One of the toughest things about being someone who cares deeply about design systems is making the case for a dedicated design system. Folks in leadership will often ask you to prove the value of it. Why should we care about good front-end development and consistency? Sure, sure, sure, they say—everyone wants a flashy design system—but is it worth the cost?
That question is tough because developer productivity, front-end quality, and even accessibility to some extent, are all such nebulous things. In contrast, this is one of the smartest things about Google’s Core Web Vitals because it puts a number on the problem and provides very actionable things to do next.
When it comes to design systems, we don’t really have metrics that we can point to and say “Ah, yes, I need to put folks on the design systems team so that we can push our design system up from a bad score of 60/100.” It would be neat if we did, but I don’t think we ever will.
Enter Sparkbox. They wanted to fix this by testing how much faster their eight developers were in a little test. They got their devs to make a form, by hand, and then do it again using IBM’s Carbon design system, which they’d never used before.
The results are super interesting:
Using a design system made a simple form page 47% faster to develop versus coding it from scratch. The median time for the scratch submissions was 4.2 hours compared to the 2 hour median time for Carbon submissions. The Carbon timing included the time the developers spent familiarizing themselves with the design system.
Now imagine if those devs were familiar with Carbon’s design system! If that was the case, I imagine the time to build those forms would be way, way faster than those initial results.
The study methodology seems a bit off. All 8 devs first coded the form from scratch and then coded it again with Carbon. This means any time spent understanding the form design is counted against the scratch trial instead of the Carbon trial.
To the previous comment. The methodology is just fine. It still makes the design system faster. If you remove the time that takes them to learn the carbon system, it’s now less than 2 hours!
I’ve been developing these at various companies for the last seven or so years. For them to work, both dev and design absolutely have to be on board with the approach and familiar with the process.
If the design team has a habit of introducing a special cases, or if dev hasn’t adopted components yet, the first iteration is going to be very rocky and costly.
It seems like Sparkbox had a lot of knowledgeable people on hand from the start, which is lucky. For most orgs, I’d suggest having the design team draft a design system and deliver it as a PDF (something very low tech), and have dev follow it manually on a small project so that everyone understands the subtleties behind keeping the design consistent. Then work up to the “living” version gradually.