In subscription software the customer is the unit of value, not the one-time sale. So three intertwined metrics decide whether a SaaS business is a money machine or a money pit: Customer Acquisition Cost (CAC) — what it costs to win a customer; Customer Lifetime Value (LTV) — the gross profit that customer throws off before leaving; and churn — the rate at which customers leave, which is the hinge that sets the lifetime in LTV. The core tension is timing and faith: a SaaS firm pays the full CAC upfront, in cash, but recovers it slowly over months or years of subscription revenue — and only if its estimate of future retention holds. Misjudge churn and every downstream number, including valuation, is wrong.

How they're calculated

CAC = total sales & marketing spend in a period ÷ new customers acquired in that period. Fully-loaded versions include S&M salaries, overhead, and ad spend; "blended" CAC (which dilutes with organic/free signups) flatters the number, so analysts prefer paid CAC.

Churn comes in two flavors, and conflating them is a classic error. Customer (logo) churn = customers lost ÷ customers at period start. Revenue churn tracks dollars, and splits further:

• Gross Revenue Retention (GRR) = retained recurring revenue ÷ starting recurring revenue, excluding upsells. GRR can never exceed 100%; churn ≈ (1 − GRR).
• Net Revenue Retention (NRR / NDR) = the same but including expansion (upsells, seat growth). NRR can exceed 100% — meaning the existing base grows even with zero new logos.

LTV. David Skok's canonical formula (the standard reference, forentrepreneurs.com) is LTV = (ARPA × Gross-Margin %) ÷ Customer Churn Rate, where ARPA is average revenue per account. The gross-margin term is not optional: Skok stresses that an 80%-margin business is worth far more than a 40%-margin one at identical revenue, so LTV must be measured in gross profit, not revenue. The division by churn comes from a constant-decay assumption: lifetime ≈ 1 ÷ churn rate (5% monthly churn → 20-month average life).

Two derived ratios do most of the analytical work:

• LTV:CAC ratio = LTV ÷ CAC. Skok's widely-cited rule of thumb is > 3 (he notes this assumes ~80%+ gross margin).
• CAC Payback Period = CAC ÷ (ARPA × Gross-Margin %), in months — how long until a customer's gross profit repays acquisition cost.

How they're used in practice

These metrics together describe unit economics — does one customer make money? An investor or operator reads them as a system. CAC payback measures cash efficiency (how fast capital recycles); LTV:CAC measures profitability of the acquisition engine; NRR vs. GRR separates two distinct questions — are customers leaving? (GRR) and are the survivors growing? (the NRR–GRR gap, which equals expansion ARR ÷ starting ARR).

For equity analysis, NRR is arguably the single highest-signal SaaS metric: a company with 120%+ NRR compounds revenue from its installed base alone, which is why high-NRR names command premium revenue multiples. Churn feeds growth modeling directly — at 2% monthly churn a firm must replace ~22% of revenue annually just to stand still ("the leaky bucket"). Analysts also watch the trajectory of these numbers across cohorts more than any single snapshot.

Adoption, debate & evidence

These metrics are near-universal in SaaS — standard in VC diligence, board decks, and S-1 filings (where NRR/NDR is now a routinely disclosed line). Published benchmarks give rough goalposts, though they vary by source and segment:

• NRR: median private B2B SaaS ~101% in 2024 (SaaS Capital); commonly cited bands put >130% as best-in-class, 100–120% good, <100% concerning. NRR scales with deal size — roughly ~118% enterprise vs. ~97% SMB per segmented benchmarks (Benchmarkit / Optifai).
• GRR: median private B2B SaaS ~88% in 2024 (down from ~90% in 2022), with top performers >90–95%; sustained GRR below ~90% is often read as a structural product-fit or pricing problem expansion can't mask (SaaS Capital).
• CAC payback: benchmark tiers commonly cited as best-in-class <12 months, "good" ~12–18, concerning 18–24, critical >24 (Benchmarkit/Optifai framing). Skok's long-standing rule of thumb is "<12 months," noting that 15–18 months "puts an enormous strain on capital." Payback lengthens with deal size — enterprise (>$100K ACV) segments commonly run ~18–24 months per segmented benchmarks. The median across private SaaS rose to ~18 months in 2024 (Benchmarkit), up from prior years.
• LTV:CAC: ">3" is the folklore standard (Skok); a ratio <1 means the firm loses money per customer, while >5 may signal underinvestment in growth.

The serious, well-documented critique comes from David Skok himself: he says his early writing pushed founders to obsess over LTV/CAC too early, before a repeatable sales model exists, making the figures "meaningless." The deeper flaw is the constant-churn assumption. The 1 ÷ churn formula extrapolates today's churn forever, which (a) overstates lifetime when churn rises with tenure or in a downturn, (b) can't be computed reliably for young firms with little history, and (c) breaks for negative-churn (NRR > 100%) businesses where naive LTV goes to infinity. Cohort-based and survival-curve LTV models are more honest but harder. Bottom line: these are decision frameworks, not precise truths — the inputs (especially future churn) are estimates, and the output inherits all their uncertainty.

Strengths & limitations

When they work: at scale (commonly cited as >$5–10M ARR) with a stable, repeatable acquisition motion and several years of cohort data, LTV:CAC and CAC payback are reliable efficiency gauges and excellent for comparing channels, segments, and cohorts.

When they fail: early-stage or rapidly-changing businesses, where churn history is thin or non-stationary. The #1 misuse is treating a computed LTV as fact — small errors in the churn denominator produce large LTV swings, and blended/un-margin-adjusted versions inflate the number. Secondary traps: using revenue (not gross profit) LTV, using blended instead of paid CAC, and reading NRR without GRR (healthy NRR can mask severe logo churn papered over by a few big expansions).

Sources

• David Skok, SaaS Metrics 2.0 — Detailed Definitions, forentrepreneurs.com — LTV/CAC formulas, gross-margin requirement, ">3" and payback guidelines, and his own critique of premature use.
• SaaS Capital, What is a Good Retention Rate for a Private SaaS Company? — 2024 median NRR ~101%, GRR ~88%; segment NRR (~118% enterprise vs ~97% SMB).
• Benchmarkit, 2024/2025 SaaS Performance Metrics; Optifai — NRR/GRR by segment, CAC-payback bands.
• Wall Street Prep — LTV/CAC Ratio and CAC Payback Period formula references and benchmark framing.
• Intercom / ChartMogul interviews with David Skok — stage-appropriate use of metrics; constant-churn caveat.

Disputes flagged: exact benchmark numbers differ by source, year, and segment — all figures above are qualified as medians/common bands, not universal constants. The LTV constant-churn assumption is genuinely contested even by its popularizer.