SaaS Growth Metrics · Cohort Analysis, NDR & Unit Economics

SaaS business health is assessed through a layered metrics framework that flows from acquisition efficiency to retention quality to expansion velocity. The four questions this framework answers: are we acquiring customers efficiently (CAC, payback period)? Are we retaining the revenue we acquire (GDR, churn rate)? Are we growing revenue within our customer base (NDR, expansion rate)? And is our unit economics structure sufficient for the growth investment we are making (LTV:CAC, Rule of 40)?

Annual Recurring Revenue (ARR) and Monthly Recurring Revenue (MRR) are foundational SaaS metrics — but their calculation is frequently inconsistent across organisations, making benchmarking and forecasting unreliable. Standard definitions: MRR is the normalised monthly value of all active subscription contracts at a point in time. ARR = MRR × 12. Neither includes: professional services fees (non-recurring), one-time onboarding fees, usage-based revenue above a contracted minimum (until it is committed), or revenue from churned customers.

The components of MRR movement — the MRR waterfall — should be tracked separately: New MRR (from new customers); Expansion MRR (upsells and seat additions from existing customers); Contraction MRR (downgrades from existing customers); Churned MRR (from cancelled customers); and Reactivation MRR (from customers who previously churned and re-subscribed). The algebraic sum of these five components should equal the change in total MRR between two periods.

Common calculation errors: annualising monthly contracts with variable pricing (treating a 3-month minimum as ARR); including one-time fees in ARR; inconsistent treatment of trials (pre-revenue trials should not be included in ARR); and inconsistent timing conventions (some organisations book ARR at contract signature; others at go-live date). Standardising these conventions is a prerequisite for reliable forecasting.

Cohort analysis groups customers by the period they were acquired and tracks their collective behaviour over time. It is the most revealing SaaS diagnostic because it separates the effect of acquisition volume changes from retention quality changes — something that aggregate metrics like total churn rate cannot do.

The foundational cohort chart: rows are acquisition cohorts (e.g., Jan 2023, Feb 2023...); columns are months since acquisition (Month 0, Month 1, Month 2...); cells show the revenue retained from that cohort at that time point, expressed as a percentage of their starting revenue. A healthy cohort analysis shows flattening curves — early churn followed by stabilisation as the retained customers settle into long-term retention patterns.

What cohort analysis reveals: whether retention is improving or worsening over time (horizontal comparison — are more recent cohorts retaining better than older ones at the same time point?); where in the customer lifecycle churn is concentrated (vertical comparison — is Month 1 churn high? Month 12 churn elevated?); and whether expansion revenue is offsetting churn within cohorts (NRR cohort analysis — do cohorts grow in revenue above their starting point despite losing some customers?).

The dangerous aggregate churn illusion: if a SaaS business is growing rapidly, a constant percentage churn rate means a growing absolute churn number — but the growing new customer volume masks this in aggregate MRR. Cohort analysis makes the retention quality visible regardless of acquisition volume changes.

Net Dollar Retention (NDR, also called Net Revenue Retention or NRR) measures how much revenue a cohort of customers generates in period N+12 relative to period N, including expansion (upsells, cross-sells, seat additions) and contraction (downgrades and churn): NDR = (Starting MRR + Expansion MRR − Contraction MRR − Churned MRR) / Starting MRR × 100.

An NDR above 100% means the existing customer base grows in revenue without any new customer acquisition — a compound growth engine. The documented NDR benchmarks for public SaaS companies (Bessemer Venture Partners State of the Cloud data): median NDR is approximately 106%; top quartile exceeds 120%; and NDR above 130% characterises the elite "land and expand" companies like Snowflake and Datadog that grow primarily through expansion.

Gross Dollar Retention (GDR) measures only the retention component without expansion: GDR = (Starting MRR − Churned MRR − Contraction MRR) / Starting MRR × 100. GDR has a ceiling of 100% (expansion is excluded). The difference between NDR and GDR is the expansion rate. GDR below 85% indicates a product-market fit or customer success problem that expansion revenue is masking; no amount of upselling compensates for high underlying churn.

CAC payback period measures how long it takes for the gross margin contribution of a new customer to recover the cost of acquiring them. It is a more operationally useful metric than LTV:CAC ratio because it measures the cash timing of acquisition investment recovery — critical for cash flow management and growth investment pacing.

The calculation: CAC Payback (months) = CAC / (MRR × Gross Margin %). If acquiring a customer costs £1,200 and they pay £100/month at 75% gross margin (£75 contribution per month), payback = 1,200 / 75 = 16 months.

Segmenting payback by channel and customer segment reveals which acquisition activities are efficient and which are not. A channel with a 10-month payback operating alongside a channel with a 36-month payback should prompt a reallocation question — even if both channels generate technically "positive ROI" at 5-year LTV, the capital efficiency difference is significant at scale. Documented benchmarks: SaaS companies targeting SMBs aim for 12–18 month payback; enterprise SaaS can sustain longer payback periods (24–30 months) because contract values and retention are higher, but longer payback requires more working capital to fund growth.

The Sales Efficiency Magic Number (first documented by Josh Kopelman and popularised in SaaS circles by Bessemer Venture Partners) measures how much new ARR is generated for each pound of sales and marketing spend: Magic Number = (Current Quarter ARR − Prior Quarter ARR) × 4 / Prior Quarter S&M Spend.

Interpretation: a Magic Number above 1.0 means you generate more than £1 of annualised new revenue for every £1 of sales and marketing spend — a strong signal to invest more in growth. Between 0.75 and 1.0 is acceptable and suggests measured growth investment. Below 0.75 suggests investigating customer acquisition efficiency before scaling spend. Below 0.5 is a warning sign that the go-to-market motion needs improvement before significant growth investment makes sense.

The Magic Number's limitation: it uses annualised revenue (×4 quarterly) which assumes linear growth patterns and consistent ACV — assumptions that may not hold for businesses with seasonal demand or large deal variation. Adjusting for gross margin (replacing ARR with ARR × gross margin %) produces a more conservative "Gross Profit Magic Number" that accounts for the profitability of the growth being generated.

Expansion revenue — additional ARR from existing customers through upsells, cross-sells, seat additions, and tier upgrades — is the highest-margin revenue in a SaaS business because customer acquisition cost is near zero. Building reliable expansion revenue requires both product architecture (pricing structures that create natural expansion triggers) and go-to-market motions (customer success and account management processes that identify and capture expansion opportunities).

Expansion pricing mechanisms: per-seat pricing (adding users drives expansion — the most common enterprise SaaS mechanism); usage-based pricing (consumption expands as adoption grows — Snowflake, Twilio, AWS); tier-based upsells (feature gates that become relevant as customer sophistication grows); and add-on module cross-selling (complementary products that solve adjacent problems for the same buyer). The most robust expansion architectures combine two of these — for example, per-seat pricing with usage-based components that expose customers to expansion triggers through natural growth.

Product-Led Growth (PLG) companies (Slack, Dropbox, Calendly, Figma) acquire and expand through the product itself rather than primarily through sales and marketing. PLG introduces a distinct metrics vocabulary: the free-to-paid conversion rate (what percentage of free users become paying users); time-to-value (how quickly new users reach the product activation milestone that predicts conversion); viral coefficient (how many new users each existing user brings in); and the Product Qualified Lead (PQL) — a free user who has reached sufficient engagement or usage to be a high-probability conversion target for sales outreach.

Activation is the most leveraged PLG metric because it predicts downstream conversion and retention. Identifying the activation milestone — the specific in-product behaviour that most strongly predicts a user converting to paid — requires cohort analysis of converting vs. non-converting users segmented by their early product behaviour. The activation event is typically reached by 40–60% of trialling users for products with good product-market fit; below 20% usually indicates an onboarding or value realisation problem.

Cohort-based revenue forecasting is more accurate than top-down forecasting for SaaS because it uses observed retention curves from historical cohorts to model future revenue from existing customers, then adds modelled new business assumptions. The forecast structure: (1) model the expected revenue from existing cohorts in future periods using historical retention curves; (2) add expected new customer cohorts based on pipeline, marketing assumptions, and seasonal patterns; (3) model expansion and contraction within both existing and new cohorts using historical rates.

Sensitivity analysis is essential: the cohort model's forecast depends heavily on assumptions about future retention and expansion rates. Running the model with optimistic, base, and pessimistic retention assumptions produces a forecast range rather than a single-point prediction — much more useful for financial planning and board reporting than a false-precision single number.

For growth-stage SaaS companies, the metrics that most determine valuation multiples (as documented in Bessemer Venture Partners' State of the Cloud data and OpenView's SaaS benchmarks): ARR growth rate (the primary driver — 3× ARR growth year-on-year at meaningful scale commands premium multiples); NDR (above 120% is the threshold for "exceptional" in investor frameworks); gross margin (above 70% is standard for software; below 60% raises business model questions); and CAC payback (below 18 months for efficient growth).

The Rule of 40 — growth rate + EBITDA margin should exceed 40% — is a useful efficiency heuristic but has documented limitations: it treats 60% growth with -20% EBITDA equally to 10% growth with 30% EBITDA, which have very different business implications. Investors increasingly use "efficient growth" frameworks that weight growth rate more heavily at earlier stages and profitability more heavily at later stages, rather than applying the Rule of 40 uniformly across growth stages.

Further Reading

Go deeper with these reference guides from the Digital Codex library.

Sources & References