Unit Economics Explained: Key Metrics and Formulas from CAC to LTV

What unit economics is and how to calculate it: CAC, LTV, AMPPU, C1 and COGS with formulas, a worked example, and a healthy LTV to CAC ratio.

Your marketer says CAC is $3. Your finance lead says it’s $45. They’re both right: the first one measured the cost of a lead, the second the cost of a buyer. What they’re really arguing about is a word: what to call CAC. Until the team agrees on that, you can’t compare unit economics with last month, let alone scale them.

Below is a glossary of the core unit economics metrics with formulas, from an acquired user to net profit. The examples come from Telegram lead generation, but the formulas work for any business. Akula Marketing covers the top of this funnel: the Telegram parser builds a list from niche chats, and 24/7 monitoring sends you Telegram alerts about messages like "looking for", "need" and "any recommendations".

What is unit economics, in plain terms

Users go into the product and money comes out. In between there are 3 kinds of costs, and most of the confusion around metrics comes from mixing them up:

  • Acquisition Costs: everything you spend before the first payment. Ads, outreach accounts, proxies, tools, and a sales rep’s time before the deal.
  • COGS (Cost of Goods Sold): the cost of serving a customer after the first payment. Production cost, delivery, support, fees.
  • Fix Costs: fixed costs that have a ceiling, such as the office, the product team, and accounting.

The first two grow with sales; the third doesn’t. That’s why, when you scale, the first two are what you watch.

One caveat. Different companies call the same metrics by different names: whoever first read an article on the topic set the vocabulary, and it stuck. Agree on definitions before you discuss numbers. That’s cheaper than redoing the reports later.

Funnel metrics: from user to buyer

These are people and their states. Count them as unique people, not actions.

MetricWhat it isIn Telegram lead generation
UsersUnique people you acquiredContacts who received your first message
Users new / oldFirst contact happened in the reporting period / before itOld = users − users new
LeadSomeone who left a request or contact details and whom you can keep talking toReplied with interest, agreed to a call
OrderA lead’s request or order; it can be paid, cancelled, or turn out to be a duplicateAsked for a proposal or an invoice
PaymentsPaid orders that weren’t refundedPaid invoices
BuyersPeople who paid at least onceCustomers
Buyers new / oldFirst payment in the period / before itOld = buyers − buyers new

The key rule: a lead and an order are different things. One lead can place several orders, and if you mix them up every conversion rate drifts. A lead is a person; an order is something that person does.

Sessions and clicks are not users

A session is a technical unit of analytics tools: by default a new one starts after 30 minutes of inactivity, and some tools also start one at midnight or when the UTM tag changes. Before buying, one person can rack up anything from one session to a hundred or more, while clicks are usually just over one per person. So never plug sessions into the formulas for other metrics: they’re noise, not people.

Conversions: C1 and conversion to lead

C1 is conversion to buyer: the share of acquired people who paid.

  • C1 = buyers ÷ users, averaged across all users;
  • C1 = buyers new ÷ users new, by cohort, without averaging.

Conversion to lead works the same way: leads ÷ users, or leads new ÷ users new. Keep in mind that not every lead in a month is new: someone could arrive in September and only leave a request in October.

The value of a conversion rate is in its inverse. If C1 is 2%, then 98% of people didn’t buy. Those are the people you build extra funnels for, to find where they got stuck.

Why "order conversion rate" is misleading

Many analytics tools show an "order conversion rate" by default: orders ÷ sessions. It isn’t a conversion rate. Break the formula down and you get conversion to lead multiplied by an error term: (orders per lead) ÷ (sessions per user). That error jumps from week to week and between channels, sometimes by a factor of 2–3, especially when returning customers place repeat orders.

Don’t draw conclusions from that number. Recalculate an honest conversion to lead using unique people.

Calculate C1 from all acquired users

A common mistake: each team only measures conversion on its own stretch of the funnel. Marketing reports 3% conversion to lead, sales reports 2% lead-to-customer conversion. Both look fine. But end-to-end C1 is 3% × 2% = 0.06%: 99.94% of acquired people don’t buy. It’s an ugly number for a report, but it shows at once that the funnel needs taking apart step by step.

