Metric Value Statistics Aggregated by User

Applicable Scenarios

Scenarios where you first aggregate a metric value per user (or other analysis subject), then run statistics over that metric value:

• The 75th percentile of each user's daily usage duration, and the average of that percentile over the past 7 days
• The average / maximum / minimum / median / variance / standard deviation of each user's daily payment amount
• The sum / distinct count / average per user of a numeric property per user per day

Hard constraint: this method is implemented via the tag model + virtual event properties. DO NOT compute everything in one shot with pure SQL (SQL breaks daily dynamic matching and historical version semantics).

Supported Aggregation Calibers (metric tag)

Values for definition_request.metric.aggregation in Step 1: