Retention & LTV

Referral Programs for eCommerce: Rewards, Fraud Controls and Measurement

Build an eCommerce referral program that pays for itself: size rewards from contribution margin, block fraud, disclose rewards and test incrementality.

TL;DR
  • Size the reward from contribution margin and a referred customer's expected repeat profit, then compare the cost per incremental customer with your paid acquisition cost, not with zero.
  • Release referrer rewards only after the friend's order ships and the return window closes, and match addresses, payment details and devices to stop self-referrals.
  • Measure referrals with a holdout or a staggered launch. Many referred friends would have bought anyway, so the true cost per new customer is higher than the dashboard shows.

What is a referral program, and when does it make sense for an online store?

A referral program rewards existing customers for bringing in new ones. The classic version is “give $15, get $15”: the friend gets a discount on a first order, and the referrer gets credit after the friend buys.

It works best when customers already talk about the product (referrals amplify word of mouth, they rarely create it), when there is a repeat purchase so store credit gets spent, and when your margin can carry two rewards per new customer.

A referral program is not a loyalty program. Loyalty rewards spend, referrals reward acquisition. Our guide to eCommerce loyalty programs covers the retention side.

How much should you pay for a referral?

Start from profit, not from what competitors offer. The question is simple: what can you afford to spend to win one new customer through this channel, compared with what you already spend through paid media?

You need four inputs:

  • First-order contribution margin: revenue minus product cost, shipping, payment fees and fulfillment on the friend’s first order.
  • Expected repeat profit: the contribution margin a new customer brings over the next 12 months, based on your own customer lifetime value data.
  • Reward cost: the friend’s discount plus the referrer’s reward, adjusted for how many referrer rewards actually get redeemed.
  • Your current customer acquisition cost on paid channels, as the benchmark to beat.

Single-sided or double-sided?

Double-sided programs reward both people. They usually convert better because the friend has a reason to act, but they cost more. Test the split: a bigger friend reward and a smaller referrer reward can produce more new customers for the same total cost.

Which reward type?

Reward type Real cost to you Strengths Watch out for
Cash or gift card Full face value Simple, highly motivating Attracts reward hunters, no repeat purchase built in
Store credit Product cost of what is bought, only if redeemed Brings the referrer back to buy Unredeemed credit can become a liability on your books, so check with your accountant
Percentage or fixed discount Margin given up on the order Easy to set up in Shopify or WooCommerce Discount codes leak to coupon sites
Free product or gift with purchase Product cost of the gift Feels generous, protects price perception Gift must be something customers actually want

Worked example: is the program cheaper than paid social?

All numbers below are illustrative.

A skincare brand has an average first order of $80. After product cost, shipping, payment fees and fulfillment, first-order contribution margin is $28. Paid social brings in new customers at a blended cost of $45 each.

The brand tests a double-sided program: $15 off for the friend, $15 store credit for the referrer.

Line item Per referred customer
Friend discount $15.00
Referrer credit ($15 x 60% expected redemption) $9.00
Referral app fees, spread across referred orders $2.00
Total program cost $26.00
Share of referred friends who would not have bought anyway (from a holdout test) 60%
Cost per incremental new customer ($26 / 0.60) $43.33

On the dashboard, referrals look almost half the price of paid social ($26 against $45). After adjusting for friends who would have bought anyway, the real gap is small ($43 against $45). The program still earns its place, but it is not the bargain it first appeared to be.

Two levers improve the math: test a $10 friend discount, or swap the referrer credit for a free product that costs $6 to make. Model the trade-offs with our LTV to CAC calculator.

When should the referrer’s reward be released?

Timing is your cheapest fraud control. Release the referrer reward when the friend’s order is shipped and the return window has closed. If you offer 30-day returns, the reward unlocks on day 31.

Customers accept a delay if you state it upfront: “Your $15 credit unlocks 30 days after your friend’s order ships.” Then send the unlock email on time. The friend’s reward is different: it has to be instant, because it is the reason the friend buys at all.

Which fraud controls does a referral program need?

Any program that pays money for new customers will attract people trying to manufacture new customers. Plan controls before launch, not after the first abuse report.

