Revenue Metrics
What money you earn, and what you keep
Revenue metrics — revenue, MRR/ARR, net and gross revenue retention, ARPU and bookings — track money earned and kept, and each has traps around timing, upgrades and refunds.
What you will learn
- Define revenue, bookings and recognized revenue distinctly
- Compute MRR, GRR and NRR with correct mechanics
- Avoid the timing and cancellation traps that break revenue dashboards
What is this?
Revenue metrics quantify the money a business earns and retains. Core ones: booked revenue (contracts signed), recognized revenue (earned per accounting rules), recurring revenue (MRR/ARR), revenue retention (GRR/NRR — how much of existing customers' revenue you keep), and ARPU/ARPA (per-customer or per-account averages). Getting revenue right is the highest-stakes metric problem in BI.
Why do you need it?
- Revenue is the number boards, investors and teams act on — wrong versions cause real damage.
- Timing and recognition differences make 'revenue' genuinely ambiguous (bookings vs recognized),
- Retention metrics expose what growth hides: churn quietly eats new signups.
- Consistent revenue definitions are the foundation of every revenue dashboard and forecast.
Mental model
A gym membership ledger
Recurring revenue
You sign a 12-month contract (bookings), count earned money monthly (MRR/recognized), and if members quit or downgrade you lose future revenue even while growth signs new people. Retention is the counter-read-out on the same gauge.
How it works
- Bookings capture contracted value at signing; recognized revenue accrues as service is delivered.
- MRR sums the recurring value of active subscriptions at month-end, prorating starts, upgrades and downgrades.
- GRR measures retained revenue from existing customers (excludes expansion); NRR includes expansions.
- ARPU/ARPA divides revenue by active customers or accounts.
- Refunds, failed payments and churn mechanically flow into the same fact model so the metrics reconcile.
How the core revenue metrics relate.
In production
- A SaaS finance team publishes MRR with expansion and contraction splits, and an exec dashboard shows GRR vs NRR to separate growth-engine from churn problem.
- A subscription company reconciles recognized revenue (accounting) with bookings (sales) so investors and tax see the same story with different lenses.
- Sales ops track ARPU by cohort to confirm pricing changes lift revenue per account, not just total bookings.
Best practices
- Separate bookings, recognized revenue and MRR as distinct measures with distinct as-of rules.
- Define churn/renewal windows exactly — cancelled mid-month vs due-date semantics change retention.
- Prorate properly: upgrades and downgrades land in the right period's MRR.
- Reconcile revenue dashboards to the billing/ledger source monthly; a mismatch is a data bug, not a rounding choice.
Common mistakes
- Reporting bookings as revenue, overstating earned value.
- Ignoring prorations so upgrade mid-month shows as a full-period jump.
- Counting canceled-but-not-refunded accounts as revenue.
- Double-counting expansion in NRR while also reporting it as new logo growth.
Common issues
MRR spiked on the 15th of every monthMid-month upgrades billed as full-period instead of prorated
Prorate changes to effective-date weighting within the MRR measure.
NRR exceeds expectations unreasonablyExpansion double-counted across accounts or churn under-reported
Define the population (all active accounts) and sum net revenue change per account.
Debugging tips
- When MRR jumps oddly, check proration and effective dates on new/upgraded subscriptions.
- When GRR looks too good, verify you excluded expansion and churned-later accounts correctly.
- Reconcile to source: MRR plus refunds minus charges-back should equal the ledger movement.
Interview checkpoint
Quick quiz
Which metric best separates growth from existing-customer decay?
A mid-month upgrade should add its full monthly value to the current month's MRR.