LTV Calculation Guide.

// 01Why LTV matters

LTV (Customer Lifetime Value) sets the boundary on four critical decisions:

• Marketing budget ceiling: LTV/3 = healthy CAC max. Without LTV, you can't tell if your budget is aggressive or conservative.
• Segmentation priority: VIP segments often have 4-8x average LTV; they deserve own campaigns, channels, content.
• Retention vs acquisition: keeping customers costs 5-25x less than acquiring; you can't make this case without LTV.
• Investor / strategic talk: LTV:CAC and payback drive SaaS valuations.

// 02Simple formula (e-commerce)

How to compute each:

• AOV: Total Revenue / Total Orders, last 12 months.
• Order Frequency: Total Orders / Unique Customers, last 12 months.
• Customer Lifespan: 1 / Annual Churn Rate. 35% annual churn = 2.86 years average.
• Gross Margin: (Revenue − COGS) / Revenue.

// 03Pitfalls of the simple formula

One-formula LTV is fast but has three big problems:

1. Past ≠ future. Markets change, products change, churn shifts. Projecting last-12-month average forward is risky.
2. Averages mask segment variance. If 20% of customers produce 60% of revenue, a single average washes two distinct populations.
3. Acquisition cohorts behave differently. Black Friday-acquired customers churn 6-9 months earlier than organic; the average hides it.

Serious marketers move to cohort analysis.

// 04Cohort-based LTV

Cohort method reveals real LTV by tracking behavior over time. Steps:

1. Group by cohort: bucket customers by first-purchase month (Jan-25, Feb-25, etc.).
2. Compute monthly cumulative revenue per cohort (with margin applied).
3. Plot the curve: X = months since cohort start (0, 1, 2... 24), Y = cumulative LTV.
4. Find asymptote: when the curve flattens, that's real LTV. If still rising, extrapolate carefully.

Step-by-step in Excel

Need: a table with customer_id, first_order_date, order_date, net_revenue. Then:

• Pivot: rows = first_order_month (cohort), cols = month_diff, values = sum(net_revenue).
• Cumulative across rows.
• Divide by cohort customer count → average cumulative LTV.
• Chart, read asymptote.

// 05SaaS LTV formula

In SaaS, customer lifespan = 1 / monthly churn. Stable ARPU + small churn = big LTV.

Churn impact

Monthly Churn	Avg Lifespan (months)	LTV ($800 ARPU, 75% margin)
5%	20	$12,000
3%	33	$19,800
2%	50	$30,000
1%	100	$60,000

Cutting churn from 5% to 2% multiplies LTV 2.5x — most SaaS get higher ROI from churn reduction than from CAC reduction.

// 06Predictive LTV (machine learning)

With mature data (1000+ customers, 12+ months history), move to predictive LTV:

• BG/NBD model: Beta-Geometric / NBD; classic probabilistic. Predicts repeat-purchase probability + frequency.
• Gamma-Gamma model: models per-order value distribution. Combined with BG/NBD = predictive LTV.
• RFM-based XGBoost: 12-month LTV prediction using RFM features + categorical signals.

Python's lifetimes library implements BG/NBD + Gamma-Gamma; 200 lines of code gives a working model. Production should sit with the data team, not marketing.

// 07LTV:CAC health benchmarks

LTV:CAC	Interpretation
< 1	Losing money on each customer.
1-2	Break-even, hard to grow.
3	Healthy SaaS standard.
4-5	Excellent, can invest aggressively.
> 5	Underspending; budget upside.

Payback period

CAC payback = CAC / Monthly Gross Margin. Target: SaaS <12 months, e-commerce break-even at first order.

// 08Common mistakes

• Skipping gross margin: gross LTV instead of net inflates by 50-70%.
• Paid-only CAC: dividing only ad spend, ignoring salaries + tools. Real CAC is 1.5-3x higher.
• Single global LTV: insufficient for strategy; calculate RFM-based segment LTVs separately.
• Bad extrapolation: projecting 24-month LTV from 3 months of data — statistically rotten.
• Ignoring acquisition channel: Meta-acquired customers behave differently than Google. Channel × cohort matrix beats one number.