Metrics That Matter for Design Decisions
Not every design decision needs a metric. But the ones that do need the right one. A guide to pairing decisions with evidence.
Not every decision needs a metric
There is a culture in product teams that every decision must be backed by a metric. This is wrong. Some decisions are matters of craft, taste, or brand. Forcing a metric onto them produces false precision and slow decisions. The question is not whether to use metrics, but when.
A design decision needs a metric when it changes user behavior in a way that affects the business. If the decision is about button color, it probably does not need a metric. If it is about removing a step from onboarding, it probably does. The test is: will this change what users do, and does what they do matter to the business?
Choosing the right metric
The right metric is the one that directly reflects the behavior you are trying to change. If you simplified checkout, the metric is checkout completion rate, not page views. If you improved search, the metric is search success rate, not time on page. The wrong metric is one that moves for reasons unrelated to your change. Page views can go up because people are lost, not because they are engaged.
Beware vanity metrics. Total signups, page views, downloads — these go up over time regardless of design quality. They tell you the product is growing, not that the design is working. The metrics that matter are rates and ratios: completion rate, retention rate, conversion rate. These isolate the effect of the design from the effect of growth.
When metrics mislead
A metric can improve while the product gets worse. If you make a feature harder to find, fewer people use it, and the support tickets for that feature drop. The metric improved. The product got worse. Always pair quantitative metrics with qualitative signal. Talk to users. Watch sessions. Read feedback. The metric tells you what happened; the qualitative signal tells you why.
