First-use completion and the time it takes
Aliases: learnability · first-use experience · onboarding cost
What it is
Learnability's operational definition: can a user with zero experience complete the goal on first use, in how much time, with how much assistance? The core metrics are unassisted first-use completion and first-use task time; supporting metrics are help requests and abandonment points.
Why it happens
First use runs on knowledge users bring, not on in-product training: interfaces that invoke universal conventions (labels are buttons, back is top-left) do well, while anything riding on product-specific conventions (hidden gestures, private vocabulary) forms first-use obstacles. Most first-use time goes to searching for the right entry and trial and error, not execution—so information architecture and label wording affect learnability more than operational complexity does.
Studying it
First-use testing demands genuinely new users: screen out anyone who has touched the product, give no instructions, assign the goal task, and record completion, time, and help-seeking. Watch the control condition—"people who know similar products" are not pure novices. First-use data is noisy; pretest the task design and criteria. Measuring first, second, and third sessions separately yields the learning curve's start and slope.
Where it stops holding
First-use performance does not predict practiced performance: a design smooth on day one can be mediocre at expert speed. First use is also contaminated by setup, registration, and other peripheral flows—outside interface learnability but inside the data. For infrequent products, first-use tests should simulate "return after an interval," which differs from a pure first session.
Applying it
- Run a small-sample first-use test before launch; if unassisted completion misses the bar, fix entries and labels before scaling traffic.
- Inventory every element that depends on a private convention and add first-use-visible hints or an alternative path.
- Monitor real first use with "new users' first-week task completion" and calibrate it against test data.