Guided first-bot flow: funnel completion 28.4% to 63.7% (+124%)
Commentary byLiftArchiveOur analysis, not theirsBlip added a step-by-step guided flow to its chatbot builder, and activation funnel completion rose from 28.4% to 63.7%, a 124% increase.
Blip's team built the flow, and it was published as an Appcues customer story.
The numbers compare the period before launch with the two months after, not a split test.
- Funnel completion: 28.4% before, 63.7% after, a 124% relative increase.
- Median time to finish the funnel: 1.1 hours before, 6.8 minutes after.
- Drop-off at the publish step fell from 55% to under 11%.
- Drop-off at the final test step fell from 14% to about 1%.
- Measured within two months of launch. No sample sizes were shared.
The change
A welcome step and a chain of tooltips guide new users to publish and test a first bot.
Simplified mockup drawn by LiftArchive from the shared screenshots.
Control
New users landed in the chatbot builder with lots of features in view and no clear path. Many dropped out at the create, publish and test stages.
Variant
A welcome step set expectations, then tooltips pointed to creating the bot, naming it and adding an image, building the first conversation flow, adding content, publishing and testing. A closing slideout encouraged users to keep exploring.
Why it likely won
Our reading of the result, not the practitioner's claim.
- The team defined activation as publishing and testing a bot, then guided users to exactly those actions.
- On a feature-heavy screen, pointing at the one next button cuts decision time.
- Each tooltip handles one small step, so users are never far from the next instruction.
- The flow covered the stages where the funnel data showed the biggest drop-offs.
Caveats: Before/after comparison over about two months, not a split test, so seasonality or other product changes could contribute. Vendor-published, with no sample sizes.
Who ran it
Blip ran this test and published the results (Appcues customer story). The numbers are theirs; the summary and commentary on this page are ours.
Read the original case study on Blip →
How to run this test on your site
- Hypothesis
- If we guide new users through each step to their first working output with targeted tooltips, then more will activate because they no longer have to work out the path on a busy screen.
- Primary metric
- Share of new users who complete the activation funnel
- Guardrails
- Flow dismissal rate
- Week-two retention
- Quality of what users build (for example, bots that get real traffic)
- Who should test it
- Builder-style or no-code SaaS where value comes from creating and publishing something. It matters less for tools that are useful right after sign-up with no setup.
Questions
How do I choose the activation event for onboarding?
Pick the action that best predicts users staying, often the first time they get real value.
Blip chose publishing and testing a first chatbot.
How many steps can a product tour have?
There is no fixed limit if each step is tied to an action the user takes.
Long tours that only describe features tend to get skipped.
Is a before/after onboarding comparison reliable?
It is a useful signal but weaker than a split test, because other changes over the same period can affect results.
Hold back a control group when you can.