Finance
Unit Economics
Tests whether a business makes money on each unit sold and each customer acquired, tracing revenue through variable costs to contribution and comparing the cost of acquiring a customer with the lifetime value they return. If the unit loses money, scale multiplies the loss.
Also known as Per-unit profitability analysis, LTV/CAC analysis, SaaS metrics. First set out by Practice-grown; codified for SaaS by David Skok among others in 2010; the primary source is cited in full below.
Where this is contested
No single originator exists. Unit economics grew out of contribution-margin accounting and 2000s venture and SaaS practice; David Skok's forEntrepreneurs essays, published from 2010, are the closest thing to a canonical codification, and the familiar 3:1 and 12-month benchmarks are folk rules distilled from investor experience rather than published research.
- Format
- Scoring model
- Level
- Business unit · Product
- Best for
- Evaluate options · Allocate resources · Assess risk
- Decision stage
- Diagnose · Decide · Review
- Difficulty
- Intermediate
- Time to apply
- A day for a first honest pass; ongoing thereafter, since cohort data changes the answer as it matures.
Plate · The model
The components
Revenue per unit
What one unit actually brings in, net of discounts, VAT, refunds and promotional pricing. For subscriptions this is usually average revenue per customer per month; for transactional businesses, average order value.
Signals of strength
Heavy introductory discounting flattering the headline price · Refunds and credits excluded from the revenue figure · Averages concealing a wide spread between segments
Variable cost per unit
Every cost incurred because this specific unit was sold: goods, packaging, fulfilment, delivery, payment processing, per-unit support. The test is whether the cost disappears when the unit does.
Signals of strength
Delivery or support costs booked as overhead to flatter margins · Costs that scale with orders growing faster than revenue · No one able to state the full cost of serving one order
Contribution margin
Revenue per unit minus variable cost per unit, the money each unit generates towards fixed costs and acquisition spend. Negative contribution means every sale destroys value and growth accelerates the destruction.
Signals of strength
Contribution positive only before delivery and payment costs · Margin eroding as the mix shifts to discounted units · Growth targets set with no contribution floor attached
Customer acquisition cost
The fully loaded sales and marketing cost of winning one new customer, including salaries, agency fees, content and incentive vouchers. Understating CAC is the commonest way unit economics are flattered.
Signals of strength
CAC computed from media spend alone · Rising CAC as cheap early channels saturate · Paid channels reported blended with organic to lower the average
Customer lifetime value
The total contribution a customer is expected to deliver before they leave, driven by monthly contribution and by how long they stay. It is a forecast rather than a fact, and it is only as good as the churn assumption beneath it.
Signals of strength
LTV built on revenue rather than contribution · A single churn rate applied to all cohorts and channels · Lifetime assumptions longer than the company has existed
LTV to CAC ratio
Lifetime value divided by acquisition cost: the return on a pound of acquisition spend. Practice treats roughly 3:1 as healthy and payback under 12 months as sustainable, benchmarks popularised by David Skok from observation of mature SaaS firms. They are folk heuristics, sensitive to every assumption upstream, and a precise-looking ratio deserves suspicion in proportion to its precision.
Signals of strength
A ratio below 1 meaning acquisition destroys value outright · A very high ratio suggesting underinvestment in growth · The ratio moving sharply when churn is re-estimated from cohorts
When it earns its keep
- You are growing on external capital and need to know whether growth is building value or subsidising every sale.
- You are choosing between acquisition channels and need each channel's true cost per customer set against the value those customers go on to deliver.
- Top-line growth looks strong but the business keeps needing cash, and you want to find where the money leaks at the level of a single order or customer.
- You are pricing or repackaging an offer and need to see how the change moves contribution per unit and payback on acquisition spend.
And when it doesn't
- The business is too early to have real retention data. LTV computed from three months of cohort history is a guess wearing a decimal point.
- The economics are driven by fixed-cost recovery rather than per-unit margins, as in capital-intensive utilities or infrastructure, where unit-level arithmetic misses the real question of utilisation.
- Units are heterogeneous and cross-subsidising by design, for instance marketplaces where one side is deliberately loss-making; naive per-unit sums mislead without a whole-system view.
