October 22, 2025/709 words/4 min read

The Shift Design Created

Learning to read an interface changes the way we build products.

I was out for drinks a few days ago with a couple of designer friends, people who say things like "spatial hierarchy" unironically, and at some point the conversation turned to work. They were going back and forth about some layout problem, and I found myself arguing about type weight and where the eye lands on a screen. Someone asked me when I started caring about this stuff and I didn't have an answer. The change had happened in the background without me noticing.

These days I spend a weird amount of time looking at screens I'm not supposed to spend too much time looking at. A restaurant website or banking app is enough to set me off. At 1am I'll still be picking apart a checkout flow when I should be asleep. I can't turn it off! Every interface is a set of choices someone made. Some of those choices are subtly wrong in ways that nobody on the team probably noticed. I've learned to notice, and sometimes I wish I hadn't, because using crappy software is more stressful than it used to be.

Five years ago I didn't see any of this. I looked at interfaces the way I looked at code. I wanted the data to flow correctly through every state, including the ugly edge cases. If it did, we were done and I wanted to ship it. I am a PM who came up through engineering and I thought understanding the machine was understanding the product. I remember sitting in meetings where the only designer at our company would raise flags about spacing and hierarchy, describing the visual weight of the screen in ways I didn't understand. I'd nod along thinking, sure, make it pretty, but the important thing is the logic underneath. I didn't say that out loud but I'm sure the guy could tell.

Then we got a new director of product who'd also been a head of design at his last place, and he had a habit that fascinated me. After product reviews he'd sit with a screen open and look at it without clicking anywhere. Then he'd say things like "this section is competing with that one" or "there's no clear entry point here." At first I thought it was just fancy designer talk. At his suggestion we started watching users navigate through the UI, and I saw exactly what he'd described. Their eyes had no anchor, and they hesitated in the spots he'd flagged. He could read the damn interface like a badly structured document, and he was consistently right about where it failed.

And I could see he was right but I couldn't do what he did. I'd stare at the same screen and just see a screen, so I started bothering him about it, probably more than he wanted. I'd pull him over before reviews and ask him to walk me through what he saw, and I was lucky that he was generous about it. He'd point at something and explain why it was pulling focus, or why two elements next to each other were creating a relationship we didn't intend. Over a few months it started clicking, like when a foreign language starts to make sense after you've been hearing it long enough. One day you understand a sentence without translating it first. I started catching things on my own. A group of actions could look sequential because of how they sat on the screen. Even a label could feel wrong against the elements around it. The layout tells the person using it where to start and what belongs together whether you planned it or not. Get that wrong and they'll feel lost, but often won't care enough to tell you why, or won't be able to. They'll just say "it's confusing" and leave, and you'll look at the ticket and think, but it works. Trust the user, it doesn't!

The work is slower now and I like it more. I spend real time figuring out where a confirmation message belongs even after the flow technically works. Five years ago that would have seemed pointless, but those tiny, boring decisions compound. I can't stop noticing them now, so government portals are ruined for me forever. Small price.