All calculators

Guide · E-commerce

Referral Rewards: Count Both Sides of the Offer

Budget referrer and new-customer rewards together, verify qualifying orders, and compare referred contribution with a stated natural-sales baseline.

By PriceAndProfit Editorial Updated 3 min read

On this page
  1. Cost one qualifying referral
  2. Move from clicks to retained orders
  3. Compare with naturally occurring purchases
  4. Identify the threshold being reported
  5. Build controls around the event ledger
  6. Use evidence for expansion decisions

A referral offer can reward the existing customer, the referred customer, or both. Calculate the cost of the complete qualifying event rather than treating the advertised reward as the whole program cost. A referral click, registration, completed order, and retained order are different events. The business needs a clear measurement rule linking rewards to the event actually used in the offer and to the final economic result of that order.

Cost one qualifying referral

Suppose a hypothetical referred customer buys an $80 order with $46 of variable product, fulfillment, and payment costs before promotional rewards. Pre-reward contribution is $34. The new customer receives an $8 price reduction, the referrer receives a $10 cash reward, and verification costs $2 per qualifying event. Contribution after those items is $14. The $8 reduction is already inside net revenue if the order ledger records $72, so do not subtract it a second time.

Contribution per qualifying referral = pre-reward order contribution − new-customer concession − referrer reward − verification cost
Hypothetical program; subtract $600 setup separately
ItemPer qualifying eventAcross 120 events
Before rewards $34 $4,080
Customer concession −$8 −$960
Referrer reward −$10 −$1,200
Verification −$2 −$240
After variable program costs $14 $1,680

Move from clicks to retained orders

Assume 150 referred checkouts occur, but 30 do not become qualifying retained orders under the explicitly stated scenario. The remaining 120 events generate $1,680 contribution after the variable reward costs. A $600 program setup expense leaves $1,080. If some rewards are paid on events that later fail qualification and cannot be recovered, record those payments separately instead of silently dropping them with the unqualified orders. The example does not establish any platform or legal reward rule.

Compare with naturally occurring purchases

Suppose a baseline model estimates that 50 of the 120 qualifying customers would have purchased at the regular price without the program. Their ordinary contribution would have been 50 times $34, or $1,700. Against that baseline, the program result of $1,080 is $620 lower. This does not mean the other 70 referrals are worthless; it shows that their contribution must also fund rewards given to customers whose purchases were not additional.

Identify the threshold being reported

Covering the $600 setup expense at $14 per qualifying event requires 42.857 events, or at least 43 whole events. That is a program-cost threshold before any natural-sales baseline. To preserve the $1,700 baseline and cover setup, the same model needs 165 whole qualifying events, because ($1,700 + $600) divided by $14 is 164.2857. Keep both thresholds labeled; describing the smaller number as proof of incremental profitability would be misleading.

Build controls around the event ledger

Record the referring account, referred customer, qualifying order, reward type, reward status, and any reversal. Investigate duplicate identities and rewards that attach to the same customer through more than one channel. Verification work can be a real operating cost even when no reward is eventually paid. Compare customer cohorts at the same age when repeat purchases matter, and include ongoing referral communications or support in the cost boundary.

Use evidence for expansion decisions

Separate observed qualifying events from forecast conversions and retention. A large number of referral links does not establish a lower acquisition cost, and reward participation does not prove that the offer caused a purchase. Test contribution against a suitable baseline where practical, and retain uncertain natural-purchase shares as scenarios. All reward amounts here are hypothetical inputs. The worksheet evaluates the entered economics without promising growth or prescribing the terms a business should offer.

Method references: OpenStax: break-even sensitivity analysis; Google Ads: conversion lift. Accessed October 22, 2026. All dollar figures are hypothetical worked examples.

Related calculators & guides