- You need a statutory or investor-grade view of profitability. Unit economics is a management lens, and it deliberately excludes overheads that accountants rightly insist on.
How to run it
Before starting, gather the inputs the analysis depends on:
- A clear definition of the unit: an order, a subscriber-month, a delivered box, a seat. Everything downstream depends on this choice being consistent.
- Revenue per unit and a complete list of the variable costs each unit incurs, including delivery, payment fees, support and refunds.
- Fully loaded sales and marketing spend, and the number of genuinely new customers it produced, to compute customer acquisition cost.
- Cohort retention or churn data, so lifetime value rests on observed behaviour rather than optimism.
- Segmentation by channel and cohort, since blended averages are where bad unit economics hide.
- 1
Define the unit
Decide what one unit is: an order, a subscriber-month, a customer. For subscription businesses the customer is usually the unit that matters, with orders as the building block. Write the definition down; most disagreements about unit economics turn out to be disagreements about the unit.
- 2
Build contribution per unit
Contribution = revenue per unit minus all variable costs per unit. Be ruthless about what counts as variable: ingredients, packaging, picking, delivery, payment fees, refunds, per-order support. A business that only reaches positive contribution by reclassifying variable costs as overhead has not fixed anything.
- 3
Measure customer acquisition cost
CAC = total sales and marketing spend in a period divided by new customers acquired in that period, fully loaded with salaries, agency fees and discounts or vouchers used as bait. Compute it blended and by channel; the blended figure hides the marginal channel you are actually deciding about.
- 4
Estimate customer lifetime value
LTV = contribution per customer per month multiplied by expected customer lifetime, where lifetime is roughly 1 divided by the monthly churn rate. Build it on contribution, never on revenue, and prefer observed cohort curves to a single churn number, because early churn is usually far higher than mature churn.
- 5
Compare LTV to CAC and check payback
Compute the LTV to CAC ratio and the CAC payback period (CAC divided by monthly contribution per customer). Practice treats a ratio around 3:1 and payback under about 12 months as healthy, per Skok's benchmarks. Treat these as folk benchmarks distilled from venture experience with mature SaaS companies, useful as conversation starters rather than laws.
- 6
Segment, stress and decide
Break the numbers out by cohort, channel and segment, then stress the churn assumption, since LTV is hostage to it. The output should change behaviour: which channels get budget, which segments get targeted, whether price or cost structure has to move before scaling.
Reading the result
A per-unit profit-and-loss showing contribution per unit, CAC and LTV by channel and cohort, the LTV to CAC ratio and CAC payback period, and a judgement about whether the business gets stronger or weaker as it scales.
- Read contribution first. If the unit loses money before any acquisition spend, no acquisition efficiency can save it; price or cost structure must change.
- Read the ratio and payback together. A healthy-looking LTV to CAC ratio with a two-year payback still means growth consumes cash for two years per customer.
- Distrust blends. A 3:1 blended ratio can hide a 6:1 organic channel and a 1:1 paid channel, and the marginal pound is going into the 1:1 channel.
- Ask what the LTV assumes about churn, and how much history supports it. The ratio inherits all the fragility of that one assumption.
A worked example
A meal-kit subscription business asks whether growth is worth funding
A Bristol-based meal-kit company delivers recipe boxes across the south west. Revenue is growing 60 per cent a year on the back of paid social advertising, but the business keeps raising to fund marketing. Before the next round, the founders work through the unit economics with the customer as the unit and the box as the building block.
- Revenue per unit
- Average order value is 42 pounds per box after discounts. The average customer takes 3.2 boxes a month, so revenue per customer is about 134 pounds a month. Introductory offers give 40 per cent off the first two boxes, which the team had been excluding; included, first-month revenue is materially lower.
- Variable cost per unit
- Per box: ingredients 16 pounds, packaging and chill-chain materials 4, picking and packing labour 3, courier delivery 6.50, giving 29.50. Payment fees and refunds add roughly 1 pound. Delivery had been sitting in overheads; moving it into the unit cost was the single biggest correction.
- Contribution margin
- Contribution is about 11.50 pounds per box, a 27 per cent margin, or roughly 37 pounds per customer per month at 3.2 boxes. Positive, but thin enough that a courier price rise or heavier discounting would bite hard.
