The decision
Move an early behavior, then measure.
Typesy VIP needed a behavior that appeared before durable learning outcomes.
The candidate was simple: start the first course within four days.
That behavior became a leading indicator. It remained separate from long-term retention.
The experiment kept one contrast.
Treatment received the activation intervention. Control retained the existing experience.
Both groups used the same first-course definition and four-day window.
| Group | Users | PAI hits | Hit rate |
|---|---|---|---|
| Treatment | 121 | 84 | 69.4% |
| Control | 57 | 33 | 57.9% |
Treatment showed a 19.9% relative lift and 11.5-point absolute difference.
What the result supports.
The treatment group showed a higher first-course rate in this experiment.
The observation supports further lifecycle testing and retention follow-up.
This public summary does not establish statistical significance or long-term retention lift.
Method notes and limits.
| Outcome | First course started within four days. |
|---|---|
| Published cohort | 178 users: 121 treatment and 57 control. |
| Observed counts | 84 treatment hits and 33 control hits. |
| Source | Typesy VIP experiment outcome supplied for publication. |
| Not public | Dates, assignment method, intervention detail, and inferential analysis. |
The rates and lifts are descriptive until the full analysis is published.
The operating lesson.
A useful PAI is not only predictive. Teams must also be able to move it.
Discovery narrows the behavior. Validation earns the right to operationalize it.