When an interface works, nobody notices it. People get done what they came for and think about something else. That's the whole point.
It sounds simple. It's the sum of hundreds of small choices.
Details are decisions
Which button matters most? What goes at the top? What can wait until the next step? Each of these choices decides what people do, and in what order. If nobody makes the choice, the interface makes it for you. Usually badly.
We use four moves more than any others:
- Hierarchy. The most important thing is biggest and comes first. People skim before they read.
- Grouping. Things that belong together sit together. Space separates better than lines.
- Contrast. Text that's meant to be read has to be readable. Grey on grey is decoration.
- One main action. Every screen has one thing people most often need to do. That gets the main button. The rest steps back.
Switch them on one at a time in the graphic at the top, and watch what happens.
Design with real content
Placeholder text lies. Real names are long, real addresses run to three lines, and error messages arrive at bad moments. We design with the content as it will actually be, early, so we see what breaks while it's still cheap to change.
The flow matters more than the screen
A screen can be pretty and still be in the way. What people experience is the path through: what happens when they tap, what they see when something goes wrong, and where they end up afterwards. We draw the flow before we draw the screens.
Prototypes before pixels
We test structure and flow with simple prototypes before we polish the details. Moving a box in a sketch is cheaper than moving a finished component in code. Five users reveal most of the big problems.
A design system is an agreement
When many teams build at the same time, you need shared building blocks. Buttons then behave the same everywhere, and people only have to learn them once. Designers get time for the problems that are actually new.
Nobody notices good design. Everyone notices bad design.