Customer Lifetime Value for Takeaways

Customer lifetime value is an estimate of future contribution, not a permanent label attached to a person. For an independent takeaway, a simple and transparent model is usually more useful than a complex prediction. The purpose is to set sensible acquisition and retention budgets while avoiding decisions that punish new or infrequent customers.

Start with contribution, not revenue

For each completed order, subtract variable costs that change with the order: ingredients, packaging, payment charges, channel commission, delivery burden, funded discount and expected refund or remake cost. The result is order contribution before fixed overhead allocation.

An illustrative historical calculation is:

Observed customer contribution = completed-order contribution across the period − customer-specific incentive and service-recovery costs.

Do not call this profit unless fixed costs and the chosen accounting treatment are included.

Define identity cautiously

A customer may order as guest, use multiple phone numbers, share an address or order through different channels. Automatic identity matching can combine unrelated people. Keep the matching rule documented, allow uncertainty and do not use sensitive or excessive data merely to improve the model.

Use cohorts instead of one universal figure

Choose a forecast horizon

Use a period supported by the data—often several months rather than an indefinite “lifetime”. Report observed value separately from projected value. State the assumptions about reorder frequency, contribution and churn, and show a conservative range.

Use CLV for budgets, not customer service

CLV can help decide how much to spend on a broad acquisition route or retention programme. It should not determine whether a customer receives a remedy for a valid failed order. Refund and complaint decisions should follow the facts and applicable rights, not a hidden value score.

Controls

• Exclude cancelled, fraudulent and duplicate orders consistently.
• Reflect refunds, chargebacks and loyalty liabilities.
• Recalculate after major pricing, menu or delivery changes.
• Keep observed and projected values separate.
• Document identity matching and data sources.
• Allow analysts and managers to challenge unexpected results.

Practical next step: calculate six-month observed contribution for one customer cohort using a low, central and high identity-matching assumption. If the result changes dramatically, improve data quality before using CLV for spending decisions.

Use scenarios rather than false precision

Show a conservative, central and optimistic case. Vary reorder frequency, contribution and churn instead of reporting a single value to the penny. Include a separate line for loyalty rewards or future discounts that have not yet been redeemed. Where delivery economics differ substantially by zone, calculate value by fulfilment pattern rather than averaging the whole customer base.

Reconcile the model to total observed orders and contribution. If the sum of customer-level values does not match the channel report within an explained tolerance, investigate identity matching and missing refunds before using the forecast.

Related guides

• Restaurant Marketing and Customer Retention — the main guide for the wider topic.
• Measuring Marketing Performance — the broader guide that frames this implementation.
• Marketing Attribution for Restaurants — a closely related operational control to review alongside this page.
• How to Measure Repeat Order Rate — a closely related operational control to review alongside this page.

Sources and date checked

Guidance checked: 24 July 2026. Platform tools and policies can change, so confirm the current process before acting.

• ICO: plan direct marketing
• ICO: legitimate interests guidance