Quick answer
A product tour is a short guided sequence inside a product that walks a user through a few key steps, using pointers, tooltips or modals. Judge it by how many users complete it and reach the target action, not by how often it is shown.
Formula: Tour completion rate = users who finished the last step ÷ users shown. Step completion = users reaching step N+1 ÷ users reaching step N. Lift = outcome rate (shown) − outcome rate (holdout).
What is a product tour?
A product tour is a short, guided sequence shown inside a product that walks a new or returning user through a few key steps, usually with pointers, tooltips or modal windows. Amplitude defines tours as guided walkthroughs that introduce users to the main features and workflows of an app, most useful during onboarding. Intercom describes its version as an interactive, multi-step guide that shows customers how to use the product.
A tour is a means, not a result. Its job is to move a person toward a moment of value, such as creating the first project or sending the first message. A tour that is shown 10,000 times and completed by nobody has done nothing, however good it looks. That is why the useful question is not "how many people saw it?" but "how many finished it, and did they do the thing it was supposed to lead to?" The sections below turn that into numbers. For the wider design of the first minutes of use, see onboarding flow and app onboarding; the first-use experience itself is covered at FTUE. This page is the definition and how to judge a tour, not a tool comparison.
Types of product tours
Amplitude groups tours into styles that differ in how much they ask of the user:
| Type | What it is | Best used for | Main risk |
|---|---|---|---|
| Interactive walkthrough | A multi-step sequence where the user acts on each step before moving on | Teaching a core workflow to new users | Long sequences lose people at every step |
| Feature callout | A tooltip or pop-up on one new or underused feature | Announcing a feature at the moment it is relevant | Interrupts someone who is mid-task |
| Checklist | A task list users complete in any order and at their own pace | Setup that has several independent parts | Easy to ignore if it is not tied to value |
| Embedded visual | A short image, video or animation that explains a screen and can be dismissed | Explaining something that is hard to describe | Passive; users watch and still do not act |
There are also two delivery styles. A linear tour shows every user the same fixed steps in order. A contextual tour appears based on what the user does or has not done, for example only after they have been on an empty dashboard for a minute. Contextual tours cost more to design but interrupt fewer people.
Metrics: judge a tour by completion and outcome, not views
Amplitude lists completion rate, time to value, a guided versus unguided comparison, and A/B results as the main measures. Written as rules you can compute:
| Metric | Formula | What it tells you |
|---|---|---|
| Start rate | Users who started ÷ users shown | Whether the first card is worth a click |
| Step completion | Users reaching step N+1 ÷ users reaching step N | Which step is losing people |
| Tour completion rate | Users who finished the last step ÷ users shown | Whether the tour is the right length |
| Dismissal rate | Users who closed it early ÷ users shown | How much it annoys |
| Outcome rate | Users who did the target action within a window ÷ users in the group | Whether the tour achieved its purpose |
| Lift | Outcome rate (shown) − outcome rate (not shown) | The effect of the tour, if the groups are comparable |
The last two rows matter most. Completion tells you the tour is bearable; only the outcome and the lift tell you it is worth having.
Worked example: a five-step tour
Sample data: a project tool shows a five-step tour to 1,000 new users. The target outcome is creating a first project within 7 days.
| Step | Users reaching it | Step completion (vs previous step) | Share of those shown |
|---|---|---|---|
| 1. Welcome card | 1,000 | — | 100% |
| 2. Create a project | 720 | 720 ÷ 1,000 = 72.0% | 72.0% |
| 3. Invite a teammate | 540 | 540 ÷ 720 = 75.0% | 54.0% |
| 4. Open the board | 410 | 410 ÷ 540 = 75.9% | 41.0% |
| 5. Finish | 330 | 330 ÷ 410 = 80.5% | 33.0% |
Tour completion is 330 ÷ 1,000 = 33%. The biggest single loss is the very first card (280 of 1,000 never got past it), so the opening, not the length, is the first thing to look at.
Completers do better, but that does not prove the tour works. Of the 330 who finished, 198 created a project within 7 days (60%); of the 670 who did not finish, 134 did (20%). That is a 40-point gap, but people who sit through five steps were probably going to do the task anyway.
A holdout is the fair test. Suppose the team keeps a random 20% of new users from seeing the tour. Sample results:
| Group | Users | Created a project within 7 days | Outcome rate |
|---|---|---|---|
| Shown the tour | 800 | 336 | 336 ÷ 800 = 42.0% |
| Holdout (not shown) | 200 | 72 | 72 ÷ 200 = 36.0% |
The lift is 42.0 − 36.0 = 6.0 points, much smaller than the 40-point completer gap. A two-proportion z-test on these numbers gives z ≈ 1.54, which is a p-value around 0.12, so with a holdout this small the 6 points could still be chance. Larger holdouts, or a longer run, are needed before deciding. See holdout group for how to size one.
Common mistakes
- Reporting impressions. "The tour was shown 50,000 times" says nothing about completion or outcome.
- Comparing completers with non-completers and calling it lift. Selection does the work, as the 60% versus 20% gap shows. Use a random holdout.
- A tour that explains every button. Five short steps toward one outcome beat fifteen steps that cover the whole interface.
- Showing it at the wrong time. A tour on first load, before the user has seen the empty state, answers questions nobody has asked yet.
- Never retiring it. When the interface changes, an old tour points at buttons that moved. Review tours when a release touches the screens they cover.
- No way out. Always give a visible skip or close, and treat dismissal as data, not failure.
Measuring a product tour in TapCub
Take the five-step tour above. The tour itself is built in your own tool or code; TapCub is where you check whether it worked.
- Send the tour's moments as events. From the tour's code, call
tapcub.trackfortour_started,tour_step_viewedwith astepproperty,tour_completedandtour_dismissedwith thestepit was closed on. Add them to All-platform analytics › Data › Tracking plan so the names are agreed before release. - Build the funnel. In All-platform analytics › Behavior › Behavior analysis, make a funnel from
tour_startedthrough each step totour_completed, which reproduces the table above and shows where users drop. - Measure the outcome. Add a second funnel from
tour_startedto your target event, for exampleproject_created, within 7 days, and compare it with the same funnel for users who never started the tour. - Save the groups. In All-platform analytics › Users › Segments & tags, create a rule segment for "completed the tour" and use it as a cohort to compare retention.
- Record the variant. If your tour tool assigns a holdout, send the assignment as an event property such as
tour_variant, then break the outcome down by it.
TapCub does not build product tours and does not randomise users into a tour or a holdout. It only measures. Do the assignment in your tour tool, and read the lift in TapCub.
See also: What is a walkthrough? Meanings, how to run one and how to measure it.
Sources
- Amplitude, Understanding product tours.
- Intercom Help Center, Product tours explained.
Was this explanation clear?
We use your answers to fix and expand this page.
Couldn’t send. Please try again in a minute.
Thanks for the feedback. We’ll use it to improve this page.You’ve already sent feedback on this page. Thanks.

Product Adoption Consultant
Explores customer onboarding, live conversations, and self-service, turning common usage questions into practical guidance and improvements.