Price the referral offer before you promote it

A referral offer should make an honest recommendation worthwhile while leaving enough contribution to serve the new customer; model both sides of the reward and test incrementality.

A company offers €50 to the referrer and €50 to the new customer. The campaign team calls it a €50 acquisition incentive because that is the number in the advert.

Finance sees a different commitment. Both rewards matter, as do fulfilment, fraud, administration and any discount the new customer would otherwise have paid. The programme needs a complete economic account before it needs a larger audience.

Referral design is a joint decision about customer value, behaviour and cost. The most persuasive offer is not necessarily the most profitable one.

Identify the recommendation’s natural moment

Ask when a customer has experienced enough value to make a credible recommendation. Completion of a useful project, a successful first order or repeated product use may provide a better moment than immediate sign-up.

The request should make clear who the product is for. A customer can then think of a relevant person rather than treating the programme as a reason to broadcast an indiscriminate link.

Dropbox’s programme documentation is a useful example of a defined invitation and completion journey. Your qualifying event may differ, but customers should understand the relationship between the action and the reward.

Avoid language that pressures people to overstate their experience. A recommendation should remain the customer’s honest judgement.

Build the full cost per completed referral

Use the cost of both rewards, expected redemption, payment or delivery charges, administration and foreseeable abuse. Add any onboarding effort specific to these customers.

Distinguish face value from cost. A service credit may cost less to deliver than its stated value, but it can also replace revenue the business would have earned. Finance should decide which treatment fits the actual offer.

For a hypothetical programme, assume €40 for the referrer, €30 for the new customer and €10 of additional processing and support. That is €80 before considering whether the referral is incremental.

If the first purchase contributes €60 before these costs, the programme depends on later contribution or another justified strategic benefit. The initial transaction alone does not cover the offer.

Model a conservative follow-up period

Estimate repeat contribution using relevant customer data rather than a hopeful lifetime multiplier. Show what happens if repeat purchase is lower, slower or more expensive to serve than expected.

Also show cash timing. Rewards may be paid before the new customer’s revenue is collected. A programme can have a plausible long-run case and still create a short-term cash problem at scale.

Research on referral customers at a German bank found favourable value differences in that setting. Treat such evidence as a reason to investigate your own cohorts, not a substitute for their economics.

The model should be easy to revise when the first actual customers complete the observation period.

Decide what counts as earned

Specify eligibility, the qualifying event, timing and exclusions in clear language. A reward for a completed purchase after a refund window is different from a reward for a lead, and customers should not discover that distinction after participating.

Consider both abuse and legitimate friction. Strict rules can protect the business while making the offer hard to understand or unfair to genuine users. Test the explanation with people who did not design it.

Provide a support route for missing rewards and a record that helps staff resolve disputes. Repeated manual investigation belongs in the cost model rather than being treated as free operational work.

If a business customer cannot accept a personal incentive, a company-level benefit or no-incentive introduction may be more appropriate. Check the customer’s procurement context before assuming a consumer-style offer transfers.

Make incentives visible where required

An incentivised recommendation is not always interpreted like an unpaid personal suggestion. The FTC’s endorsement guidance addresses material relationships in US advertising; other markets have their own requirements.

Have the real campaign reviewed for its jurisdictions and channels. Give participants usable disclosure language where needed, and do not design the programme around concealing the reward.

This is also a trust decision. Recipients should be able to understand why an invitation is being made without feeling that a relevant financial interest was hidden from them.

Test the additional value

Credited referrals are not automatically additional customers. Some people may already have intended to buy and simply use the reward link at checkout.

Where feasible, compare eligible groups under a planned test of the offer. Keep the observation window and outcomes fixed, and examine contribution after reward costs rather than sign-ups alone.

If a credible experiment is not possible, report the limitation and use conservative assumptions about incrementality. Avoid turning every credited order into a claim of newly created demand.

Start small enough to inspect support cases and customer fit. Expand when the programme shows a repeatable route from a fair recommendation to a valuable customer relationship. An offer that cannot survive that scrutiny should be redesigned before its promotional budget grows.