Cohort targeting
Users sometimes don't see a guide or survey you expected them to receive, or they see it when you thought a cohort exclusion would block them. When your targeting includes a cohort filter, the cause is often cohort sync timing rather than a misconfigured trigger or limit.
Guides and Surveys audience targeting is a subset of Feature Experiment targeting. It supports remote evaluation only and follows the same asynchronous cohort sync rules as Experiment. For a deeper explanation of how cohort targeting works in Experiment, go to Cohort Targeting.This article assumes you know how to configure targeting in the builder. If you need a refresher, review Setup and Targeting before proceeding.How cohort targeting works in Guides and Surveys
When you add a cohort to a targeting segment, Guides and Surveys evaluates the user through remote evaluation. Amplitude syncs the cohort to the Amplitude Experiment destination and checks whether the user belongs to that cohort at evaluation time.
Guides and Surveys doesn't support local evaluation cohort targeting. You can't download cohort membership to the SDK for client-side evaluation the way server-side Experiment SDKs do.Dynamic cohorts don't update in real time. Amplitude syncs them to the Experiment destination on an hourly schedule by default. If you target a cohort of users who performed a Sign Up event, a user who just signed up might not match the cohort until the next sync completes.
Large cohorts can also take longer on the initial sync or after a significant membership change. The first sync of a 10-million-user cohort takes longer than later incremental syncs.
Remote evaluation supports cohorts up to 10 million users.
Troubleshooting checklist
Use these questions when a user doesn't match your cohort-based targeting as expected.
Does the cohort use the Amplitude Experiment sync destination?
For remote evaluation cohort targeting, the cohort must sync to the Amplitude Experiment destination. Open the cohort in Analytics and confirm the sync destination includes Amplitude Experiment. Without this destination, Guides and Surveys can't evaluate cohort membership during remote evaluation.
Is the user in the cohort right now?
Open the cohort in Analytics and confirm the user appears in the current membership. If the user isn't in the cohort yet, the guide or survey won't target them regardless of other settings.
Did the user join the cohort after the last sync?
Dynamic cohort membership updates asynchronously. If the user entered the cohort recently, check the cohort's sync history. The user might have been added in a sync that completed after the evaluation occurred. Wait for the next hourly sync and test again.
Are you targeting new users with a cohort?
Cohort membership updates with a delay, so a cohort-based new-user filter doesn't match users on their first visit. Use the built-in First Seen property instead when you need real-time new-user targeting. Go to First Seen timestamp for setup details.Is the user's identity consistent across sessions?
If a user's identity changes between visits, Amplitude treats them as a different user for targeting purposes. This affects cohort membership the same way it affects limits and history. To learn how Amplitude resolves user identity across devices and sessions, go to Track unique users. For identity troubleshooting steps, go to Guide or Survey seen too often.Did a share link or force trigger bypass targeting?
Direct share links override audience targeting, including cohort filters. SDK triggers, CTAs, and urgent-priority experiences can also display a guide outside normal targeting rules.Situations that look like cohort targeting failures
Some expected behaviors can feel like cohort targeting isn't working.
The user just qualified for the cohort
Behavioral cohorts update on a schedule, not at the moment the qualifying event fires. A user who completes the action that defines the cohort might not match the filter until the next sync. This is expected behavior for dynamic cohorts in remote evaluation.
If you need immediate targeting based on a user property or event, pass the property explicitly through remote evaluation or use a real-time property like First Seen instead of a cohort.
The cohort membership changed significantly
Adding or removing a large number of users from a cohort can extend the next sync duration. Users at the edge of membership changes might evaluate against stale membership until the sync finishes.
You're excluding a cohort
Excluding a cohort works the same way as including one, with the same sync delays. If a user recently left an excluded cohort, they might remain excluded until the next sync updates membership on the Experiment destination.
Learn more in Feature Experiment
Guides and Surveys reuses Experiment's remote evaluation pipeline for cohort targeting. For full details on sync destinations, supported cohort sizes, and advanced troubleshooting, go to:
Was this helpful?