The journey a new user goes through between signing up and being able to use your product on their own.
In practice
Onboarding is the path, activation is the destination — the two are often wrongly treated as the same thing. The most common mistake is explaining everything at once on the first visit instead of revealing things progressively, right when the user actually needs them. A good onboarding is measured by one question: how much time separates sign-up from the aha moment — the shorter that gap, the better calibrated the path.
The formula
Time-to-value = median time between sign-up and the aha moment (and activation rate = the share who get there at all)
- Median, not mean
A few users who take three months would drag a mean to nonsense. The median tells you what the typical new user lives through — and the 75th percentile tells you what the slow half does.
- Sign-up
The start of the clock is account creation, not first visit: onboarding owns what happens after someone decided to try, not before.
- Aha moment
The end of the clock — the defined action from the Activation stage. An onboarding without a defined destination cannot be measured, only redesigned on taste.
Onboarding is the path; activation is the destination. The two are confused constantly, and the confusion has a cost: teams measure "completed the onboarding flow" (clicked through the tour) instead of "reached the value", and optimise a tutorial nobody needed.
Worked example
Cutting a path in half without adding a feature
- An analytics tool's aha moment: the first chart built on the user's own data. Before: sign-up → 5-step profile form → product tour (9 tooltips) → empty dashboard → "connect a data source" buried in settings. Median time-to-value: 4.5 days. Activation: 22%.
- Change 1: the first screen after sign-up is "connect a source", with a sample dataset one click away for those who can't yet. Change 2: the tour is removed; the two tooltips that mattered move onto the chart builder itself. Change 3: the profile form is asked for later, when the user invites a colleague.
- After one month of cohorts: median time-to-value 38 minutes. Activation: 41%.
- Nothing was added. Three things were removed or moved. The product didn't change; the distance between the door and the value did.
Most onboarding work that pays is subtraction. The question to ask of every step is not "is this useful?" but "does the user need this before the moment, or could it wait?".
Orders of magnitude
- Time-to-value should match the product's cycle: minutes for consumer apps, within the first session for most SaaS, a day or two when real data has to be connected, longer only when a human has to be involved — and then the onboarding's job is to make that human show up fast.
- Drop-off in the first session is the largest single loss in most funnels: it is common for half or more of sign-ups never to return after day one. Every step removed between sign-up and value moves that number.
- Product tours and coach marks are among the most-skipped elements in software; the value of guidance lies in showing up at the moment of need, inside the task, not in a sequence up front.
How to improve it
- Start from the aha moment and count backwards: list every screen between it and sign-up, then justify each one. Cut what can't be justified, defer what can wait.
- Replace empty states with a starting point: sample data, a template, an import. An empty dashboard is the most common place a new user quietly leaves.
- Ask for information when it becomes useful, not at the door. Name, company, role, team size can all wait until the product needs them for something the user wants.
- Watch five new users go through it — recordings or a live session. You will find the step that loses people faster than any funnel chart will show it.
- Iterate on cohorts, one change at a time. "Never touched since launch" is the 0-point answer in the Tour because an onboarding is a hypothesis about your users, and hypotheses age.
In the Tour
Has your onboarding been tested or iterated on at least once?
- Yes, iterated based on real data20 pts
- Tweaked a little, informally7 pts
- Never touched since launch0 pts
"Has your onboarding been tested or iterated on at least once?" is the third Activation question — 20 for iterated on real data, 7 for tweaked a little, informally, 0 for never touched since launch. It sits after the two questions that make iteration possible (a defined moment, a measured share who reach it): an onboarding can only be improved towards a destination someone has named.
Questions people ask
What's the difference between onboarding and activation?
Onboarding is everything you build between sign-up and the first value: screens, emails, defaults, sample data. Activation is whether — and how fast — users reach that value. One is the design, the other is the measurement. Good onboarding is judged only by activation; it has no score of its own.
Should onboarding include a product tour?
Rarely as a sequence up front. Guidance works when it appears at the moment a user is about to need it, inside the task; a tour before the user has any task teaches things they'll have forgotten by the time they matter. If you keep one, make it skippable, three steps at most, and measure whether people who skip it activate less — they usually don't.
How long should onboarding take?
As long as the path to the first value requires, and not a screen more. The right question isn't duration but distance: how many decisions and how much typing stand between the account and the moment. Measure time-to-value, then remove steps until the median stops dropping.