Active Days
IntermediateActive days is the number of distinct days on which a user interacted with the product at least once within a defined window; it measures habit, not volume.
Active days answers how many days a user came back, not how many actions they took; ten events in one day still count as a single active day. Its value is that, unlike a raw event total, it is not inflated by one heavy session and it reflects the quality of a habit. The common trap is ignoring the window length: 5 active days in a week and 5 active days in a month are two very different behaviors, so the number only means anything relative to its time window. Read over a fixed 28-day window this same metric becomes the basis of the L28 metric and the power-user curve, and it is its distribution — not its average — that separates casual users from the core.
How it is calculated
Active days = count of distinct calendar days in the window on which the user had at least one interaction
The output is a count, not a percentage, and has no denominator — but it only means anything relative to the window length. To turn it into a rate, divide by the total days in the window to get the share of active days.