The Tooltip Tour
A sequence of overlaid callouts pointing at interface elements, and why it is usually a patch.
The problem it solves
A screen has grown past the point where it explains itself, and there is no budget to redesign it. A tour is bolted on to narrate the confusion.
Attention across tour steps
Start with the uncomfortable part. A tooltip tour is almost always a symptom. Something on the screen failed to communicate, and instead of fixing the label, the layout or the default, the team added a layer of text on top that points at it. The tour ships in a week. The underlying screen stays broken for years.
That does not make tours useless. It makes them a last resort with a narrow brief. If you are going to build one, build it knowing what it is covering for.
The diagnostic question
For each step in your planned tour, ask: could this be solved by changing the thing the tooltip points at. A tooltip that says This is where your projects live can be replaced by a heading that says Projects. A tooltip that says Click here to invite your team can be replaced by a button that says Invite your team. Work through the list honestly and most tours shrink to one or two steps, which is the right size.
Tour as a patch
- This is your workspace switcher.
- Here you can see all your items.
- Click the plus button to add something.
- Use filters to narrow down results.
- You can access settings from here.
Fix the screen instead
- Show the workspace name in the switcher rather than an icon.
- Name the list with a heading and an item count.
- Label the button Add task instead of using a bare plus.
- Show the filter bar with the active filter stated in words.
- Nothing. Settings does not need a tour step.
What survives the diagnostic is worth a callout: a genuinely novel interaction with no convention behind it, a destructive control that needs a warning once, or a feature whose value is invisible from its affordance: A visual cue that tells you what an element will do if you act on it.. Three steps is a ceiling.
- 1Cut every step that could be a better label, a heading, a default or a layout change.
- 2Cap what remains at three steps, on one screen.
- 3Make every step dismissible with a visible close control and the Escape key.
- 4Let the user interact with the real interface during the tour rather than blocking it behind an overlay.
- 5Never show it twice, and never show it to users who arrived through an invite from an existing account.
If the interface needs a tour to be understood, the tour is documentation for a bug.
One more trap. Tours are easy to measure and the numbers flatter them, because the people who finish a five step tour are the motivated ones who would have figured it out anyway. Completion rate tells you about the users, not about the tour.
When it fits
- A single genuinely novel interaction with no existing convention, such as a new canvas gesture.
- A one time warning before a destructive or irreversible control.
- A major layout change shown once to existing users who already know the old arrangement.
- A short bridge while a real redesign is in flight, with a date to remove it.
When it backfires
- When it narrates an interface that should have been relabelled, which hides the problem from the people who could fix it.
- When it blocks the interface behind an overlay, so an eager user cannot start doing the thing being described.
- When it runs longer than three steps, at which point people click Next to make it stop and retain nothing.
- When completion rate becomes the success metric, since the users who finish tours were already going to activate.
Products using it
Figma
New canvas interactions are introduced with a brief pointer near the relevant tool while the canvas stays fully usable.
Google Maps
Interface changes have been announced with a single callout on the moved control rather than a multi step walkthrough.
GitHub
Redesigned navigation has been introduced with one dismissible pointer at the changed element, not a tour of the page.
The psychology under it
The takeaway
Run every tour step through the question of whether a better label would do the job, and ship only what survives.
Finished the teardown? Bank it and the day counts toward your run.
