Quick answer
An MRR forecast is useful when you need a planning-grade bridge from today's recurring revenue to a likely future run-rate. It stops being enough when retention behavior changes by cohort, when expansion is uneven, or when you need to separate acquisition growth from existing-customer health.
Why a bridge model is useful
An MRR forecast estimates future recurring revenue by bridging starting MRR with new MRR, expansion, contraction, and churn each month. A bridge model keeps the levers explicit: how much MRR comes from new customers vs how much comes from retaining and expanding existing customers.
Core monthly bridge
MRR_next = MRR + new MRR + expansion - contraction - churn (computed monthly).
How to set inputs (practical defaults)
- Starting MRR: current month recurring run-rate (exclude one-time revenue).
- New MRR: use trailing 3-month average if growth is volatile.
- Expansion/churn rates: start with your trailing monthly revenue retention behavior (or approximate from NRR/GRR).
- Horizon: 6-12 months for execution, 12-24 months for strategy scenarios.
How to interpret results
- Ending MRR and ARR run-rate show where the business lands if assumptions hold.
- CMGR helps compare scenarios (growth rate compounded monthly).
- Implied monthly NRR/GRR reflects existing-customer health independent of new MRR.
When a simple bridge stops being enough
- If expansion and churn differ a lot by cohort or segment, a single blended rate can hide where future MRR is really breaking.
- If the forecast looks good only because new MRR is masking weak existing-customer retention, inspect NRR, GRR, and an MRR waterfall next.
- If planning decisions depend on activation timing, seasonality, or pricing changes, move from a simple bridge into cohort and scenario analysis.
Common mistakes
- Mixing monthly and annual rates (e.g., annual churn used as monthly churn).
- Double counting: including expansions inside 'new MRR' or vice versa.
- Forecasting long horizons without scenarios (small rate changes compound a lot).
- Using this instead of cohort curves when you have meaningful seasonality or changing retention by cohort.