Unit Economics for Startups: CAC, LTV and Payback, With Worked Examples
What unit economics means, the formulas for customer acquisition cost, lifetime value and payback, worked examples for SaaS and low-priced apps, and how to do the sums before you have revenue.
By Suren, Founder & Growth Strategist at Gufy · Published
Unit economics are the revenue and costs of a business measured for one unit, usually one customer. They answer one question: do you make more from a customer than it costs to win and serve them? If the answer is no, growth makes the problem bigger.
This guide gives the formulas, works through examples for a SaaS product and a low-priced app, shows how much the answer changes with churn, and explains how to do the sums before you have any revenue.
The four numbers
| Number | What it tells you | Formula |
|---|---|---|
| Customer acquisition cost (CAC) | What it costs to win one paying customer | Total sales and marketing cost ÷ new paying customers |
| Lifetime value (LTV) | The gross profit one customer brings over the time they stay | Monthly revenue per customer × gross margin × months they stay |
| LTV:CAC ratio | Whether a customer is worth more than they cost | LTV ÷ CAC |
| Payback period | How long until a customer has repaid their acquisition cost | CAC ÷ (monthly revenue per customer × gross margin) |
Two notes on the inputs:
- Months they stay is usually estimated as 1 ÷ monthly churn. If 4% of customers cancel each month, the average customer stays about 25 months.
- Gross margin is what is left of revenue after the direct cost of serving the customer: hosting, payment fees, support, and any usage-based costs such as messaging or AI.
A worked example: a SaaS product
Take a product at $99 a month with a 75% gross margin, where 4% of customers cancel each month.
Lifetime value
- Months a customer stays: 1 ÷ 0.04 = 25
- Gross profit per month: $99 × 0.75 = $74.25
- LTV: $74.25 × 25 = $1,856
Customer acquisition cost
In one month the company spends $6,000 on ads and $2,000 on fees and tools, and wins 20 paying customers.
- CAC: $8,000 ÷ 20 = $400
Ratio and payback
- LTV:CAC: $1,856 ÷ $400 = 4.6 to 1
- Payback: $400 ÷ $74.25 = 5.4 months
On these numbers every customer is worth more than four times what they cost, and the cost comes back in under six months. That is a business you can put more money into.
Churn changes everything
Keep the same product and change only the share of customers who cancel each month.
| Monthly churn | Months a customer stays | LTV | LTV:CAC at $400 CAC |
|---|---|---|---|
| 2% | 50 | $3,712 | 9.3 to 1 |
| 4% | 25 | $1,856 | 4.6 to 1 |
| 8% | 12.5 | $928 | 2.3 to 1 |
| 15% | 6.7 | $495 | 1.2 to 1 |
Nothing about acquisition changed between the first row and the last. The same ads, the same cost, the same price. Retention alone moved the business from excellent to barely breaking even. This is why we look at what happens after sign-up before recommending more ad spend: finding the aha moment and fixing onboarding often improves unit economics more than any change to the ads.
What "good" looks like
These are rules of thumb, not laws.
- LTV:CAC of about 3 to 1 is the usual target. Below 1 to 1 you lose money on every customer.
- Payback under 12 months is the common benchmark for subscription businesses. With little funding, aim for much less, because you are financing every month of payback yourself.
- Early numbers are rough. With twenty customers and three months of history you do not know your churn. Use a cautious estimate and update it. Before product-market fit, these figures tell you what to fix, not how fast to grow.
Count the whole cost
Founders often quote the cost per customer from the ad platform. That is one part of CAC.
| Version | What it includes | Use it for |
|---|---|---|
| Paid CAC | Ad spend ÷ customers from ads | Comparing campaigns and channels |
| Fully loaded CAC | Ads, agency or freelancer fees, salaries, tools ÷ customers from that work | Deciding whether the channel pays |
| Blended CAC | All sales and marketing cost ÷ all new customers, including referrals and word of mouth | Judging the business as a whole |
Two things follow.
A fixed fee weighs more as spend falls. If you pay a fixed monthly fee for someone to run a channel and then halve the ad budget, the fee is now a much bigger share of each customer's cost. A founder we work with made exactly this point when reviewing a channel: the cost of a customer was the ads plus the management fee, and at a low budget the fee could not be justified. He was right, and we stopped the channel.
Blended CAC can hide a weak channel. Referrals cost almost nothing, so they pull the average down. Look at the paid figure for each channel as well, or free customers will make an expensive channel look fine.
Break it down by segment and channel
An average hides the detail you need. In one account we manage, the cost of a lead in one industry segment fell by more than a third in a week, while two other segments stopped producing leads at all. The overall figures gave little sign of it. The decision (move budget to the segment that was working and rewrite the message for the others) only showed up in the breakdown.
The same applies to channels, with one addition: judge a channel by the customers it keeps, not the sign-ups it brings.
| Channel A | Channel B | |
|---|---|---|
| Cost per install | $1.50 | $4.50 |
| Still active after a week | Lower | Higher |
| Which is better? | You cannot tell yet | You cannot tell yet |
This is a real pattern from an app we worked on. One channel brought installs at a third of the price. The other brought fewer people who came back more often. Neither figure answers the question until you know what an active user is worth. Until then, the honest position is that you do not know, and the next job is to find out. Our comparison of AppsFlyer and Mixpanel explains the tools that connect a channel to what users do afterwards.
Low-priced subscriptions and consumer apps
A cheap subscription leaves very little room, and the sums are worth doing before you build.
Take an app at $50 a year.
- App stores keep 15% to 30% of subscription revenue, so you receive roughly $35 to $42.
- If it costs $30 in ads to win one subscriber, you have $5 to $12 left from the first year.
