May 19, 2023

Discount design systems…

I'm going to spend the latter half of our 30 days focused on how to build out different types of design systems, but there are a few more things worth talking about before we get there… “measure twice, cut once,” and all that.

…

Almost 35 years ago, Jakob Nielsen made the case for an approach to usability testing he called discount usability engineering.

The video I linked to above is more recent and isn’t very long if you want to go watch it. Here’s the gist:

It’s usually better to start with a greater number of simpler low-fidelity tests with a handful of participants than to wait for one large, exhaustive study with many participants. This approach brings good heuristic insights much more quickly, which means you can iterate faster and test solutions more frequently. In other words, the number of iterations is greater than the quality of iterations.

This approach lowers the cost — for testing as well as for iterating on designs — yet still provides valuable learnings because of the sheer number of tests being run. And simpler testing methods require less experienced people, which again reduces costs and increases the likelihood usability testing will actually happen.

…

I would like to make a similar case for “discount design systems.”

Instead of a starting your DS work with a huge design system effort that has a lengthy development cycle and a big target release months from now, start with several narrowly-scoped releases with aggressive timelines and gather feedback after each release.

For example, instead of spending months building a huge set of complex Figma components to distribute to all the designers in your org, start with something small like a shareable library of styles with your brand’s primary colors or something highly reusable like a global feedback widget that appears on all of your sites.

Instead of spending months pulling all of your common code components out into a separate library and abstracting them so they can be reused across all your platforms, start with 2-3 of the most used or frequently updated. Better yet, offer to help build 2-3 of those components for the new project that has a tight deadline and build them so that they can easily reused in other products.

There are several benefits to this discount approach:

• Fast iterations keep your work highly practical (and less prescriptive)
• Quick wins demonstrate the value of the work you’re doing
• Smaller scope lets you “soft launch” and iron out wrinkles early when risk is low
• Consensus-building efforts are easier with a smaller group of people involved
• Teams that you help can become advocates for your DS initiative

Maybe after talking with your digital product teams you came away with a strong sense that they’d love a huge set of connected design system tools. I still suggest breaking the work up into sprints, or at least multiple phases. This keeps your efforts nimble and makes it much easier to adapt to feedback from your users, org changes, market changes, or changing stakeholder priorities.

I know of several organizations that worked on huge design system efforts for months, only to have the work scrapped before launch because of a rebrand, a reorg, or simply because stakeholders hadn’t really been paying attention.

There may be some rare scenarios where it makes sense to build large at first, but if you're just getting started, consider starting with a discount design systems mindset.

P.S. This approach may not work as well in orgs where some part of a design system already exists. More on that tomorrow.

—
Cheers,

Up Next: When to launch...

← Full Archives