Skip to main content
Two Query features go past a single measure function: rolling measures, which compute a measure over a trailing window for every day, and measure filters, which drop rows by their measure value. This page covers both, with formulas built from rolling measures.

​
When to use

​
Rolling measures

A rolling measure computes a measure over the last X days for every day in the date range. It answers questions like “What are my weekly active users on a daily basis?” or “What’s the moving average for each day?”

​
Rolling options

There are three rolling options. Each one takes a number of days: pick a preset, or type your own.
Measure By dropdown for an Average measure with Rolling selected, showing a Rolling Average option set to 7 days and an Apply button next to the list of measure functions.

​
Build a rolling measure

  1. Build your measure in Query.
  2. Open the Measure By dropdown under the selected event. Click Rolling at the top right of the dropdown.
  3. Select a rolling option. Select a preset number of days, or enter your own, and confirm it (the screenshot shows an Apply button).
  4. Click Execute.

​
Compound rolling measures

You can build several rolling measures in one report and combine them in a formula.
Report with Measure A as Unique Users with Rolling Average 7 days and Measure B as Unique Users with Rolling Window 7 days, a formula F1 set to A/B, and the resulting daily ratio plotted as a line.
Formula covers formula syntax and limits.

​
Measure filters

A measure filter keeps or drops rows by the value of a measure. Use it to exclude outliers, or to avoid division by 0 when you break down a ratio.

​
How it works

In this example, a breakdown by status message has a long tail of small values.
Data table of event counts by Status Message, with three large values at the top and a long tail of values from 416 down to 0.
To focus on the top groups, add a measure filter of A > 1000. Under Report Filter, click Add filter > Measure Filter, then enter the condition. The table then keeps only the rows where measure A is over 1000.
Report Filter with a measure filter A > 1000 and the Add filter menu open showing Global Filter, Cohort Filter and Measure Filter, and a data table that now lists only three status messages.

​
MCP prompt examples

“Show me a rolling 7-day average of unique users per day for the last 30 days”
“Create a line chart of daily DAU/WAU: rolling 7-day average of unique users divided by a rolling 7-day window of unique users”
“Show event counts by status message for the last 7 days, only where the count is over 1000”