Why a Visual Refresh Didn't Move the Metric
Your team spent a month redesigning a SaaS product's dashboard: new typography, a tighter color system, better spacing, a cleaner component library. Stakeholders love how it looks. But the metric that triggered the project — the percentage of new users who reach their "first successful action" within a session — didn't move at all after launch.
- Using the UX/UI distinction, explain why a successful UI overhaul can leave a UX problem completely untouched.
- What information would you want, that a purely visual redesign process wouldn't have surfaced, to diagnose the real problem?
- Name two categories of underlying issue (not visual ones) that commonly cause this exact pattern — activation metric flat despite a UI overhaul.
- What would you change about how this project was scoped from the start to avoid this outcome next time?
1. Why UI polish doesn't touch the UX problem
UI is the visual/tactile surface; the activation metric is a function of the whole experience — information architecture (can they find the thing that leads to success?), interaction design (does the flow match their mental model of the task?), content (does the copy tell them what to do next?), and onboarding sequencing. A redesign that only changes typography, color, and spacing changes none of those. If the actual blocker is "new users don't understand what the primary action even is because the IA buries it three clicks deep," making that same buried structure prettier does not un-bury it. The team solved the surface the interviewer's "UX is research, UI is visuals" answer would call "UI," while the metric depended on the parts of UX that live above the visual layer entirely.
2. What information was missing
A purely visual redesign process typically skips Discover-phase research: session recordings or funnel analytics showing exactly where new users drop off before reaching the first successful action, support tickets or onboarding survey responses describing confusion, and a task-based usability test of the old flow with new users to see where they actually got stuck (as opposed to where the team assumed they got stuck). Without that, the redesign optimized for what stakeholders could see and critique (does it look modern) rather than what was actually broken (can a new user find and complete the task).
3. Two likely underlying issue categories
- Information architecture / navigation: the entry point to the first successful action isn't discoverable — it's mislabeled, buried in a menu, or requires a setup step the user doesn't understand is required first.
- Content and onboarding sequencing: the product never tells the user what the first successful action is or why it matters, so even users who could technically find it don't know to look — this is a UX-writing/onboarding gap, not a visual one.
4. How to scope it differently
Start the project with a Discover phase scoped to the actual metric: pull the funnel data to find exactly where drop-off happens, run 4-5 moderated sessions watching new users attempt the task, and write a "how might we" problem statement grounded in that evidence before any visual work begins. If the diagnosis turns out to be genuinely visual (e.g., the button that triggers the action is visually indistinguishable from decorative elements), a visual redesign is the right fix — but that should be a finding, not an assumption made before Discover.
Share this question