Skip to main content

How to choose components

Check that a component means what your content says

Users work out what a component means before they read the words in it. A warning icon means "there is a risk". A green panel means "you're done".

When your content backs that up, users know what they’re looking at, and they trust that component the next time they see it. When it does not, you tell the user something that is not true. Users learn what components mean from every service they use, so a warning used where there is no risk makes the next real warning easier to ignore.

5 steps to choose a component

  1. Turn the data into information

    Decide what the user needs to know before you choose anything. A caseworker needs a total and a payment date, not every part of an award.

    Designing data interfaces

  2. Name what the user is doing

    Describe the moment from the user's side, such as "taking a risky action", not what the page shows.

    Find the task in check a choice

  3. Pick a component and check what it means

    Start with the component you think fits, then check that your content backs up what it means. Warning text means "there's a risk to act on", so the content inside it needs to be a risk the user can do something about.

    What a component means

  4. Ask what happens if the user misses it

    What is at stake is minor, moderate or serious, depending on whether the user can put it right. Count harm to other people too, such as the claimant a caseworker is acting for.

    If the user misses it

  5. Check how much attention it draws

    The more that is at stake, the more attention the content needs. Users learn to ignore components that always stand out, and miss content that is too quiet.

    How much attention it draws

A suggested pairing passes two checks: the component means what your content says, and it draws the attention the content needs.

Examples of choosing a component

See the steps worked through on an example decision, and the mistakes they help you avoid.

Where this guidance comes from

We work with teams across DWP and see how components are used in prototypes and services. This guidance brings together what we’ve learned. One pattern stands out. Components with one clear job, such as panel, timeline and quick reference, tend to keep their meaning across services. Flexible ones, such as inset text, details and warning text, are more likely to lose it.

The guidance also draws on established research. The idea that a design means something before anyone reads the words comes from speech act theory, and from the way users build a mental model of how things work. Jakob’s Law explains why that meaning carries from one service to the next. Research on selective attention and the Von Restorff effect shaped the attention check. The minor, moderate and serious levels come from the severity scales used in risk management.

It’s a starting point, not a set of rules. We expect it to change as we learn more, and we’d like it to grow from your evidence too.

Something missing or does not belong here?

Tell us what you have found when choosing components, or anything this guidance does not cover. Let us know.