A referral programme (“bring a friend”) turns satisfied customers into an acquisition channel. Well designed, it brings cheaper, more loyal customers. Badly designed, it attracts benefit hunters, generates costs and tires customers. The key lies in the offer, simplicity and measurement.
The essentials in a definition
A referral programme (refer-a-friend) offers benefits to those who refer and to those referred. It works when the benefit is clear and fair, the process simple and the products good. It requires transparency on conditions (UCPD), lawful handling of referred people’s data (GDPR) and measurement of invitations, acceptances and purchases with events.
When it works
When the product is good enough to be recommended, satisfaction is high and the customer has social reasons to share. It works well in services, subscriptions, e-commerce with repurchase, health and wellbeing. It does not fix a weak product: it amplifies the existing experience.
People recommend what they appreciate.
Designing the offer
A clear benefit for the recommender and the person recommended (discount, credit, product, experience). Balance cost against customer value. Simple, transparent rules: when the benefit is granted, limits, validity. Consider non-monetary benefits. Comply with advertising and data-protection law.
Both sides should gain.
Referral programmes: designing and measuring
- Ensure a product worth recommending
- Offer a benefit to both sides
- Ease sharing
- Measure quality and prevent abuse
Easing sharing
A personal link or code, one-click sharing by message, email and social networks, a pre-written message the customer can personalise, a clear landing page for the invitee. Ask for the recommendation at the right moment: after a good experience (delivery, result, positive review). Remind without insisting.
Friction kills referral.
A good referral programme does not create satisfaction; it amplifies it.
Measure and protect
Measure participation, invitations sent, conversion, cost per customer brought, retention and value of those customers. Prevent abuse (fake accounts, self-referrals). Adjust the offer based on results. Publicly recognise the best ambassadors. Review the programme each quarter.
Compare the quality of referred customers.
Decision table
| Element | Good practice | Risk |
|---|---|---|
| Benefit | Clear, balanced for both | Hidden conditions |
| Process | Few steps, own code or link | Excessive friction |
| Data | Only what is necessary, with clear information | Contacts used without a legal basis |
| Fraud | Rules and verification | Abusive self-referrals |
Common mistakes to avoid
- A confusing benefit. People do not take part.
- Sending messages to contacts who did not consent. It breaches privacy and communications rules.
- Not measuring. You do not know the cost per new customer.
How to measure the result
Track participation rate, acceptance rate, cost per new customer and retention of referred customers.
Limits and exceptions
Participation varies greatly by product and community. Rules on incentives and communications vary; this content is not legal advice. The effect may partly come from customers who would have come anyway.
Illustrative example
Illustrative example: a store offers both referrer and referred a discount on the next purchase, with a personal link and clear rules. It measures invitations sent, acceptances, purchases and cost per new customer. It compares the retention of referred customers with that of ad customers.
In practice: checkpoints
Test two versions of the benefit with the same audience and compare participation and cost per new customer. Ask referred people to say who invited them, to record the source. Review results every month, remove those who abuse the rules and thank the best promoters. Compare the retention of referred customers with that of others, and adjust the benefit if the cost does not pay off. Tell people clearly how the invited friends’ data will be used, and let them ask for removal. Make sharing easy with a ready-made message and link, and show progress to participants, so the programme feels simple and fair.
Frequently asked questions
What benefit should I give?
Something valued and that the business can sustain against customer value.
When do I ask for the referral?
Right after a good experience.
How do I avoid abuse?
With clear rules, limits and verification.
Which benefit works?
The one that has value for both and is simple to use; test.
How do I avoid fraud?
With rules, per-person limits and purchase verification.
Design referral programmes with a clear, balanced benefit, simple process, lawfully handled data and measurement of cost per new customer and retention.
Sources and recommended reading
- Unfair Commercial Practices Directive (2005/29/EC)
- General Data Protection Regulation (GDPR)
- Analytics Help: GA4 recommended events
Last editorial review: October 2026. Informational content; always confirm current conditions in the official sources before deciding.