Blog
Referral marketing: how to design a program that pays off
Referral marketing means your current customers recommend your business to people they know, usually in exchange for a benefit given once the referred person becomes a customer. A referral program gives it a structure: who can refer, what each side gets, when it’s paid and how each referral is recorded.
It differs from affiliate marketing in who does the recommending. An affiliate promotes for a commission, often to an audience they don’t know. In a referral program, customers who already bought recommend you to people around them, and that relationship is part of what makes it work.
What research shows about referred customers
A study published in the Journal of Marketing followed about 10,000 customers of a German bank for almost three years. Schmitt, Skiera and Van den Bulte found that customers who came through the referral program had a higher contribution margin, though the difference shrank over time, and a higher retention rate, a difference that persisted. Their average value was at least 16% higher than that of non-referred customers with similar demographics and time of acquisition. The difference varied across segments, which is why the authors recommend a selective program.
A later study by the same researchers explained why it happened at that bank. Referrers know the business and their contacts, so they tend to bring in people who fit it better: the referrer’s margin and the referral’s margin were related, and more experienced customers brought in higher-margin referrals. Referred customers also churned less only as long as the person who referred them remained a customer. Their practical recommendation was to recruit referrers among customers acquired at least six months earlier, with high margins and a low risk of churning.
These are results from one bank and its program. They show mechanisms worth using, not a figure that will repeat in every business.
The decisions behind a program
| Decision | Options | What to consider |
|---|---|---|
| Who gets rewarded | Only the referrer, only the new customer, or both | A benefit for the new customer lowers their risk in trying you; one for the referrer gives them a reason to do it |
| What the reward is | Discount, credit, product or cash | Record what it costs you: an extra gift costs what it takes to make and ship it; a discount or credit that reduces a payment the customer was going to make anyway costs its full value |
| When it’s paid | At sign-up, after the first purchase, or after the return window | Pay once the condition has already left you margin; that way you don’t reward fake accounts or returned orders |
| Who you invite to refer | Everyone, or customers with tenure and good margins | The studies cited point to stable customers: their referrals left more margin and stayed longer |
| Which rules prevent abuse | A cap per person, no self-referrals, no repeated address or payment method | Write them into the terms before launch; fixing them later annoys honest customers |
| How it’s disclosed on social media | Stating that the referrer receives a benefit | In the US, the FTC requires clear disclosure of that material connection; check the advertising rules in your country |
How much you can pay per referral
The cost is the easiest part to underestimate. The reward is paid for every referral that meets the condition, including people who would have bought anyway because they were already planning to. If half the referrals would have come regardless, each customer the program actually brings in costs twice the reward.
Fictional case. A coffee subscription estimates that a new customer leaves USD 60 in contribution margin over their first twelve months, before rewards. It offers the referrer a free bag of coffee, on top of their subscription, that sells for USD 12 in the store and costs the business USD 8 including shipping. The new customer gets 10% off their first USD 40 box, that is, USD 4. Both benefits depend on the same purchase: the first box, paid and not returned. It estimates that half the referrals would not have bought without the program.
- Cost per referral: 8 + 4 = USD 12.
- Cost per incremental customer: 12 ÷ 0.5 = USD 24.
- Margin left per incremental customer: 60 − 24 = USD 36.
- Maximum rewards per referral to break even: 60 × 0.5 = USD 30.
Referral reward calculator
Work out what each customer the program actually brings in costs you, and how high the reward can go. Use margin over one period before rewards, for example a customer’s first 12 months.
The hardest input to estimate is the incremental share. Asking at checkout whether the person would have bought without the recommendation gives a first clue, although answers tend to be optimistic. If you can, compare periods or regions with and without the program. Until you measure it, use a cautious share: at 25%, the same program still leaves USD 12 per incremental customer.
A complete example of program terms
Here’s how the fictional subscription’s program could read:
- Who can refer: subscribers active for at least six months.
- What each side gets: the new customer, 10% off their first box; the referrer, a free bag of coffee.
- When it’s granted: the discount applies when the first box is paid; the gift ships once that box is past the return window.
- Limits: up to five gifts per person per year; a new account with the same shipping address or card doesn’t count.
- Tracking: a unique code per subscriber, stored with every order.
- Social media: if someone shares their code publicly, the terms ask them to say they receive a gift for each new subscription.
With the code stored on every order, the subscription can see after six months whether referred customers stay longer than everyone else, as we explain in cohort analysis. That isn’t enough to set the reward: at twelve months, compare the referred customers’ actual contribution margin over that same window with the USD 60 you assumed, and revisit the incremental share. Only if both support it is a bigger reward worth it. The customer lifetime value of each group helps with that comparison.
Plan your next post in HeyMark.
Keep the idea, review the draft with your team, and see how it performed in the accounts you connected.