Unit economics: the per-customer P&L, with LTV, CAC and payback worked for three segments

Unit economics is the profit and loss of one unit of the business, usually one customer: what it pays each month, what it costs to deliver and serve, what it cost to win, and how long it stays. Four numbers carry it: monthly gross margin per customer, customer acquisition cost (CAC), CAC payback in months and lifetime value (LTV). Growth creates value when each new customer returns its acquisition cost several times over, and quickly.

Monthly gross margin per customer = Revenue per account per month × Gross margin %

CAC payback (months) = CAC / Monthly gross margin per customer

LTV = Monthly gross margin per customer / Monthly churn rate

LTV:CAC = LTV / CAC

Each ratio has its own page: customer acquisition cost for what goes into the spend, how to calculate LTV for the lifetime value inputs, and the LTV to CAC ratio for the comparison and the 3:1 convention. This page uses their figures and puts them into one per-customer P&L.

What a unit is

In subscription software and most B2B businesses the unit is the customer account. Elsewhere it may be an order, a unit sold, a delivery route or a store location. Choose the unit your decisions are made on: if sales and marketing budgets are set by customer segment, the unit is a customer in that segment; if a retailer decides where to open stores, it is the location.

Whatever the unit, define it once and keep it. The SEC's guidance on key performance indicators in MD&A expects a metric disclosed to investors to come with "a clear definition of the metric and how it is calculated", and with disclosure when the calculation changes. Internal unit economics deserve the same discipline.

The per-customer P&L

Revenue per account comes first, then the costs a single customer causes, in the order a P&L lists them. For each of three segments of one software company, per account per month (USD):

Gross margin is what the customer leaves after the cost of delivering the product. Contribution goes one line further and takes off the costs that vary with the customer but sit in operating expenses, such as a named account manager. Gross margin vs contribution margin vs net margin sets out where each cost belongs.

Below the line come the two figures that turn a monthly P&L into an investment case: what the customer cost to win (CAC) and how long it stays (churn, and so lifetime).

Worked example: three segments

The same company, with the segment figures from how to calculate LTV and the CACs from the LTV to CAC ratio page. Monthly figures in USD.

Mid-market, step by step: 900 × 0.75 = 675 of gross margin a month. LTV is 675 / 0.015 = 45,000. LTV:CAC is 45,000 / 15,000 = 3.0. Payback is 15,000 / 675 = 22.2 months. With revenue per account in B2, gross margin in C2, churn in D2 and CAC in E2:

=B2*C2/D2          LTV
=B2*C2/D2/E2       LTV:CAC
=E2/(B2*C2)        CAC payback in months

Uncapped, enterprise looks best by a distance: five times its acquisition cost back. It also has the slowest payback, more than two years of gross margin before the account has repaid what it cost to win.

On contribution instead of gross margin the paybacks lengthen: 4,000 / 240 = 16.7 months for SMB, 15,000 / 585 = 25.6 for mid-market and 70,000 / 2,200 = 31.8 for enterprise. Enterprise carries the heaviest cost to serve, so it loses the most.

Cap the lifetime

LTV divides by churn, so the lifetime it assumes is 1 / churn. Enterprise churn of 0.8% a month implies 125 months, more than ten years. Nobody has observed a ten-year customer in a business that has not existed for ten years, and most of that LTV sits in years six to ten. A lifetime that long is an assumption, not evidence.

Count only the gross margin expected in the first 60 months, allowing for the customers who leave along the way:

LTV capped at 60 months = SUM over t = 0 to 59 of Monthly gross margin × (1 − churn)^t

which equals Monthly gross margin × (1 − (1 − churn)^60) / churn

In Excel, with monthly gross margin in B2 and churn in B3:

=B2*(1-(1-B3)^60)/B3

Enterprise: 2,800 × (1 − 0.992^60) / 0.008 = 133,843. Capped, the ranking reverses: SMB has the best ratio, enterprise falls from first to second, and none of the three reaches 2. The segment that looked like the best use of acquisition money is the one whose value depends most on years nobody has seen. A discount rate would cut the long-lived segments further; how to apply one is in how to calculate LTV.

The check that proves it

Unit figures are averages, so they must multiply back to the ledger.

Revenue per account × accounts = ledger revenue. With 300 SMB, 120 mid-market and 20 enterprise accounts:

The 298,000 must equal the month's recurring revenue in the general ledger, and the 221,000 its gross profit (a blended margin of 74.2%). If they differ, revenue per account or margin is wrong before any ratio is.

CAC × new customers = attributed sales and marketing spend. For the quarter: 4,000 × 30 = 120,000 for SMB, 15,000 × 14 = 210,000 for mid-market and 70,000 × 4 = 280,000 for enterprise, a total of 610,000. That must tie to sales and marketing expense on the P&L, less any cost left out on purpose, such as account management for existing customers.

Unit economics without subscriptions

Distributors, manufacturers and services firms rarely have a monthly churn rate: customers do not cancel, they stop ordering. The P&L is the same; the period is a year and retention comes from cohorts in the ledger.

Payback is 20,000 / 13,500 = 1.48 years, about 18 months; lifetime contribution is 13,500 × 6.5 = 87,750. The average relationship of 6.5 years comes from purchase history, the span from first to last order for customers who have stopped. Customer lifetime value from the ledger works that method in full, and B2B churn without a subscription covers how to decide when an account has stopped.

Where it goes wrong

• Uncapped LTV from very low churn. A 0.8% monthly churn rate claims a ten-year customer. Cap it and discount it: the enterprise ratio fell from 5.0 to 1.91.
• CAC that counts only paid media. Leaving out sales salaries, commissions and tools makes every segment look cheap to acquire, and enterprise, which is mostly sales time, cheapest of all.
• Blended unit economics across segments. One company-wide ratio hides a segment that loses money on every customer it wins.
• Revenue instead of gross margin in LTV. Mid-market LTV on revenue is 900 / 0.015 = 60,000 instead of 45,000, overstated by the cost of delivery.
• Logo churn and revenue churn mixed up. Customer counts set the lifetime; revenue retention shows expansion and contraction. State which one you used, and track both.

A widely quoted benchmark puts good subscription businesses above 3:1; David Skok's SaaS Metrics 2.0 says "the best SaaS businesses have a LTV to CAC ratio that is higher than 3". It is a convention from practitioners, not a standard, and it was set on uncapped figures.

The per-customer P&L from your own ledger

Unit economics by segment still hides the customers inside each segment, some well above the average and some below zero. Covirage's customer profitability tools compute gross margin and cost to serve per customer from your ledger and activity files, rank the customers, and reconcile them to the totals; the external AI model explains the ranking and never does the arithmetic. See customer profitability to build the per-customer P&L from your own files. For the per-unit margin, see contribution margin ratio; for cost to serve by activity, activity-based costing; for what acquisition spend does to cash, burn rate.