Field note · November 2025
When the tutorial is the leak
Product tours are easy to instrument and easy to celebrate. They produce a completion event that looks like learning. In Onboarding Drop-Off Analytics they are often the opposite: a corridor that delays first value until the user has already spent their patience.
Completion is not comprehension
If your activation definition includes “finished the tour,” you have built a metric that rewards sitting still. Interview overlays from our Qualitative Drop-Off Interviews intensive keep turning up the same sentence: “I clicked through so it would stop.” That is not activation. That is hostage negotiation with a tooltip.
Tours hide the real first value
In a consumer wallet case we documented on the voices page, tutorial completers were simply people with fewer device constraints. Funding — the thing finance cared about — tracked a permission prompt, not the tour. The tour was not harmless decoration. It occupied the first session window in which the permission explainer should have lived.
A quieter alternative
Put one real task in the first ten minutes. Export a list. Add a beneficiary. Take a photo that actually proceeds. Offer help beside that task, not before it. Instrument the task attempt and the task success as separate events. If you still want a tour, make it skippable on the first screen and do not count it toward activation.
When a tour is still justified
High-stakes consoles with irreversible actions sometimes need a forced orientation. Even then, the tour should end on a successful dry-run of the dangerous action, not on a carousel of feature names. If you cannot dry-run, you probably need a sandbox, not a slideshow.
We teach this restraint in Experiment Design for Onboarding: do not A/B the tour colour while the tour itself is the leak. Kill or postpone the tour, then measure first value again. The chart may look worse. That is information, not failure.