A repeat direct customer is worth the contribution from their later orders: the money left after the variable costs of making and fulfilling those orders. That value depends on how much they order, what each order costs to serve, and how often they return within a reasonable time horizon.
Start with one typical repeat order, then estimate how many later orders a group of first-time customers might place. Use a range rather than a single lifetime-value number. The result can help you decide how much to spend bringing guests back or acquiring new direct customers, but it is not profit until fixed costs and other excluded expenses are covered.
Define the repeat-order cohort
A cohort is a group of customers who share a starting point. For this decision, use guests who placed their first direct order during the same month. Count only their later direct orders in the repeat-order model. Their first orders have their own revenue and costs and should be evaluated separately.
Set a follow-up period before counting. Six months can be a useful starting point for a restaurant with frequent orders; a seasonal operation may need a longer window. The point is to compare customers who have had roughly the same opportunity to return, rather than mixing last week's new guests with people who have ordered for years.
Decide what makes an order direct. An order placed through your own restaurant website may qualify; an order placed through a marketplace generally belongs in a separate channel calculation. Also decide how you will identify returning customers. A phone number, email address or account may help, but guests can use different details or share an account. Ask your ordering provider what customer and order data you can export and how repeat customers are identified.
Gather a short list of inputs:
- Order and customer records: first-order dates, later order dates, sales, guest-paid charges and discounts. Your ordering records or POS may provide these.
- Food and packaging costs: typical ingredient and container costs for the orders in the cohort. Use supplier invoices and recipe or portion estimates.
- Fulfillment costs: extra labor, delivery or dispatch charges, and payment processing. Use schedules, invoices and processor statements.
- Retention assumptions: the share returning and the timing of later orders. Start with your own history where available; otherwise, test a cautious range.
Keep pickup and delivery separate if their costs differ materially. A single average can conceal the fact that a profitable pickup reorder is subsidizing a weak delivery order.
Subtract the costs of each later order
Start with food and drink sales after discounts, add any guest-paid delivery charge, and subtract the costs that arise when the order is placed. Include ingredients, packaging, incremental labor, processing and the amount charged for delivery or dispatch. If a discount has already reduced sales, do not subtract it a second time.
Hypothetical example: A repeat delivery order has $25.75 in food and drink sales before a $1.75 discount, leaving $24 in net food sales. The guest also pays a $2 delivery charge. The restaurant pays $5 for delivery, so the delivery cost net of the guest charge is $3. The remaining assumed variable costs are $7 for food, $1.25 for packaging, $2 for incremental labor and $1 for payment processing.
That produces $8 of contribution: $24 in net food sales minus $7, $1.25, $2, $1 and $3. Equivalently, total guest payments of $26 minus $18 in variable costs also leave $8. The $2 charge helps offset delivery expense, but it is still a cost to the guest and could affect whether they reorder.
This example includes ordinary variable costs and the discount. It excludes fixed costs such as rent and subscriptions, and exception costs such as refunds, remakes and support time. It also assumes the delivery and processing figures are the restaurant's actual costs, not just advertised rates. Use your own invoices and include any per-order fees that apply.
Labor deserves particular care. If an extra order requires extra paid preparation or packing time, include it. If existing staff can handle it without changing paid hours, do not automatically assign the full hourly wage as an incremental cost. Revisit that assumption when order volume requires another shift or a longer schedule.
For an order that receives a refund or remake, calculate the actual result separately. Then estimate how often those exceptions occur and include an allowance when you make a spending decision. A $1 processing charge, for example, is not the all-in cost of operating the ordering channel.
Model frequency over a defined horizon
Contribution per order answers only part of the question. A guest who returns once creates a different amount of contribution from a guest who returns several times. Use a period you can observe, and count actual orders rather than assuming every customer returns at the same pace.
For a simple forward estimate, divide the follow-up period into three reorder opportunities. In the hypothetical scenario above, assume each later order contributes $8. Also assume that a fixed percentage of the customers who ordered in one opportunity return in the next. This is a simplifying model, not a prediction of individual behavior.
For example, at a 60% return rate, 100 first-time customers generate an estimated 60 orders in the first opportunity, 36 in the second and 21.6 in the third. That is 117.6 expected later orders, or $940.80 in contribution. Divided by the original 100 customers, it is $9.41 per acquired customer, rounded to the nearest cent. Fractional orders represent an average across many comparable cohorts, not a literal partial order.
Real customers may skip a period and return later, order more than once in a period, or change their basket size. The model does not capture those patterns. If you have reliable history, replace the repeated percentage with observed order counts for each period. Keep the time horizon visible so a six-month estimate is not mistaken for value over a customer's entire relationship with the restaurant.
Test retention before setting a spending limit
The table applies the same hypothetical $8 contribution and three-opportunity horizon to 100 first-time customers. Each rate is applied to the group that ordered in the preceding opportunity. The first order is excluded, as are fixed and exception costs.
| Return rate per opportunity | Expected later orders | Later-order contribution | Contribution per first-time customer |
|---|---|---|---|
| 40% | 62.4 | $499.20 | $4.99 |
| 60% | 117.6 | $940.80 | $9.41 |
| 80% | 195.2 | $1,561.60 | $15.62 |
The range is wide because a change in return rate compounds across opportunities. The low case is not necessarily your forecast; it is a useful stress test. If your historical results support something closer to 40% than 80%, do not budget as though the high case is assured.
To use the estimate for acquisition, combine it with the separately calculated contribution from the first order. Then subtract the acquisition cost and any fixed or exception costs you want the decision to cover. Keep customer acquisition spending distinct from a discount used on a later order; otherwise, the same incentive may be counted twice.
For example, if a first order contributes an assumed $3 before acquisition spending, the low repeat scenario adds about $4.99, for roughly $7.99 before fixed and exception costs. A $10 acquisition cost would exceed that amount. Under those assumptions, do not approve that spend on repeat value alone. A higher spend would need support from other measured contribution or a different cost structure, not just the hope that retention improves.
Update the model when menu prices, portions, delivery charges or incentives change. A promotion may bring guests back but reduce contribution per order. Compare the additional contribution from the orders you reasonably expect with the full cost of the promotion; do not treat every discounted order as proof that the discount caused it.
If you are also comparing where first orders originate, the guide to comparing direct and marketplace contribution covers that separate decision. The online ordering buying guide can help you evaluate the tools and reports needed to track direct orders.
Next step: Pull one month of first-time direct customers, calculate the contribution on a representative later pickup and delivery order, and check how many customers returned over your chosen horizon. Use the checklist above to evaluate your current setup, then ask Fleksa about restaurant online ordering and verified US availability if it fits your needs.
Frequently asked questions
Is repeat customer value the same as order revenue?
No. Revenue is what the guest pays. Repeat-order contribution subtracts the variable costs of those later orders; it is not final profit because fixed and other excluded costs still need to be covered.
Should I include the first order?
Keep it separate when measuring repeat value. Add its contribution back only when evaluating the total economics of acquiring a new customer, and subtract acquisition spending separately.
What if I cannot identify every returning guest?
Use the records you can match consistently and treat the result as an estimate. Ask your provider how repeat customers are identified and exported, and check whether account sharing or different contact details could distort the count.
