not design galleries. What ships as a native CSS or HTML feature is what becomes normal two years later — and it arrives with the performance and accessibility work already done.
Leave a Reply