How to Calculate Customer Lifetime Value

Customer lifetime value is one of those metrics that lives in every pitch deck and gets calculated carefully by almost no one. The standard formula is simple; the honest version requires four adjustments that most implementations skip. As of 2026-Q2, the standard academic treatment is Gupta et al. 2006^[2], and the contemporary SaaS benchmarks come from OpenView and ChartMogul^[3][4].

1. The formula, done honestly

Start with the simple form:

CLV = ARPU × Gross Margin × Average Customer Lifetime

For a subscription business, Average Customer Lifetime = 1 / Monthly Churn Rate, in months. So:

CLV = (ARPU × Gross Margin) / Monthly Churn Rate

Worked example. ARPU = $100/month, gross margin = 75%, monthly churn = 2.5%. CLV = (100 × 0.75) / 0.025 = $3,000.

This simple form has three hidden assumptions: churn is constant over the customer lifetime, margin is constant, and future dollars are worth the same as today's. All three are wrong in most businesses. The corrections are below.

2. Always use gross margin

Revenue CLV is a vanity number. Gross margin CLV is what you can actually invest in acquiring the next customer. For a subscription business with 75% gross margin, revenue CLV overstates recoverable value by 33%.

Compute gross margin correctly. Include:

• Hosting and infrastructure directly attributable to customer workload.
• Payment processing fees (typically 2.5–3.5% on card, 0.5–1.0% on ACH).
• Third-party software costs resold to customers.
• Customer support headcount allocation.
• Onboarding and implementation labor for the first period it serves.

Exclude: sales commissions (those are CAC), R&D (operating expense, not COGS), and general overhead. In 2024 SaaS benchmarks, gross margins typically ran 70–80% for products with good cost discipline; below 65% suggests a cost-structure problem worth fixing before pricing or growth work^[3].

3. Cohort churn, not aggregate churn

Aggregate churn averages fresh and tenured customers together. In most SaaS businesses, churn is front-loaded: the first 90 days sees disproportionate drop-off, after which retention stabilises.

Using aggregate churn in the CLV formula double-counts the early-churn effect for your mature customer base and undercounts it for new cohorts. The correction: compute CLV separately for cohort age bands — 0–90 days, 90–365 days, 365+ days — and weight appropriately^[1].

Numerical illustration. Aggregate monthly churn = 3%. But decomposed: months 0–3 = 8%, months 4–12 = 3%, months 13+ = 1.5%. A customer who makes it past year one has a 67-month expected remaining lifetime, not the 33 months implied by the aggregate figure. That difference materially changes LTV:CAC economics for products with annual contracts and strong retention after the first year.

4. Discount future cash

A dollar received 36 months from now is not worth a dollar today. The discounted CLV formula:

CLV = Σ [(Margin per period) × (1 − churn)^t] / (1 + r)^t

Where r is the discount rate per period. For most small businesses, r should reflect your cost of capital or opportunity cost — typically 10–15% annually for venture-backed startups, 15–25% for bootstrapped businesses that could deploy capital to other initiatives^[2].

Practical impact: at 15% annual discount rate and 2.5% monthly churn, the undiscounted CLV of $3,000 from the earlier example becomes roughly $2,400 in present-value terms. A 20% haircut is typical.

5. The LTV:CAC ratio is where teams overclaim

The standard target is 3:1 — every dollar of CAC should return three dollars of lifetime margin. This is a reasonable heuristic, but only if both sides of the ratio are calculated cleanly.

Common overstatements:

• CAC excludes sales compensation or loaded-cost. Fully loaded CAC includes sales salaries, commissions, marketing spend, sales tools, and allocated overhead for these functions.
• Organic acquisition treated as zero CAC. Acceptable if it's truly word-of-mouth, but most "organic" includes content marketing, SEO, and community investments that have real costs.
• LTV on revenue, CAC on fully loaded cost. Mixing bases inflates the ratio mechanically.

Payback period is often more decision-relevant than the ratio. CAC Payback = CAC / (ARPU × Gross Margin) in months. OpenView 2024 data shows median payback of 18–24 months for mid-market B2B SaaS, 12–18 months for SMB^[3]. Payback over 24 months in a cash-constrained business is a runway problem regardless of the LTV:CAC number.

CLV is most useful as a comparison metric over time. The absolute number has known errors; the trend — is cohort CLV improving, flat, or declining — is the signal that drives decisions.

