A $20 order lands and you clear about $5 on it. Look at the business that way and spending $10 to win a customer looks insane, a $4 loyalty reward looks generous, and a $6 refund feels like theft. All three judgments are wrong. They price the order, when the thing you're actually buying is the customer, and that customer is worth something like eighty times tonight's margin.
The formula, with real numbers plugged in
Lifetime value in profit terms is four numbers multiplied:
LTV = average order value × orders per month × net margin × months retained
Take an ordinary delivery regular: $20 AOV, ordering twice a month, at a 25% margin on direct orders. That's $40 of revenue and $10 of profit per month. A converted regular (someone past their third order, using your app, enrolled in loyalty) tends to stick with a brand for years. Call it 40 months. Multiply it out: $10 × 40 = $400 of profit from one customer who spends $20 at a time. In revenue terms the same person is a $1,600 relationship.
The honest caveats before you tattoo $400 on the office wall
Retention curves aren't flat. Half of first-time customers never order twice, and the 40-month lifespan belongs to the ones who make it past roughly the third order. So compute LTV on cohorts rather than averages: take everyone who first ordered in a given month and watch their spend decay over time. Watch the margin term too. It's 25% on a direct order, but push the same customer through a 30% marketplace and it collapses, dragging LTV down with it. The appetite is identical; the value is a fraction of it, and the math behind that sits in our delivery margin breakdown. And if you'd rather not bet on the far future, cap the lifespan at 24 months. For a single-location operator that keeps the number conservative and still lands around $240 per regular.
Bring three months of order history. We'll compute your cohorts on a 30-minute call.
What changes when you price customers, not orders
Acquisition budgets stop feeling reckless. Against a $5 order margin, a $12 cost per acquisition is a loss; against a $400 LTV it's a 30× return, provided you retain the customer. That's how chains outspend independents on ads without bleeding money. They know the denominator.
Refunds turn into retention spending. A $6 refund on a late order buys back goodwill on a $400 asset, which makes it one of the cheapest retention tools you have. Resolve the complaint quickly and that customer often stays longer than one who never hit a problem at all.
Loyalty math flips too. A 10% cashback reward that lifts frequency from 2.0 to 2.3 orders a month adds roughly $50 of lifetime profit per regular, for a fraction of that in rewards paid out.
How to measure it without a data team
You don't need a warehouse of dashboards for a usable LTV. The minimum version takes an afternoon. Pull every customer whose first order landed 6–12 months ago; that's one cohort. Count their total orders and revenue since, then divide by the cohort size. What you get is realized value per acquired customer, with no forecasting involved. Run the same sum for customers who reached a third order and you'll see the value of a converted regular, which usually runs several times the blended average.
Two habits keep the number honest. First, measure in profit rather than revenue: apply your real per-order margin by channel, since a marketplace-heavy cohort is worth far less than its top line suggests. Second, track payback period next to LTV. If a customer costs $12 to acquire and contributes $10 a month, you've recovered the cost in five weeks. Payback under two months means you can reinvest hard; payback over six means fix retention before you scale spend.
Four multipliers, four levers
The formula points straight at the work. Frequency moves fastest, and the standard tools are push notifications tied to something real (a new item, bad weather, the day someone usually reorders) and cashback campaigns. Lifespan is really an experience problem, decided by delivery speed and order accuracy, which is why delivery time belongs in your retention reporting and not just the ops dashboard. Margin comes down to channel mix: every regular you move from a marketplace into your branded app multiplies whatever lifetime profit they have left. AOV is menu work, mostly bundles and add-ons offered at checkout.
All four depend on one thing: being able to see the customer. Marketplaces keep the identity for themselves, so you get an order ID and no way to tell that tonight's #4411 is last week's #3902. Any LTV work has to start from data you actually own.
Do this next
Work out one number this week: profit per customer per month for customers with three or more orders, using your own channel data in reports and analytics. Multiply it by your average regular's lifespan, even a rough guess. Then put that number somewhere the person who answers complaints will see it every day, and watch how differently they treat a $6 refund.
Branded app, loyalty, push campaigns, and cohort analytics — launched in about two weeks.