Unit economics
Unit economics measures the revenue and costs of a single unit, usually one customer, to show whether each sale makes or loses money.
Unit economics looks at one customer (or one account, or one seat) and asks whether the business makes money on them. Total revenue and total profit can hide a broken model, especially in a fast-growing company. Per-customer numbers show the truth sooner.
David Skok frames it as a single question: can you make more profit from your customers than it costs to acquire them? He names two metrics as the core of the answer, customer lifetime value and customer acquisition cost.
The pieces
- Revenue per customer. Monthly ARPA or annual contract value.
- Cost to serve. Hosting, support and payment fees, which sit in COGS and give you gross margin.
- Cost to acquire. CAC.
- How long they stay. Driven by churn, which turns monthly profit into lifetime value.
A worked example
A tool sells for $50 per month. Serving each customer costs $10 (hosting, support, card fees), so gross profit is $40, an 80% margin. Monthly churn is 4%, so the average customer stays 25 months. Acquiring one customer costs $300.
LTV calculator
Open the full calculatorEstimate customer lifetime value from revenue per account, gross margin and churn, and compare it with what a customer costs to acquire.
Customer lifetime value (LTV)
$1,333
- Gross profit per account per month
- $40.00
- Average customer lifetime
- 33.3 months
- LTV:CAC ratio
- 4.4x
- CAC payback
- 7.5 months
At or above 3, the benchmark most SaaS investors and operators use.
Under 12 months, a common target for small SaaS.
This simple LTV assumes steady churn and no expansion revenue. Use gross profit, not revenue, so LTV compares fairly with CAC.
Each customer returns about $700 over their life after paying back acquisition. The LTV:CAC ratio and payback period are the two summary numbers most people quote.
How Skok describes the cash curve
Skok points out that a business can have healthy unit economics and still lose money for a while, because you pay for acquisition up front and recover it over months. Cash flow turns positive once the profit from existing customers covers the investment in new ones, as long as you do not keep raising acquisition spend. That is why a growing company with great unit economics can look unprofitable on paper.
Common mistakes
- Using revenue instead of gross profit when calculating lifetime value.
- Leaving the founder's time out of CAC.
- Mixing segments. Unit economics for $10 plans and $500 plans should be separate.
- Trusting churn from only a few months of data.
- Calling it done once the ratio looks good, without checking that you can still afford the cash gap.
When cost to serve stops being small
The worked example above assumes each customer costs $10 a month to serve, whatever they do. For most classic SaaS that was close enough. Elena Verna, who works on growth at Lovable, wrote in October 2026 that the assumption breaks for AI products. In traditional SaaS the marginal cost of another customer was usually low and predictable. With AI, each extra interaction can carry a real inference cost, and an agent can do very different amounts of work to deliver the same result for two customers.
We’re also struggling to predict what those actions will cost us.
Elena Verna / What on earth is going on with AI pricing and packaging? (October 16, 2026)
Pricing is shifting to follow those costs. Kyle Poyar reports in Growth Unhinged (October 7, 2026) that 76% of AI-native companies already use some form of variable or hybrid pricing, compared with 49% of SaaS companies. Automation startup Gumloop went further: it passes tokens and compute through at cost and adds an 8% orchestration fee, so its margin is set on every unit of usage.
Restricting usage is not the only response. Verna also wrote in September that AI made companies panic about gross margin, while Lovable still treats giving its product away for free as one of its biggest growth levers. If you do the same, count the cost of free usage as part of CAC so the per-customer math stays honest.
If your product has AI features:
- Measure cost to serve per customer or per plan, not as one average. A few heavy users can have negative gross profit while the average looks fine.
- Recheck gross margin whenever you switch models or ship a feature that calls one.
- Put limits or usage-based pricing on the expensive actions, so revenue and cost move together.
For small SaaS
If you run a small product, redo this math every quarter and by channel. Raising your price by $10 or lowering churn by one point often moves the numbers more than any marketing change. Unit economics also sets your ceiling: if the numbers are weak, more customers just means a larger loss.
Questions to ask each quarter
- Does a new customer repay CAC before the typical customer churns?
- Which plan or segment has the best ratio, and can you sell more of it?
- If you raised prices by 10%, how many fewer customers could you afford to win?
- What happens to gross margin if usage doubles?
Answering these with real numbers is more useful than debating whether a ratio is good in general. Write the answers down, and your pricing and channel decisions will start to follow from them.
Listen, watch and read next
- Read: What on earth is going on with AI pricing and packaging?, Elena Verna (October 16, 2026). A candid look, from inside an AI company, at why credits, tokens and usage limits are confusing and why the cost side is hard to predict.
- Watch: Masterclass on monetizing AI profitably, Kyle Poyar at Metronome's Monetize 2026 (October 8, 2026). He draws on his 2026 survey of 230 B2B software and AI companies to show how they price and package AI.
- Read: The AI monetization debate shifts again, Kyle Poyar (October 7, 2026). Outcome-based, pass-through and hybrid pricing, with recent examples. The first half is free; the rest is for paid subscribers.
- Listen: The Messy Middle of a SaaS Pivot, Startups for the Rest of Us, episode 851 (September 22, 2026). Rob Walling and Nathan Tyler; see "Pricing migrations, hosting, and AI costs" (32:10).
Related terms
- LTV:CAC ratio
- CAC (Customer acquisition cost)
- CAC payback period
- Gross margin
- Contribution margin
- LTV (Customer lifetime value)
Sources
- SaaS Metrics 2.0: A Guide to Measuring and Improving What Matters, David Skok, For Entrepreneurs
- 16 Startup Metrics, a16z
- What on earth is going on with AI pricing and packaging?, Elena Verna, Elena's Growth Scoop, October 16, 2026
- How to give away free product and make money doing it, Elena Verna, Elena's Growth Scoop, September 8, 2026
- The AI monetization debate shifts again, Kyle Poyar, Growth Unhinged, October 7, 2026