6. Segment CLV, don't aggregate

Aggregate CLV hides the economics. In most businesses, CLV varies by 3–10x across customer segments. Segment by:

• Acquisition channel. Organic customers typically have CLV 2–3x higher than paid-search customers in the same category, driven by both higher retention and higher expansion.
• Initial plan / entry point. Customers who enter on the highest tier often have CLV 5–10x those entering on the lowest tier, driven by larger ARPU and meaningfully better retention.
• Cohort vintage. Newer cohorts may be different — better onboarding, worse product-market fit in a new segment, pricing change impact. Track CLV by acquisition quarter to catch structural shifts.
• Geography. Churn and ARPU vary by country. Aggregate numbers from a US-heavy base can mask problems in secondary markets.

The decision-relevant question is almost never "what's our CLV?" It's "which channel/segment/cohort has CLV high enough to justify what we're spending to acquire them?" Answer that honestly and most of the budget reallocation decisions become obvious.

7. CLV is a statistical estimate, not a point number

Every CLV calculation makes assumptions with non-zero error. The standard formula assumes constant churn and constant margin — both approximations. The academic literature on CLV modeling includes Pareto/NBD, BG/NBD, and Gamma-Gamma models that produce distributions of CLV, not point estimates^[1].

For most small businesses, a formal probabilistic model is overkill. What matters is treating CLV as a range: "Median CLV is $2,400, with most customers falling between $1,200 and $4,800 over three years." That range maps onto acquisition decisions more honestly than a single point estimate of $2,400.

Rule of thumb: publish CLV with an implicit ±20–30% error band in internal reporting. Match CAC investments to the lower bound (the conservative case), not the headline number. This discipline keeps you from overspending on acquisition based on optimistic LTV projections — a common failure mode in the 2021–2022 cohort of venture-backed SaaS that looked healthy on paper and reached cash crisis 18 months later^[3].

8. Numeric worked example — honest vs inflated CLV

Same SaaS, two CLV numbers. A $99/month product, 4,000 active customers, claimed "aggregate monthly churn 2.5%". Inflated version uses revenue and aggregate churn; honest version uses gross margin, cohort-weighted churn, and a 15% annual discount.

Inflated CLV
  Revenue/mo        $99
  Churn             2.5% (aggregate)
  Lifetime          40 months
  Revenue CLV       $99 × 40 = $3,960
  LTV:CAC @$900 CAC   4.4x  (looks excellent)

Honest CLV
  ARPU × GM         $99 × 0.73 = $72.27
  Cohort weights    mo 0–3 8% / mo 4–12 3% / mo 13+ 1.5%
  Discount rate     15% annual (≈1.17% monthly)
  Discounted CLV    ~$2,150 across cohort bands
  LTV:CAC @$900 CAC   2.4x  (below 3:1 target)

Same customers, same spend. The inflated math says "scale the paid-acquisition program aggressively"; the honest math says "acquisition is below payback-target, fix margin or channel mix before scaling." The gap is a 45% CLV overstatement driven entirely by definition choices, not by any forecasting disagreement^[2].

9. Failure modes worth naming

• CLV projected on 6 months of data. Any retention curve estimated from less than 12 months has a wide error band. Fader and Hardie's BG/NBD models explicitly need repeat-purchase observation windows long enough to estimate the latent dropout rate^[1]. Under 6 months of cohort data, publish a range not a point — or don't publish at all.
• Expansion revenue double-counted. If your CLV uses NRR-expanded ARPU at each period, and your CAC also captures the cost of the CS team that drove that expansion, you risk booking the gain twice. Either keep CLV at initial-ARPU with separate expansion accounting, or loaded-cost the CS team into CAC — not both and neither.
• Payback period ignored entirely. A 4:1 LTV:CAC with 36-month payback is worse for a cash-constrained business than a 2.8:1 ratio with 9-month payback. Payback gates cash survival; LTV:CAC gates eventual profitability. Both matter, they are not substitutes. ChartMogul 2024 cohort data shows the median SMB SaaS payback moved from 14 to 16 months between 2022 and 2024^[4].

As of 2026-Q2, the standard-of-care recommendation is: compute CLV on gross margin, bucket by cohort age, apply a 12–18% annual discount, and report both payback period and LTV:CAC alongside any absolute number.