Money per customer: AMPPU, AMPU, LTV

A unit is whatever you want to scale. In digital products it’s almost always the buyer: a paying user in B2C, a customer in B2B. Two metrics show how efficient your unit economics are: margin per buyer and margin per acquired user.

AMPPU (Average Margin per Paying User) is how much you earn from one buyer over a period, after the costs of each sale. It depends on three values:

  • Av. Price: the average amount the customer pays. If you only keep a commission, still record the customer’s full amount here: it shows whether the price matches the value they get.
  • COGS or Margin: how much goes on costs or, the other way round, how much you keep. With a revenue share model it’s easier to record margin: say you keep 20% of the price, so COGS is 80%.
  • APC (Average Payment Count): the average number of payments per buyer over the period. In subscriptions it’s called LifeTime, the number of paid months.

AMPPU = (Av. Price − COGS) × APC, or Av. Price × Margin × APC, depending on the business model. People used to calculate ARPPU instead, which is revenue without costs subtracted. AMPPU is the more honest name: it says up front that it’s about margin.

AMPU (Average Margin per User) is margin per acquired user, including those who didn’t buy:

  • AMPU = C1 × AMPPU;
  • AMPU = Gross Profit ÷ Users.

Both metrics are calculated over a fixed customer lifetime: AMPPU 30 and AMPU 30 cover the first 30 days, AMPPU 180 the first six months, AMPPU 365 the first year. Over the whole lifetime, AMPPU is called CLTV (Customer Lifetime Value) and AMPU is called LTV. The relationship is the same: LTV = C1 × CLTV.

Acquisition metrics: CAC, CPUser, CPC

  • Acquisition Costs = Users × CPUser: all scalable costs before the first payment, including VAT.
  • CPUser (Cost per User): the cost of one acquired person. In advertising it’s usually higher than CPC, because on average a person clicks more than once and you pay for every click. In outreach it’s the cost of one contact who received your message: accounts, proxies, tools and a rep’s time, divided by the number delivered.
  • CPC (Cost per Click): the cost of a click on an ad platform.
  • CAC (Customer Acquisition Cost) = Acquisition Costs ÷ buyers: the cost of acquiring a buyer. Compare it with AMPPU. As long as CAC is higher than the margin per customer, every new sale adds to the loss.

A historical mix-up: for years, mobile app teams used "CAC" for the cost of an install. That now has its own name, CPI (Cost per Install). When a colleague says "CAC", check who they mean: a user, a lead or a buyer.

Do sales costs go into acquisition or COGS?

If your sales team works separately from marketing (cold-calling a list, say), you can put its costs in either bucket: both scale, and profit comes out the same. By default, put everything spent before the first payment into Acquisition Costs. Your economics will look worse than you thought, but you’ll see the problems instead of hiding them.

Profit metrics: from revenue to net profit

  1. Revenue = Av. Price × payments: turnover before any costs. Revenue targets are the most common road to inefficient scaling: turnover grows and the money doesn’t.
  2. Gross Profit = AMPPU × buyers = Revenue − COGS: gross profit from the customer flow, before acquisition and fixed costs.
  3. Profit = Gross Profit − Acquisition Costs = Users × (AMPU − CPUser): profit from the user flow. This is the core unit economics formula. If AMPU is below CPUser, scaling only grows the loss.
  4. Fix Costs: fixed costs you can cap: rent, product team salaries, servers if they don’t grow with your customer count.
  5. Net Profit = Profit − Fix Costs: what the company keeps.

How do you tell COGS from Fix Costs?

Ask one question: what happens to this cost if sales grow 100 times? If it grows many times over with no ceiling, it’s COGS. If you can cap it, it’s Fix Costs.

