What is progressive disclosure — and why a hidden button makes an interface clearer

When a screen feels crowded, the first instinct is "let's fit it all in, but prettier". Here's the twist: removing half the buttons from the first screen doesn't make the interface poorer — it makes it clearer. And not because there's more whitespace.

Jakob Nielsen described the mechanism back in 2006: the very fact that something sits on the initial screen tells people it's important. Your first screen isn't a showcase of everything you can do. It's your statement about priorities.

The technique is called progressive disclosure.

Progressive disclosure in one sentence

Show a few of the most-needed options first. Everything else appears on request, behind one clearly labelled control.

The classic example is the print dialog. Copies and printer are right there. Scaling, reverse page order and the rest of the exotica live behind "Advanced". Most people never go there — and that's a designed outcome, not a lost feature.

You see it every day: "Advanced settings" in a deploy form, "Show parameters" next to a prompt box, a collapsed shipping block in a cart.

What it buys you (and why it isn't a matter of taste)

By Nielsen's account, the technique improves three of usability's five components: learnability, efficiency of use, and error rate.

• Beginners get fewer places to go wrong, and don't burn time studying features they don't need.
• Experienced users get speed too: no scanning a long list every time to reach the two settings they actually use.

The standard objection goes: "if you hide features, people won't understand the system as a whole." Nielsen answers directly — research says those are groundless worries. People understand a system better when you help them prioritise.

Two conditions, or it breaks

Nielsen names exactly two — and both are homework, not styling.

1. The right split. Everything frequently needed stays up top. If people regularly dig into the second level, you split it wrong. You don't settle that by taste: use frequency of use, analytics, or watching two real people.

2. An obvious way down. The control must honestly promise what's behind it. "Advanced" is a weak promise — it could be anything. "Shipping settings" is a strong one. That's exactly the work we covered in microcopy: what decides the outcome isn't collapsing things, it's the label on the door.

One more limit: more than two levels is almost always bad. If you need a third, the problem isn't disclosure — it's that you have too many features.

How do you find the split with no analytics yet? Take your last ten real usage scenarios — real ones, not imagined — and mark which settings came up. Anything that never appeared goes to the second level. It's a rough measure, but it beats intuition: a designer's intuition overrates rare features, because they remember how much work went into them.

Staged disclosure is a different thing

There's a relative: staged disclosure — a multi-step wizard showing part of the fields at each step. Two differences:

• With progressive disclosure most people never go to the second level; it holds the rare stuff. With staged disclosure everyone walks through every step, or the task isn't done.
• Progressive is a hierarchy (down and back). Staged is a line (step after step).

Nielsen describes a real case from usability-testing 46 web applications: a hotel site squeezed the whole reservation onto one screen. For comparing rooms and dates that worked well — everything side by side. But address and card fields on that same screen got in the way: while you're still choosing, those fields are simply noise. The right answer wasn't one screen or five. It was two.

Hence the practical rule: split by the shape of the task, not by field count.

Where you'll hit this tomorrow

The moment you build your first settings screen or a longer form — especially one an AI generated for you. A model happily dumps every parameter at once: it has no idea which ones matter once a year. Your job is to keep three or four frequent fields on top and move the rest behind a clear control. The same principle scales up: an MVP is basically progressive disclosure stretched across releases.

There's also an inverse mistake nobody warns you about: hiding what a person needs right now. Price, cancellation terms, delivery time — those aren't "advanced". When the important stuff moves down a level for the sake of a clean first screen, the technique stops being a technique and turns into a dark pattern. The difference is simple: hide what's rare and you're helping; hide what's inconvenient and you're deceiving.

How many levels are acceptable?

Two. Frequent on the first, rare on the second. Designs with three or more, in Nielsen's observation, tend to have poor usability: people lose track of where to look.

Won't a hidden feature get lost?

It will, if the entrance is labelled vaguely. A feature exists for a person exactly as much as the label leading to it is clear. So write the label first, then hide things behind it.

Is an accordion progressive disclosure?

Often yes: collapsed FAQ or settings blocks are the classic case. But if people have to open every block in turn to find the one they need, the technique has turned into extra work — a plain list is better then.