Unit economics: CAC, LTV, and payback

Before you fund growth, prove that one customer — one order at a time — is actually worth more than it costs to win.

The idea

Unit economics is the business shrunk to a single customer. Start at one transaction: revenue minus the costs that vary with that transaction gives contribution margin — the cash each order actually leaves behind. Multiply by how often a customer buys, then by how long they stay, and you get LTV, their lifetime value.

Set LTV against CAC, what you paid to acquire them. Two numbers decide if growth is worth funding: the LTV:CAC ratio (rule of thumb: aim for ~3:1 or better) and the payback period (how many months until the customer has repaid their CAC). And keep profitable per order separate from profitable per customer — they can disagree.

How it works

Build it in two stages — the order, then the customer. Every arrow is real cash; nothing here is top-line revenue dressed up as profit.

per order
  fees collected                 $12.00
  - driver pay                  - $5.00
  - support & processing        - $1.50
  = contribution / order         = $5.50

per customer
  orders / month                 4
  monthly contribution           $5.50 x 4      =  $22.00
  monthly churn                  8%   ->  lifetime = 1 / 0.08 = 12.5 mo
  LTV = monthly contribution / churn         = $22 / 0.08 = $275
  CAC                                          =  $60

  LTV : CAC  = 275 / 60                    =  4.6 : 1   (≥ 3, fundable)
  payback  ~ CAC / monthly contribution     = 60 / 22  ~  2.7 mo
            (churn makes the true payback a touch later, ~2.9 mo)

The two right-hand numbers answer different questions. LTV:CAC asks “is a customer worth more than they cost, ever?” Payback asks “how long is my cash tied up before I get it back?” A business can pass one and fail the other.

When to use it

Reach for unit economics when…	The trade-off / limit
Deciding whether to pour money into paid acquisition — LTV:CAC and payback say if the math funds itself.	Needs a believable churn number; early on, lifetime is a guess and small churn errors swing LTV hugely.
Comparing acquisition channels — per-channel LTV:CAC shows which to scale and which to cut.	Blended figures hide a broken channel inside a healthy average.
Sanity-checking a “grow now, profit later” plan — payback shows how long you carry the cost.	Contribution isn’t profit: it ignores fixed costs, overhead, and the time value of money.

Watch out for

• Revenue in place of contribution margin. Using top-line revenue (not revenue minus variable cost) inflates LTV several-fold and makes a losing model look fundable.
• Blended CAC. Averaging cheap organic signups with expensive paid ones flatters the paid channel. Split CAC by channel or you’ll scale the wrong one.
• Churn treated as fixed and low. New cohorts churn faster than mature ones. Borrowing a seasoned churn rate overstates lifetime and LTV.
• Confusing per-order with per-customer. A healthy order margin still loses money on every customer if CAC exceeds LTV — and a negative order margin can’t be rescued by volume.
• Ignoring payback and cash. A glittering 5:1 ratio with a 30-month payback can still run you out of cash before the lifetime value ever arrives.

Worked example

In a case interview: “A food-delivery startup is growing fast but burning cash. Are the unit economics sound?” Don’t reach for growth rate — build the unit. Fees per order about $12; subtract driver pay ($5) and support and processing (~$1.50) for roughly $5.50 contribution per order. A customer orders ~4 times a month, so ~$22/month; at 8% monthly churn they stay ~12.5 months, giving LTV ~ $275. If CAC is $60, that’s a 4.6:1 ratio and a ~3-month payback — the units are sound, and the burn is a funding-of-growth story, not a broken-economics one. But if CAC were $300, each customer would be worth less than they cost: profitable per order, unprofitable per customer — and no amount of volume fixes that.