An example from online education: developer and designer salaries are fixed, because you can put an upper bound on the product team. Teachers, and the managers who hire them, grow with the number of students, so they’re COGS. Servers work the same way: for a video or game-streaming service they grow with every user and belong in COGS. One of the product team’s jobs is to build the product so that as many costs as possible end up fixed.

What usually goes into COGS:

  • production cost, packaging and delivery;
  • payment processing fees;
  • support that grows with your customer count;
  • per-deal bonuses for sales reps;
  • taxes charged on every sale, such as sales tax or VAT;
  • one-off costs on the first sale: implementation, onboarding, first-month support. These are sometimes tracked separately as COGS 1st sale.

How to calculate unit economics: a Telegram outreach example

A hypothetical B2B campaign; the numbers are for illustration, not a benchmark:

  • Users: 2,000 people who received the message;
  • Acquisition Costs: $600 for accounts, proxies, tools and a rep’s time. CPUser = $0.30;
  • Leads: 80, a 4% conversion to lead;
  • Buyers: 10, so C1 = 0.5%;
  • Av. Price: $400, COGS: 25% of the price, APC over a year: 1.5 payments.

The math:

  • AMPPU 365 = ($400 − $100) × 1.5 = $450;
  • CAC = $600 ÷ 10 = $60, 7.5 times less than the margin per customer;
  • AMPU 365 = 0.5% × $450 = $2.25 against a CPUser of $0.30;
  • Profit = 2,000 × ($2.25 − $0.30) = $3,900 before fixed costs.

Hand the same numbers to each team separately and the picture falls apart: marketing sees "a lead for $7.50", sales sees "12.5% of leads bought". Unit economics combines them into one answer: is it worth acquiring the next thousand contacts? We cover how to measure the stages between a reply and a deal in our article on Telegram outreach conversion rate.

Run the Telegram parser and build a list from niche chats →

What LTV to CAC ratio is considered healthy

The benchmark you’ll see most often in SaaS and venture writing is that margin per customer should be at least 3 times the cost of acquiring them. Below 1:1, every new customer loses you money. Well above 5:1, the business is probably underinvesting in acquisition and growing slower than it could.

Compare values at the same level. In English-language articles, LTV in this ratio usually means margin per customer, which is CLTV in this glossary’s terms. So divide CLTV by CAC (both per buyer) or LTV by CPUser (both per acquired user). In the example above, AMPPU 365 is $450 against a CAC of $60, a 7.5:1 ratio within the first year.

Treat 3:1 as a guideline. With a long sales cycle or expensive servicing you need a bigger margin of safety; for a quick channel test it’s enough that first-month margin covers acquisition.

Metrics that are less useful than they look

MAU, WAU, DAU: the number of active users per month, week or day. Usually "active" means someone opened the product at least once, and for money that number says almost nothing. If you do use it, count as active only people who took a target action: exported a list, sent a request, paid.

The same goes for revenue without margin, sessions, and the "order conversion rate" from analytics tools. They can stay on the dashboard, but don’t base decisions on them.

Frequently asked questions

What’s the difference between LTV and CLTV?

CLTV is the margin from one buyer over their whole lifetime. LTV is the margin from one acquired user, including those who didn’t buy. LTV = C1 × CLTV, so LTV is smaller than CLTV unless every acquired user buys.

Can I calculate CAC from leads instead of buyers?

You can, but then it’s cost per lead (CPL), and you can’t compare it with margin per customer. CAC is acquisition costs divided by buyers, nothing else.

Over what period should I calculate AMPPU?

Over the period you’re willing to wait for payback. For quick decisions, AMPPU 30; for subscriptions and B2B with repeat payments, AMPPU 180 or 365. Compare campaigns over the same period, or a campaign measured over 30 days will lose to one measured over a year even when the results are equal.

If your bottleneck is CPUser or conversion to lead, start with list quality: the Telegram parser collects people who recently posted in niche chats, not every subscriber in sight. For why outreach goes unanswered and what to fix, see 24 reasons prospects ignore cold outreach.

Build your outreach list and track the funnel stage by stage →