A “paid customer” can include multiple unique users, so comparing MAU to accounts isn’t apples-to-apples.
Peak-month MAU
Figma defines quarterly MAU as the single month in the quarter with the highest usage (i.e. a peak number, not an average) so it can overstate engagement.
Cohort filter
Figma calculates retention using only Paid Customers with > $10K ARR, dropping smaller, higher-churn accounts and inflating stickiness.
The $10K filter for NDR is typical of companies who have a large self serve motion. They are trying to show real companies spending a meaningful amount and remove anyone paying for their own license on plastic
It does include downsell already. As customers who may have downsold are still active, their metric will reflect the downsell. What bothers me is that this is effectively a Net Expansion metric. Customers who have churned completely in the past 12 months don’t show up in the metric. If they lost, say Walmart, 2 months ago with $500k ACV, you would not see it in this metric…
Some caveats:
Paid customers vs. users
A “paid customer” can include multiple unique users, so comparing MAU to accounts isn’t apples-to-apples.
Peak-month MAU
Figma defines quarterly MAU as the single month in the quarter with the highest usage (i.e. a peak number, not an average) so it can overstate engagement.
Cohort filter
Figma calculates retention using only Paid Customers with > $10K ARR, dropping smaller, higher-churn accounts and inflating stickiness.
Good stuff.
The $10K filter for NDR is typical of companies who have a large self serve motion. They are trying to show real companies spending a meaningful amount and remove anyone paying for their own license on plastic
And everyone overstates MAU lol
Is this Net Dollar Retention calculation common? Looks like they are excluding churned customers…
It's not 100% normal but at least they clearly defined it for us. It should include both churn and downsell to be completely accurate
It does include downsell already. As customers who may have downsold are still active, their metric will reflect the downsell. What bothers me is that this is effectively a Net Expansion metric. Customers who have churned completely in the past 12 months don’t show up in the metric. If they lost, say Walmart, 2 months ago with $500k ACV, you would not see it in this metric…
Nice write up! Their measurement of retention is misleading.
“Survivor Bias” as Dave Kellogg calls it.