Fraud pattern Signal Control
Self-referral with a second email Same shipping address, payment card or device as the referrer Block or hold rewards when these match
Code posted on coupon or deal sites Spike of referred orders from one referrer, often with no prior relationship Cap rewards per referrer per month, use unique links instead of public codes
Buy, get reward, return the order Referred orders returned at a much higher rate than normal orders Release rewards only after the return window
Existing customers using a friend’s code Friend email or address already in your customer list Restrict the friend discount to first orders from new customers
Fake account rings Many new accounts with similar emails, same IP range, low-value orders Minimum order value for eligibility, manual review above a reward threshold

Most referral apps for Shopify and WooCommerce include some of these checks. Ask vendors which run automatically and what happens to a flagged referral (blocked, held or paid anyway).

If the return rate of referred orders runs well above your store average, something is being gamed.

How and when should you ask customers to refer?

Ask right after a good experience: when the order arrives, after a second purchase, or after a high review score. The order confirmation page comes before the customer has received anything.

  • Your post-purchase flow: add a referral email a few days after delivery. Our post-purchase email flow guide covers where it fits.
  • After a positive survey answer: a customer who rates you highly on a net promoter score question is a strong candidate.
  • Account page and package insert: a permanent link or QR code for motivated customers.

Keep email and SMS sharing compliant

Let customers share their own link through their own channels: copy link, text it from their phone, or post it. Be careful with features where your system sends the message to the friend.

If your platform emails a friend because a customer entered the friend’s address in exchange for a reward, the FTC has said that a seller offering something of value for forwarding a commercial message is generally responsible for CAN-SPAM compliance, including opt-outs. For texting, do not let customers enter a friend’s phone number to trigger marketing texts from your brand. Marketing texts generally require the recipient’s own prior express written consent under the TCPA, and a number typed in by someone else does not provide it. Our SMS compliance guide covers the consent rules.

What disclosure rules apply to referral rewards?

The FTC’s Endorsement Guides expect a material connection between an endorser and a brand to be disclosed clearly. A customer who earns a reward for recommending your product has that kind of connection. The FTC’s business guidance also says you cannot make an incentive depend on a review being positive.

In practice:

  • Pre-fill the referrer’s share message with a short, plain disclosure, such as “I get $15 credit if you order through my link.”
  • Do not tie referral rewards to positive reviews or star ratings.
  • Publish clear program terms: reward amounts, eligibility, timing and when you can cancel a reward.

This is not legal advice. Have counsel review your terms.

How do you know if the program is actually incremental?

Referral dashboards report referred orders, not how many of those friends would have bought anyway. That gap is why referral programs look better on paper than in the P&L. Three ways to measure it:

  1. Customer holdout group: randomly exclude 10% to 20% of eligible customers from all referral asks for a set period. Compare new customers referred by the asked group with new customers who name a friend as their source in the excluded group (through a post-purchase survey).
  2. Staggered launch: turn the program on for one region or one customer segment first. Compare new customer growth against the control.
  3. Survey triangulation: ask referred friends “Had you heard of us before your friend sent the link?” It is not a clean test, but it shows how much volume was already in motion.

See incrementality testing for test design.

Metric What it tells you Target direction
Share of customers who refer at least once Program awareness and appeal Up
Referred friend conversion rate Whether the offer is strong enough Up
Cost per incremental new customer True acquisition cost Below paid CAC
Referred customer 12-month contribution margin Quality of customers you are buying At or above store average
Referred order return rate Fraud and fit At or below store average

Referral program launch checklist

  • [ ] Reward sized from contribution margin and expected repeat profit, not competitor offers
  • [ ] Friend discount limited to first orders from new customers
  • [ ] Referrer reward released after shipping plus the return window
  • [ ] Self-referral checks on address, payment and device turned on
  • [ ] Monthly reward cap per referrer
  • [ ] Program terms published, with reward timing explained
  • [ ] Share message pre-filled with a plain reward disclosure
  • [ ] No system-sent texts to friends who have not opted in
  • [ ] Referral email added to the post-purchase flow
  • [ ] Holdout group or staggered launch set up before go-live

FAQ

What is a good referral reward for an eCommerce store?

There is no universal number. A good reward keeps the cost per incremental new customer below your paid acquisition cost, with referred customers as profitable over 12 months as your average. Start smaller than you think and raise it only if conversion is the bottleneck.

Should referral rewards be cash or store credit?

Store credit and free products are usually cheaper to fund and bring the referrer back to buy. Cash motivates more strongly but attracts people who only want the payout.

How long until a referral program shows results?

Expect a few months. Customers need time to receive orders and refer friends, and your holdout test needs enough new customers on both sides. Judge it on cost per incremental customer and referred customer profit, not early referred order counts.

Turn this into a plan for your store

The Free Profit Audit scans your store, sizes your biggest profit leaks and builds a 90-day plan.

Free Profit Audit