Most subscription businesses track MRR as a single number. That’s the mistake. The intelligence isn’t in the total, it’s in what’s moving underneath it.
Signal
Revenue looks healthy on paper, but you are constantly surprised by churn & struggling to forecast.
Stakeholders
Founders, CFOs, and commercial leads at subscription businesses who need to move from reporting revenue to predicting it.
Strategy
Decompose MRR into its component movements, track cohort-level behaviour over time.
What MRR Actually Measures
Monthly Recurring Revenue is a point-in-time figure: the normalised monthly value of all active subscriptions right now. It is not recognised revenue. It is not cash collected. It is the run rate, what the business would generate next month if nothing changed.
That predictability is the whole point. One-time revenue tells you what happened. MRR tells you what’s coming. But only if you decompose it properly.
Every MRR figure is the result of four distinct movements from the prior period. New MRR is the monthly value added from newly acquired customers. Expansion MRR is the incremental value from existing customers upgrading or adding seats. Contraction MRR is the value lost from customers who downgrade or remove seats but remain active. Churned MRR is the value lost from customers who cancel entirely.
Beginning MRR, plus new, plus expansion, minus contraction, minus churned, equals ending MRR. A single growth number obscures all of this. A waterfall shows you the whole story.
Why Each Movement Matters Differently
New MRR reveals the health of your acquisition engine. Consistent growth signals effective go-to-market. Volatility, particularly dependency on large deals, creates planning problems that compound downstream.
Expansion MRR is arguably the most valuable component because it compounds from customers you’ve already paid to acquire. High expansion indicates that customers are finding growing value in the product. Timing matters here: businesses where expansion happens within 90 days of acquisition can invest more aggressively in new customer acquisition because payback periods are short.
Contraction MRR is the early warning signal most businesses ignore. Customers reducing spend whilst remaining active are often signalling disengagement before cancellation. Some contraction is healthy; customers rightsizing to appropriate plans, but concentrated contraction in specific segments or following specific triggers almost always means something needs fixing.
Churned MRR is the most straightforward to measure but the most consequential to misunderstand. A 5% annual churn rate does not sound alarming until you calculate what it costs in replacement acquisition spend. Because different churn has different causes; early churn usually means onboarding or acquisition-fit failures, late churn usually means competitive displacement or product drift, aggregate churn rates without cohort segmentation are nearly useless for diagnosing root causes.
Expert help is only a call away. We are always happy to give advice, offer an impartial opinion and put you on the right track. Book a call with a member of our friendly team today.
MRR & Customer Value
Net Revenue Retention (the percentage of MRR retained and expanded from a cohort over a period) has become the single most important subscription metric precisely because it captures expansion, contraction, and churn in one figure. NRR above 100% means your existing customers are growing in value faster than you’re losing it. That creates compounding dynamics where growth accelerates without proportional increases in acquisition spend.
The shape of cohort curves over time reveals whether you have a structural problem or a tactical one. If churn concentrates in the first few months and then drops sharply, you have an onboarding problem but a fundamentally strong product. If churn remains elevated even for customers who’ve been with you for years, the issue is the product itself or how you’re positioning it.
Segmentation transforms this analysis. Enterprise MRR growing 30% through expansion whilst SMB MRR declines 15% through churn isn’t a 7.5% net growth story it’s two entirely different businesses that happen to share a P&L. The strategic question becomes which one you’re actually building.
Common Reporting Errors
The most consequential mistake is conflating MRR with recognised revenue. They measure different things. Recognised revenue accounts for partial-month contributions from customers who joined or left mid-period. MRR at month-end reflects only what’s active at that moment. Mixing the two creates inconsistencies that undermine confidence in every number downstream.
Annual contracts require normalisation. A £2,400 annual contract contributes £200 to MRR, not £2,400 in month one and nothing thereafter. Companies that add the full annual value directly to MRR create artificial spikes in months with large deal closes, followed by apparent stagnation in months without them. That pattern destroys the forecasting value MRR is supposed to provide.
One-time charges (implementation fees, professional services, setup costs) should never enter MRR calculations. Similarly, customers on promotional pricing should have their MRR recorded at what they actually pay, with the uplift captured as expansion MRR when the promotion ends. Recording promotional customers at standard pricing creates fictional MRR and distorts both current performance and future forecasts.
Structuring MRR For Forecasting
Effective forecasting requires MRR data structured at three levels: customer (acquisition date, channel, segment, initial MRR), subscription (plan type, billing period, renewal date), and movement (effective date, type, amount, reason). That movement-level detail is where forecasting models find their signal.
Cohort analysis is the mechanism that converts historical behaviour into forward projections. A cohort acquired 18 months ago with 82% MRR retention and consistent expansion tells you, with reasonable confidence, what a cohort acquired today will likely look like at the same age. Segment-specific cohort analysis is more powerful still, enterprise cohorts and SMB cohorts often require entirely different models.
Leading indicators extend this further. Product usage declining typically precedes churn. Feature adoption typically precedes expansion. Connecting MRR data to these signals enables models that forecast movements before they appear in the MRR waterfall — turning MRR from a reporting tool into an early warning system.
The prerequisite for all of this is consistency. MRR’s forecasting value depends entirely on methodological stability over time. If categorisation rules change month to month, trend analysis becomes impossible and the comparisons that enable good decisions stop working entirely.
Get In Touch
Our friendly team are always on hand to answer questions, troubleshoot problems and point you in the right direction.
