What a new user decides before they read a word of your copy, and why most onboarding work is aimed at the wrong moment.
A first-time user is not reading. They are scanning for a shape they recognise, and deciding whether this product belongs to a category they already understand. That judgement lands in the first few seconds, well before the value proposition has been processed as language.
This is why onboarding copy so often underperforms. It is answering a question the user has not asked yet.
Three things, in order: say what kind of thing this is, show one piece of real state, and offer a single obvious next move. Everything else competes with those three and should be deferred.
Real state matters more than it sounds. An empty dashboard with illustrations reads as a demo. The same dashboard with one seeded row reads as a working tool the user has just joined.
Guided tours are a way of postponing the design problem. They add a layer that explains an interface instead of making the interface explain itself, and they are skipped by most users on the first tap.
Where a tour seems necessary, it is usually pointing at a screen that is trying to do too much. Splitting that screen is cheaper than narrating it.
Not completion of the onboarding flow — that number goes up when you make the flow shorter, whether or not anything improved. Measure the first action that only a committed user takes, and how many days pass before it happens.
On the last three engagements where we made this change, the honest number was lower than the team expected and far more useful than the one they had been reporting.
“What a new user decides before they read a word of your copy, and why most onboarding work is aimed at the wrong moment.”