Global control group
The global control group receives no marketing messages for a fixed percentage of users. Comparing these users' conversions with everyone else's measures the total impact of your messaging. Decision record:…
The global control group sends no marketing messages to a fixed percentage of users. Comparing these users' conversions with everyone else's
measures the total impact of your messaging. Decision record: docs/adr/0060-genel-kontrol-grubu.md.
When it applies
| Send | Control group |
|---|---|
| Campaign and journey messages | Not received |
| In-app messages | Not shown |
| API and transactional sends, test sends, Live Activity | Not affected |
| Web pop-ups | Not affected |
If the Exempt from the global control group option is turned on for a campaign, journey or in-app message, the group receives that message too. Use this option for service announcements.
Campaign deliveries to the control group are recorded with status suppressed and reason global_control.
Setup
Dashboard → App settings → Global control group, or the public API:
curl -X PUT https://api.pushizo.example/v1/apps/$APP_ID/global-control \
-H "Authorization: Bearer $API_KEY" -H "Content-Type: application/json" \
-d '{"enabled":true,"percent":5}'- The percentage is an integer between 1 and 50. For small audiences, a low percentage does not give meaningful results.
- All of a user's devices are in the same group. Membership is computed without storing records, so new users are included automatically.
- When the percentage increases, existing members stay members; when it decreases, some of them leave the group.
- Enabling, changing the percentage, and resetting restart the measurement.
Report
curl "https://api.pushizo.example/v1/apps/$APP_ID/global-control/report?days=30&event=purchase" \
-H "Authorization: Bearer $API_KEY"If event is not given, events marked as conversions in the event definitions are used. For each group, the report gives the users, the users who converted at least once, and the rate;
it also gives the lift percentage, estimated additional conversions, the significance result and the number of messages not sent. The result is a sample comparison.
A/B testing and holdout (Phase 12)
Compare content variants in campaign and journey sends, measure incremental impact with a control (holdout) group, and pick the winner automatically or manually. Design decisions: ADR-0019. Shared code:…
Feature flags
Turn a feature in your app on and off without shipping a new release, and roll it out gradually to a percentage of users. Text, number or JSON values also serve as remote configuration. Decision record:…