MRR

An annual plan sold in January collects 12 months of cash in January and contributes one twelfth of it to each month’s MRR. That single convention is what makes the metric useful, and it is also why MRR and the bank balance disagree all year.

The normalization forces a set of choices, and the choices are where MRR goes wrong. A subscriber on a discounted first year can be counted at the discounted rate or the list rate. Store commission can be netted out or left in, and the answer changes the figure by double digits. A subscriber in billing retry is either still active revenue or already gone. Free trials that have not converted are usually excluded, though not always. None of these has one correct answer, but a business that answers them differently in two quarters has a trend line that describes its own accounting rather than its subscribers, which is the most common way this metric misleads the people relying on it.

Where MRR earns its place is in decomposition. The month’s change breaks into new subscriptions, upgrades, downgrades, and churn, and those four move for different reasons and respond to different work. A flat month hiding strong new business against heavy churn is a retention brief. The same flat month with weak acquisition and no churn is an acquisition brief. Reading the decomposition next to what those subscribers were shown, per platform, is the version that connects the revenue line to something a team can change, and it is why subscription-aware reporting in Insights sits beside the journey rather than in a separate finance tool. Averaged flat across the base, MRR also hides the plan mix, which is what ARPPU is for.