A test to apply before adopting anything
Design trends circulate faster than evidence about them. Before adopting one, ask which specific user task it makes faster, clearer or less error-prone. If the answer is only that it looks current, it is decoration — which is fine, as long as it is a deliberate choice and not mistaken for a usability improvement.
The trends below are sorted by that test rather than by how often they appear in design roundups.
Worth adopting
Progressive disclosure over long forms
Breaking a long form into a short first step and revealing the rest afterwards consistently improves completion, because the initial commitment looks small. It works particularly well for quote requests, where the first step can be one question about what the person needs.
The failure mode is hiding something the user needs in order to decide, such as price or required documents. Disclose progressively; do not conceal.
Genuine dark mode, not an inverted palette
Users increasingly expect a dark option, and on OLED screens it measurably reduces battery drain. Doing it properly means designing a second palette with its own contrast relationships, not applying a filter.
Pure black backgrounds with pure white text cause halation and are harder to read than a dark grey with slightly softened text. Reduce saturation on accent colours for the dark palette, since colours that read well on white tend to vibrate on black.
Accessibility treated as a requirement
This is not a trend so much as an expectation that has finally arrived, helped along by regulation. It is also the cheapest quality improvement available, because most of it is discipline rather than effort.
- Contrast ratios that meet WCAG AA — 4.5:1 for body text
- Every interactive element reachable and operable by keyboard, with a visible focus ring
- Form fields with real labels, not placeholder text standing in for one
- Motion that respects
prefers-reduced-motion - Alt text that describes the image rather than repeating the keyword
Speed treated as a design decision
The most consequential UX decision on most projects is how much the interface weighs. An elegant design that takes six seconds to appear on a mid-range phone loses to a plainer one that appears in one. Designers who specify image budgets alongside spacing scales produce better outcomes than those who hand off without them.
Adopt with care
AI-assisted personalisation
Adapting content to behaviour can help, particularly in commerce where recommendations reduce search effort. It can also produce a confusing interface where two users see different things and neither can find what they saw yesterday.
Keep navigation stable and personalise content within it. Personalising the navigation itself makes an interface impossible to learn.
Conversational and voice interfaces
Useful for narrow, well-defined tasks — checking an order, booking a slot. Poor as a replacement for navigation, because users cannot tell what the system is capable of. Offer conversation as a shortcut alongside a conventional interface, not instead of one.
Bento grids and heavy scroll animation
Bento layouts organise a dense homepage well and have earned their popularity. Scroll-triggered animation is more mixed: a single well-timed reveal draws attention, while a page where every element flies in slows reading and irritates returning visitors. If you use it, keep it subtle and honour reduced-motion preferences.
Mostly decoration
Several widely-featured trends are aesthetic choices with no usability benefit, and some carry a cost.
- Glassmorphism — attractive, but translucent panels routinely fail contrast requirements over busy backgrounds
- Brutalist and anti-design layouts — strong for portfolios, actively harmful for anything transactional
- Custom cursors — they break expected behaviour and add nothing
- 3D hero scenes — heavy, slow on mobile, and rarely explain the product better than a screenshot would
- Infinite scroll on content sites — it removes the footer and makes returning to a position impossible
What has not changed
Underneath the trend cycle the fundamentals move very slowly. Clear hierarchy, generous spacing, legible type, obvious affordances, honest labelling and fast loading were the right answers ten years ago and still are.
A useful sanity check on any redesign: can a first-time visitor tell what you do, whether it applies to them, and how to start, within about five seconds? Most sites that fail are not failing on trends.
Planning a redesign?
Our UI/UX engagements start with the tasks your users actually need to complete, then design around those — wireframes and prototypes before any visual polish.