Why
- The first screen is looked at longest. What’s visible without scrolling gets 57% of viewing time, the first two screens 74% (NN/g, 2018). In 2010 the first screen got 80%: people scroll more now, but the top of the page still decides whether they’ll scroll.
- The decision is quick. In Gitte Lindgaard’s study (2006), ratings of a page shown for 50 ms nearly matched ratings after 500 ms: the impression forms before reading.
- Every extra option costs time to choose. By Hick’s law, choice time grows with the number of options; three equal calls to action on one screen are three “so where do I go?” questions.
Where it’s used
| Where | One idea means | What it looks like |
|---|---|---|
| A page’s first screen | what this is, who it’s for and what’s next | a heading, a 1–2 sentence lead, one main action |
| A slide | one conclusion | the title is a conclusion as a sentence: “Latency grew after the 3.7 update”, with the evidence below |
| A dashboard screen | one problem or one question | “is the service up”: status, latency, errors; disk and backups on the next screen |
| A chat message | one request | the first line is the question or the conclusion; it shows in the notification |
| A section of a page | one reader question | the heading is the point, the first sentence the answer (Answer first) |
A dense screen is one idea too
A one-screen cheat sheet of 30 commands is one idea if every command is about git: people keep it open and search it
by eye, and the density helps. What breaks the rule is something else — two different questions side by side, with no
clue which to start with.