SaaS user reporting
Signups, activation and the steps between them. Funnels keep whole people when data is sampled, so a conversion rate still means what it says.
Signups, activation, and the gap
Signup counts are the number everybody has and the number nobody acts on. The useful report is the step after it: how many of those accounts did the thing your product exists to do, and how long it took them.
Funnels that keep whole people
When data is sampled, sampling is decided per subject rather than per event, using a hash of the subject id. Either all of a person’s events are kept or none are. Sampling events independently is what turns a funnel into nonsense, because a person can then appear at step three and not at step one.
Cohorts by the properties you send
Filter a whole report by plan, source or any property you attach to the events, and the URL carries the filter, so the narrowed view is the thing you paste into a conversation rather than a screenshot of it.
Comparisons that admit their limits
A week-on-week comparison against a week that did not exist says so. For a young product, that is most of the comparisons, and inventing percentages for them is how a launch looks like a rocket.
Start with signups
One event when somebody signs up is enough to get a real report, and the funnel follows once you send the step after it.
Start free