When a user falls under an active experiment, they can be served one of the pre-defined experiment treatment prices, or a personalized price.
Few things to note about how Corrily treats experimental prices:
- Corrily’s machine learning models determine what price a user gets assigned
- A uniquely identifiable user, whether using an
custom IDis locked to an experiment treatment i.e. as long as Corrily can identify them, they will be served the same experiment treatment.
- Certain experiment prices may be disabled during an active experiment if they are under-performing (when traffic allocation is set to automatic)
- To check which users have been assigned what prices, anyone with
supportor higher privileges on the Corrily dashboard can key in the
User IDon the Users page to see what experiment treatment they were assigned.
- You can edit and override a user’s price here so they no longer see an experimental treatment
- For any user, when you call Corrily with the
override:trueflag, it resets the old price and a new price may be assigned at that point in time. The flag is local to the call i.e. next time you wish to reset a user’s price, you would need to pass the flag again.
- Once an experiment is completed, experimented-on users will continue seeing control or experiment prices. You can move them to a new price using Corrily’s dashboard.