Hello! Our team has 2 questions, hope you can help us 🙏
1. How does the conversion window work when we:
a. Have an SQL query for the first 7-day engagement rate which calculates how many users were in the app in the first 7 days after creation. We do that calculation inside the query. Should we use the conversion window of 0? If we set the conversion to, for example, 24 hours, would it extend our calculation 1 day more meaning that we would actually calculate the engagement of the first 8 days?
2. How does the PHASE filter work?
a. We have a case: engagement of first 7 days metric and the timestamp for this metric is first_event date, but the denominator for this metric is All Experiment Users. In this case, we have 2 different timestamps (first_event_date and experiment users date). How does the filter in GB work? Does it filter both of the timestamps or only one of them?
10/30/2023, 12:43 PM
@blue-sundown-32899 has marked the issue as closed. Reopen Issue button
11/01/2023, 7:12 AM
Hello, pinging in case it was forgotten 🙏
11/01/2023, 7:31 AM
1. the conversion windows are used for the gap between the experiment exposure event, and the metric event (or at least the datetime in the metric query). So if your metric is not related to the experiment exposure event, you can leave the initial delay to 0, and the conversion window to something high. If you do want to see how they use the app within the next 7 days after exposure to the experiment, you can set this to 7 *24.
2. Phase filters apply a date range. So we’ll ignore all experiment exposure events that happened before that time window.