There's a classic scene in digital product work: the team presents the new interface in a meeting, everyone applauds, the aesthetics get praise. Three weeks later, support is buried, conversion dropped, and users are complaining that 'the screen got weird'. Design won the meeting and lost the operation.
Pretty and functional aren't enemies, but they're also not the same thing. Confusing them is what produces interfaces that look amazing in a portfolio and frustrate in daily use.
How 'pretty' gets in the way
The symptoms tend to be the same. Too many visual elements pulling focus from the main button, making the user hesitate where they should act. Low contrast, the kind of gray icon on a gray background that looks elegant in Figma and is unreadable on a phone screen in sunlight. Slow microinteractions that delight in a 30-second demo and infuriate when someone has to do that action 40 times a day.
The real user isn't in a sound booth looking at a 4K screen. They're packed into a subway, battery at 12%, burning through their data plan, out of patience. Design that ignores that is design made for the designer, not for the user.
How to decide on an interface based on something beyond taste
Don't start with visuals, start with the task. Before thinking about color or typography, define what action this screen needs to enable and in how few steps. Then use what data you already have: heatmap showing where the cursor gives up, analytics revealing where the funnel breaks, a conversation with five real users. Standardize via design tokens (reusable colors, spacing, typography) so you stop reinventing the wheel on every screen. And test before polishing: an ugly clickable prototype answers more questions than ten beautiful Figmas.
How to know if it actually worked
Three numbers settle most of the doubt. Task completion rate for the key actions (how many finish signup, checkout, upload). Average time to complete critical flows (down? up?). And honest qualitative feedback, an open question in a moderated interview or in-app survey. If any of those three got worse after the redesign, the problem isn't users getting used to it, it's the design decision.
“Good design isn't what wins awards. It's what makes the user not notice it exists.”
Pretty is an ally when it simplifies and amplifies the task. It's an enemy when it becomes the goal itself. Mature teams decide every interface based on data, conversation, and testing, and treat the praise in the meeting as a clue, not as proof. The proof comes from what happens when the product is running on its own.