- Customer acquisition cost
- Fully loaded monthly sales and marketing spend of 87,000 pounds (paid social, agency, content, introductory discounts) produced about 600 genuinely new customers, a blended CAC of 145 pounds. By channel it splits into roughly 90 pounds for referral and organic and over 210 pounds for paid social, the channel taking most of the budget.
- Customer lifetime value
- Naive churn of 9 per cent a month implies an 11-month lifetime and an LTV near 410 pounds. Cohort curves tell a harsher story: nearly half of customers lapse within three months, and only the survivors churn at 9 per cent. Cohort-weighted LTV comes out closer to 310 pounds.
- LTV to CAC ratio
- Blended: 310 / 145 = 2.1, below the 3:1 folk benchmark. Paid social alone is about 1.5 with payback near six months; referral is above 3.4. Payback on the blended numbers is 145 / 37, roughly four months, which is comfortable and the one genuinely reassuring figure.
The read. The business is viable per unit but the marginal pound of marketing is going into its worst channel. Paid social at 1.5:1 is buying growth the next funding round will have to pay for, while referral at 3.4:1 is starved. The analysis says fix early churn first, since half the acquisition spend is buying customers who leave within a quarter, shift budget towards referral, and only then scale. The 3:1 benchmark did its job as a prompt, but the cohort curve, and what it reveals about the first ninety days, is the finding.
Pitfalls
- Computing LTV on revenue instead of contribution, which can double or triple the apparent ratio and is the most common flattering error.
- Understating CAC by counting media spend only, leaving out salaries, agency fees, content and the discounts used to convert.
- Trusting a single blended average. Channel and cohort segmentation is where the decisions live; blends are where problems hide.
- Projecting long lifetimes from short history. A business eighteen months old claiming a five-year customer lifetime is reporting hope.
- Treating the 3:1 ratio and 12-month payback as laws. They are folk benchmarks distilled from mature SaaS companies and transfer imperfectly to other models and stages.
- Forgetting fixed costs entirely. Healthy unit economics with crushing overheads is still an unprofitable company.
What the critics say
Bill Gurley argued that the LTV formula is routinely abused as a justification for aggressive spending: its inputs are estimates under management control, the formula is applied to unproven businesses as if predictive, and companies come to run themselves in service of the formula rather than the customer.
Gurley, B. (2012) 'The Dangerous Seduction of the Lifetime Value (LTV) Formula', Above the Crowd, 4 September 2012.
McCarthy and Fader's work on customer-based corporate valuation shows that publicly reported unit economics are frequently inflated by optimistic retention assumptions, and that rigorous cohort modelling of acquisition, retention and spend often yields far less flattering figures than headline LTV to CAC ratios.
McCarthy, D. and Fader, P. (2020) 'How to Value a Company by Analyzing Its Customers', Harvard Business Review, 98(1), January-February 2020.
The framework has no canonical definition, so metrics are not comparable across companies: what counts as variable cost, which spend loads into CAC and how lifetime is bounded all vary by teller. The benchmarks in circulation, including Skok's own, derive from mature SaaS businesses and were never validated for other models or earlier stages.
Work it through
Work at the level of one customer. Enter monthly revenue and variable cost per customer, the fully loaded acquisition cost, and monthly churn. Contribution, lifetime value, the LTV to CAC ratio and the payback period compute below. Read contribution first: if the unit loses money, no acquisition efficiency can save it.
Inputs
Result
- Contribution per customer / month
- £37
- Expected lifetime
- 11.1 months
- Lifetime value (LTV)
- £411.11
- LTV to CAC
- 2.8:1
- CAC payback
- 3.9 months
Sources and further reading
- Skok, D. 'SaaS Metrics 2.0: A Guide to Measuring and Improving What Matters', forEntrepreneurs (Matrix Partners). ↗
- Gurley, B. (2012) 'The Dangerous Seduction of the Lifetime Value (LTV) Formula', Above the Crowd. ↗
- McCarthy, D. and Fader, P. (2020) 'How to Value a Company by Analyzing Its Customers', Harvard Business Review, 98(1), January-February 2020. ↗