Frequently Asked Questions

Date of slack thread: 4/22/24

Anonymous: We’re looking to do a scheduled rollout of a feature gate against company IDs as the unit ID, so e.g. phase 1 will be x % of company IDs for a feature. Can we easily pull which company IDs are passing for e.g. the first phase? I see the View pulses page, but couldn’t seem to extract this info.

Tore (Statsig): What do you need this for? If you increase the rollout %, only new company IDs will be included, the ones that were passing previously will continue to pass. There is no static list of companyIDs that will pass for a rollout - it’s a deterministic hash of the ID. So given an ID, I can tell you if it passes or not. And given a list of IDs, I can tell you which ones pass. But on its own, Statsig does not track a list of which IDs are in which group in real time. The best you could do is the next day, you can export exposures from the console to get the set of all passing IDs.

Anonymous: Thank you. Is that hash function something you expose or no? The rationale here is that for certain feature rollouts, our customer success teams can have touch points if they’re aware of who’s having it turned on when, if for example we have rollouts over a week. We could certainly create manual buckets, but were wondering if we could still leverage the pass % functionality and get that info.

Tore (Statsig): Are you able to persist it on your side? You could use an evaluation callback to track who has been exposed to which features. Alternatively, if you are okay just knowing which group they would be, you can use any SDK/HTTP API - just make sure you disable exposure logging for that check so it doesn’t record it in any experiment.