Hello team! We are looking for a feature flag management system to replace our custom one. Can someone please confirm if there's a solution available that has the ability to manage feature flags in groups based on user tier? Specifically, we need a system where users with pricing tier A would have 5 default features enabled, while users with pricing tier B (higher tier) would have 7 features enabled. Additionally, if this grouping is possible, we'd like to know how these groups would behave with experiments, and if it is possible to add overrides? Thanks!

This is possible. How you implement it will depend a bit on your setup. You can setup attributes for each group and change the flag behavior based on it (while allowing for overrides). You can see the targeting docs for more info: https://docs.growthbook.io/features/targeting Saved groups can also help here: https://docs.growthbook.io/features/targeting#saved-groups This behavior extends to experiments, too. I can also say that when we use GrowthBook internally, we often target by organization rather than by user.

Thank you @strong-mouse-55694, Quick follow-up: if we need to set attributes, should we use the Frontend SDK to set them up? Does it have some sort of security measure to prevent users from simply overriding it on the frontend side and receiving higher-tier flags?

Hi. You'll need to pass them in your SDK. Since it's client-side, it's possible for users to edit the javascript to change the value. How big of a concern that is, is up to you. You can also use secure attributes which helps to obscure the values: https://docs.growthbook.io/lib/js#secure-attributes or remote evaluation: https://docs.growthbook.io/lib/js#remote-evaluation (this requires more setup). And, yes, you'd create your feature and the then target by group/attributes.

Hello @strong-mouse-55694, security is a huge concern for us, as I mentioned, we are planning to use pricing tiers with feature flag groups, so this will be sensitive information. Secure attributes are set on an FE, so it would not solve our problem. Remote-evaluation might be the right thing, but I need more info on how this should be implemented ( I don't see a manual on a private server side to do the evaluation). Can you help me with that? Thank you!

For remote eval, you need to: • Update your SDK connection and toggle on Remote Evaluation. • Set up a remote eval server. We provide GrowthBook Proxy to handle this, but you can also use a custom solution Depending on your setup, there are a few other options to consider: if you can implement things server-side, then you can avoid sharing any with the client. You can also check our Edge solution, which renders the experiment variations before they reach the client