How do you keep something powerful approachable?
I spent years being told that good design means less on the screen, and the best interface I ever worked on was one of the densest things I have seen.
Every few years somebody shows me a redesign of a professional tool where the main achievement is that there is less on the screen.
It photographs beautifully. It goes down badly with the people who use the thing for eight hours a day, and the reason is not nostalgia.
Who is density actually for?
The person who is still there at four in the afternoon.
I spent six years on Eikon at Thomson Reuters, designing across desktop, tablets and wearables for people who read a financial terminal the way the rest of us read a sentence. The screen was crowded by any magazine standard. It was also the most fluent interface I have worked on, because the density was doing work.
Everything visible at once is something the user does not have to remember, look for, or navigate to. Take it off the screen and it does not disappear. It moves into the person's head, or into a click they now make four hundred times a day.
This is the trade nobody states when they present the cleaner version. We did not remove complexity. We relocated it, from a place where it costs pixels to a place where it costs attention, and we did it on behalf of people who were not in the room.
Several years of HR and talent software taught me the same lesson from the other direction. The systems that decide who gets hired, onboarded and promoted are enormously powerful and are used by people who were given no training and have four other jobs. Stripping those interfaces does not make them kind. It makes them mysterious.
What does "simplify" usually mean in practice?
Hiding. Almost always hiding.
Behind a hover. Behind a menu. Behind a disclosure triangle, a modal, a second screen, an ellipsis nobody notices. The information is still required, so the design has not removed the requirement. It has made the requirement conditional on an interaction.
And every hiding mechanism has a bill attached, paid by somebody specific. Hover excludes anyone on a touchscreen and anyone using a keyboard. Content behind interaction is invisible to a person scanning a page with a screen reader unless someone thought carefully about announcing it, and mostly nobody did. A modal takes a person who was orienting themselves and moves them somewhere else.
So "we simplified it" often means "we made it harder to find, for everyone, and impossible to find, for some."
I am not arguing for clutter. I am arguing that removal is a decision with a victim, and we have a professional habit of presenting it as a decision without one.
Where the craft question and the access question turn out to be the same
Here is the thing that reorganized how I think about this.
An expert reading a dense screen is navigating by position. They do not read the label; they go to the place where the number lives. That is real skill and it is worth designing for.
Now put that same screen in front of someone magnifying it four hundred percent, or someone on a phone between shifts. The layout rearranges, and every piece of positional knowledge they built is gone. They are not reading a denser version of the interface. They are reading a different interface, one nobody designed, made of the same parts in an order determined by a stylesheet.
Which means the question that makes a dense tool humane and the question that makes it usable rearranged are one question: does the meaning survive being reordered?
If the answer is yes, your density is structural and it will hold up for the expert, the newcomer and the person at four hundred percent. If the answer is no, your density was a picture, and it only ever worked for people who could see all of it at once.
What I ask instead of "can we simplify this?"
- What does this person already know? An interface for a recruiter who does this daily and an interface for a manager who does it twice a year are not the same product wearing different themes.
- What did we move, and to where? If it left the screen, name the place it went. Memory, a click, a second visit, a support call.
- Does each region stand on its own? A panel that makes sense only in its position is a panel that stops making sense the moment anything reflows.
- Is anything only discoverable by pointing at it? Hover-only content is a decision about who gets the information, and it is rarely made deliberately.
- What happens to the longest realistic value? Not the mock. A name in full, a job title from a hospital, a department that has been renamed twice.
- Who reads this on a phone, standing up, in bad light? In HR software the answer is most of the workforce, and it is almost never the person the mock was drawn for.
What I have stopped believing
That approachable and powerful sit at opposite ends of one slider, and our craft is choosing where to put it.
They are not on the same axis. A thing is approachable when its structure is legible, which is a property you can design into something dense. It becomes unapproachable when its structure is arbitrary, and you can achieve that with a very clean screen indeed.
The next time a redesign arrives with less on it, ask where the missing parts went and who is carrying them now. If nobody can answer, what you are looking at is not a simplification. It is a transfer, and somebody downstream has already been billed.