Test an onboarding change before you ship it to everyone
Find out whether a shorter onboarding flow gets more people through, measured against success criteria set before the first user sees it.
See the record →Use cases · Product teams
Decide whether to commit a quarter to a roadmap item
Before two engineers spend a quarter on it, find out whether the roadmap item would fix the problem it was proposed for.
The situation
Sales says the missing Salesforce integration is costing deals. Building it would take two engineers a full quarter. If deals are lost for another reason, the quarter is gone and the real problem is still there.
What you find out
- Whether the item would fix the problem it was proposed for, before engineering starts.
- Criteria that product, sales and engineering agreed on up front, so the debate ends on a result.
- Why the item was dropped, in a record anyone can pull up next quarter.
Decision Recordworked example
- assumption
We lose deals mainly because we don’t integrate with Salesforce.
- experiment
Review the last 30 lost deals for the reason given, and add a “Connect Salesforce” button in settings that opens a waitlist. Three weeks.
- success criteria · locked
Salesforce is named in at least 8 of 30 lost deals, and at least 15% of active admins click the button. Locked before the review starts.
- evidence
Named in 3 of 30 lost deals. 6% of admins clicked.not met
decision
Drop it
The quarter goes to the next item on the roadmap, and the reason stays on record for the next time the integration comes up.
Head of product decides; sales and engineering review the criteria; agent drafts the experiment.