I'm exploring using Firebase's "User in random percentile" condition to target different segments of a customer-base for A/B testing and have a few questions after reviewing the Firebase Remote Config Parameters and Conditions documentation:
Each app instance is persistently mapped to a random whole or fractional number, according to a key defined in that project.
Assuming that like
setUserProperty
this value persists for all generated percentiles (i.e. those without keys and with keys) until the app in uninstalled. Is this a correct assumption?If a condition maps to a random percentile that is <= 50 and I then change that condition to only target <= 10 (but do not modify the key associated with the random percentile), I expect the 11-50 percentile to now match on the next condition (could be the default) but the original 0 to 10 percentile should still match on the condition. Are the these assumptions correct?
Once a random percentile is created for a user, does that value persist across any mutations to conditions so long as the key for that percentile does not change? For example, for key
"firebase_test"
user A's percentile value is 47%. So long as user A does not uninstall the app that user's percentile value will always be 47% for"firebase_test"
.Is the FirebaseRemoteConfig singleton doing any persistence of values? I'm assuming a fetch after the cache expires will ignore any values returned previously by
getValue
methods and simply return the latest value configured.Is there any way to print out the percentile value a user is assigned for given keys (and no key as well)? For example calling
mFirebaseRemoteConfig.getPercentile("firebase_test")
would return1.234
.
Thank you!