Blog / Build with AI /

Designing settings screens people understand

Most settings screens are a junk drawer. Every time someone says “can we make that configurable?”, another toggle goes in. Eventually you have forty options, no obvious order, and a user who changes nothing because they’re afraid of breaking something.

My argument: a good settings screen is mostly about what you leave out. Fewer options, clearer words and sensible defaults beat a long list of switches, and the designer’s job is to make decisions so the user doesn’t have to.

Reason 1: every setting is a decision you’ve handed to someone else

When you add an option, you’re saying “I couldn’t decide, so you do it”. Sometimes that’s fair. People have different needs, and a dark mode switch or a language choice is a genuine preference.

Often it isn’t. “Enable fast mode” or “Use optimised loading” are options nobody should have to think about. If the fast option is better for nearly everyone, make it the default and don’t ask.

A test: for each setting, name a real person who would choose the non-default value, and say why. If you can’t, the setting probably shouldn’t exist.

Reason 2: people read labels, not manuals

Settings screens get a few seconds of attention. The words have to work on their own.

Bad: “Enable async script deferral for non-critical resources”.
Better: “Load page scripts later so pages appear sooner”.

Write labels that say what happens, in the language of the person using it. Then add a short line of help text underneath for anything that isn’t obvious, including what could go wrong. If you build WordPress plugins or browser extensions, this alone separates the ones that feel friendly from the ones that feel risky.

Other label habits:

  • Phrase toggles consistently. “Show tab colours” (on = shown) is clear. “Disable colours” (on = hidden) makes people do mental backflips.
  • Avoid double negatives. “Don’t hide the toolbar” is a trap.
  • Use verbs on buttons. “Save changes” beats “OK”.
  • Say what happens to existing data before an action that affects it.

Reason 3: grouping tells people where to look

Related settings should sit together under a plain heading: Appearance, Notifications, Data and privacy, Advanced. Order the groups by how often people need them. The thing 90 per cent of users touch goes first; the thing nobody touches goes last or behind an “Advanced” section.

“Advanced” is a perfectly good place to hide rarely used controls. It isn’t a place to hide dangerous ones without a warning. If an option can delete data or break the site, say so in plain words and confirm before acting.

A rough structure for a small tool:

  1. The few settings that change how it looks or behaves day to day.
  2. Account or connection details (API keys, logins).
  3. Data: export, import, reset.
  4. Advanced.

Reason 4: show the effect, not just the switch

If a setting changes how something looks, show a preview. A font size slider with no sample text makes people guess. A colour picker beside a live example makes them decide in a second.

If a setting can’t be previewed, make it easy to undo. A “Reset to defaults” button and a visible “Saved” confirmation reduce fear. Fear is the real enemy of settings pages.

The counter-argument

“But power users want options.” That’s true, and they have a point. Some people love control, and taking options away from them can feel patronising. Developers, editors and administrators often need fine-grained choices.

The answer isn’t no options. It’s layers. Give everyone good defaults, a short list of common settings, and an advanced area for those who want more. If you’ve got a lot of power users, a settings export or a config file may serve them better than a hundred more checkboxes.

The other fair objection: defaults are opinions, and your opinion may be wrong. True, so watch what people actually change. If most users switch the same setting, your default is wrong. Change it.

Accessibility belongs here too

A settings screen is a form, so basic form rules apply: every control needs a visible label, enough contrast, and keyboard access. Don’t rely on colour alone to show on or off. Accessibility checks for small tools is a good short list to run through.

A mini-audit you can do in ten minutes

Open your settings screen and ask:

  1. Which settings has nobody ever asked for?
  2. Which labels would you struggle to explain to a friend?
  3. Could any two settings be merged into one?
  4. What happens if a user presses every toggle? Is anything dangerous?
  5. Is there a way back to the defaults?
  6. Does it work on a phone?

If you’re building with AI, you can paste your settings code in and ask it to review labels and grouping. AI is quite good at suggesting clearer wording. Just don’t let it add options. Ask it to propose which ones to remove.

Settings in a WordPress or Shopify context

If you build themes or plugins, you’ll know the temptation: a customiser with eighty options feels generous. In practice most buyers use a handful of them, and every extra option adds support questions, edge cases and testing work. A dozen well-chosen options usually beat eighty.

For store owners picking tools, the same rule helps in reverse. A tool with a clear, short settings page tends to be easier to trust than one that hands you a cockpit.

Key takeaways

  • Each setting is a decision you hand to the user. Add them reluctantly.
  • Write labels in plain words that describe the effect.
  • Group by how often things are used, and put risky controls behind a warning.
  • Show previews and give people an undo.
  • Layer options for power users rather than piling them up.

Think of a settings page you’ve struggled with. What would you have removed first?