- That has to cover hosting, support, the team and every user who cancels after a month.
If a subscriber pays monthly and leaves after the first month, you have spent $30 to earn a few dollars. On a low price, a customer who churns early is a loss, not a smaller profit.
An install is not a customer
For apps the cost builds up at each step. With illustrative numbers:
| Step | Rate | Cost per person at this step |
|---|---|---|
| Install | $3 | |
| Creates an account | 1 in 2 | $6 |
| Still active after a week | 1 in 4 | $24 |
| Pays | 1 in 20 of those | $480 |
A $3 install sounds cheap. A $480 paying customer on a $50 plan does not work. The same caution applies to headline user numbers: an app can have tens of thousands of installs and a small fraction of that in people who open it each month. Count the active ones.
Check how others in your market really make money
A founder we spoke with set his price by taking a large competitor's subscription and halving it. Two problems. First, that competitor had millions of users and several sources of income, so its price said little about what a new entrant could charge. Second, his real competition was free: search, AI assistants, social groups, friends. When the alternative costs nothing, the question is not whether your price is lower than a rival's. It is whether having everything in one place is worth paying for at all.
Advertising revenue has the same catch. Advertisers pay for audience size, so it only becomes real income once you have a large number of active users, which you have to pay to acquire first.
Before you have revenue
You cannot measure LTV with no customers. You can still work out whether the plan is plausible.
1. Work out the most you can afford to pay
Start from the price and work backwards.
- $99 a month, and a cautious guess that a customer stays 12 months: $1,188 in revenue
- At a 75% gross margin, that is an estimated lifetime value of about $890
- Divide by three, to match the 3 to 1 rule of thumb: about $300
That $300 is your allowable CAC. It is a budget for one customer, and you can spend it in different ways: on ads that send people to a self-serve trial with good onboarding, or on cheaper leads plus a person who demonstrates the product and closes the sale. Price decides which is realistic. A low-priced product cannot pay for a salesperson, which is why it has to sell itself.
2. Find out what it costs to reach someone who pays
Asking people is a start. If you ask 200 people in your target group and one says they would pay, the next question is what it costs to reach 200 of them. If that is more than your allowable CAC, the plan does not work as it stands.
Saying yes is also not paying. A survey can confirm the problem. It does not confirm the price. The reliable test is a payment: a pre-order, a paid pilot or a discounted early sign-up. The SaaS idea validation playbook covers how to ask.
3. Test cheaply before you build
A landing page and a small ad test can tell you whether strangers will pay, for a small fraction of what a full build costs. We have worked with founders who were about to commit a six-figure sum to development and learned, from a test costing a few thousand dollars, that people had the problem and would not pay to solve it. That is a good result. It is the cheapest point at which to find out.
It is also worth asking whether an app is the cheapest way to test the idea. A paid group on an existing platform, or a simple web page with a payment link, can prove the business model before any code is written. What Is a Minimum Viable Product? covers the options, and Marketing Budget for a Tech Startup Without Revenue gives realistic test budgets.
How to improve the numbers
In rough order of effect for most early products:
- Keep customers longer. See the churn table above. Onboarding and lifecycle messaging are usually the cheapest lever.
- Sell to the right customer. A narrow, well-chosen first segment converts better and stays longer. See What Is an ICP?
- Fix conversion before buying traffic. A landing page that converts twice as well halves CAC.
- Raise the price, or the value per customer. Annual plans, higher tiers and add-ons all lift LTV. A price framed against the outcome is easier to raise; see our guide to the positioning statement.
- Cut the channels that do not pay, using fully loaded cost.
Common mistakes
- Counting ad spend only. Fees, salaries and tools are part of the cost.
- Using revenue in LTV instead of gross profit. It flatters every ratio.
- Trusting a lifetime you have not observed. Three months of data cannot support a five-year lifetime.
- Judging channels by sign-ups or installs. Follow them through to retained, paying customers.
- Copying a large competitor's price. Their costs, scale and income are not yours.
- Doing the sums after the build. They take an afternoon and can save a year.
Common questions
What are unit economics?
Unit economics are the revenue and costs of a business measured for one unit, which for most startups means one customer. They answer a single question: does the business make more from a customer than it spends to win and serve them? The main figures are customer acquisition cost (CAC), lifetime value (LTV), the ratio between them and the payback period.
What is a good LTV to CAC ratio?
Three to one is the common rule of thumb: a customer should be worth about three times what it cost to win them. Below one to one you lose money on every customer. A very high ratio can mean you are spending too little on growth. Early-stage figures are rough, so treat the ratio as a direction, not a pass mark.
What is a good CAC payback period?
For a subscription business, under twelve months is the usual benchmark, and shorter is better when cash is tight. A startup without much funding should aim to recover acquisition cost within a few months, because every month of payback is cash it has to fund.
How do you calculate CAC?
Add up everything spent on winning customers in a period (ad spend, agency or freelancer fees, sales and marketing salaries, tools) and divide by the number of new paying customers in the same period. Counting ad spend alone gives a flattering number that will not match your bank balance.
How do you work out unit economics before you have revenue?
Work backwards. Take your planned price, margin and a cautious guess at how long a customer will stay to estimate lifetime value, then divide by three to get the most you can afford to pay for a customer. Compare that with what your first small test shows it costs to reach someone who will pay. If the gap is large, change the price, the customer or the channel before building more.
Where Gufy fits
Gufy helps SaaS and mobile-app startups see these numbers clearly and improve them: tracking and product analytics that connect a channel to retained, paying customers, paid acquisition judged on fully loaded cost, and onboarding and lifecycle work that lengthens the time customers stay.
If you are spending on growth and cannot say what a customer costs or what they are worth, bring the numbers you have.