Removing "Experimental" and "Observed" labels from the design system

[martinwake]

We’ve decided to remove the “Experimental” and “Observed” tags from the elements of the DWP Design System. We’re keeping the “Retired” tag so that people know not to use these patterns in new services.

We hope this will make the design system easier to understand and use. Basically: if something is published in the design system, you can use it in your service.

Background

Sometimes it’s clear that just having a standard approach to a common problem would be useful, but there isn’t a lot of research available to help choose one design from a number of options. We originally followed the GOV.UK design system’s practice of tagging these patterns as “Experimental”. We thought it would be useful to know that these things should be used with a bit more care.

Other patterns, though, have been used in several live services with thousands of users. These aren’t really “Experimental”, but would still benefit from coordinated and focused user research specific to that pattern. For these we introduced the “Observed” status.

Why we’ve changed our approach

We’ve had consistent feedback that the existing system isn’t clear about a few things:

• how you should use a pattern differently depending on its status
• when and why a pattern’s status changes from “Experimental” to “Observed”
• what comes after “Observed” - is there a point where we have enough research that we don’t need any more?

We considered adding a third status of “Recommended”: but we would still be left with a lack of clarity about what these statuses meant in practice when using the design system. We decided that as there wasn’t a clear difference in how we wanted people to use things with various statuses, it would be better to remove the labels altogether.

What we’re doing instead

If we don’t think we have strong enough evidence to recommend a pattern for use, we won’t publish it in the design system at all. (In practice this was true for Experimental patterns anyway, but the labels didn’t make that clear.)

We still want to be clear and open about the research basis for our patterns, so that designers can make informed judgments about how and when to use them. So from now on we’ll also document this in the guidance for every new component and pattern, and work our way through existing patterns to do the same.

We hope that by making the evidence and reasoning behind our decisions clear we’ll also open them up to challenge from contributors, which will improve the long-term quality and robustness of the design system as a whole.

Related stories

#350