A referral program needs more than an offer left running in a footer. A campaign gives the team a specific audience, a reason to ask, an owner, and a decision to make after the results come in. This guide turns those decisions into a working brief.
The seven-part campaign framework
An evergreen campaign can use the same discipline as a seasonal promotion. Define the following before changing creative or increasing the reward.
- 1. Goal
- Choose one primary outcome: activate advocates, acquire qualified new customers, encourage repeat sharing, or support a product launch. Write the decision the result will inform.
- 2. Audience
- Define who should advocate and who qualifies as a referred friend. Include exclusions, the source of the audience data, and what happens when a customer belongs to several segments.
- 3. Incentive
- Specify the advocate reward and friend offer separately. Document minimum purchase, excluded items, approval delay, expiration, and the margin guardrail.
- 4. Moment
- Choose where and when the ask appears: post-purchase, after delivery, in an account, at a loyalty milestone, or in lifecycle email. A recent unresolved support issue may be a reason to suppress the ask.
- 5. Creative
- Write the customer-facing proposition, call to action, friend landing message, and sharing copy. Explain the qualifying action without making readers hunt through fine print.
- 6. Controls
- Agree on attribution, reward approval, returns, fraud review, and support handling. A tracked referral and an earned reward are different events.
- 7. Measurement
- Set the baseline, primary KPI, guardrails, review date, and minimum evidence needed to decide. Name the person who can pause the campaign.
Five campaigns worth testing
These are planning examples, not customer results or promises of lift.
Post-purchase activation
Invite first-time buyers to refer after they have experienced the product. Test a delivery-based ask against your current timing. Watch participation and support complaints, not just clicks.
VIP advocate campaign
Give repeat buyers a more personal invitation. Test recognition or early access before assuming a larger discount is necessary. Exclude customers already receiving another referral promotion.
Seasonal acquisition
Match the friend landing page to a gifting moment or product launch. Confirm that the offer, inventory, and redemption deadline remain consistent across email, landing page, and checkout.
Subscription milestone
Ask after a successful renewal or usage milestone. Suppress failed payments and cancellations. Confirm that the qualifying subscription event actually reaches your referral integration.
Re-engaging past advocates
Make repeat sharing easy for customers who have already referred successfully. Test a fresh sharing message or placement before changing the economics of the entire program.
A worked campaign hypothesis
For first-time buyers whose order has been delivered and who have no open service issue, a post-delivery referral email will generate more qualified new-customer orders per eligible customer than the current ask, without exceeding our approved reward-cost guardrail.
Keep the friend offer unchanged, compare consistent customer cohorts, and use the same observation window. This isolates the timing decision. Changing the audience, offer, and landing page together makes the result harder to interpret.
Measure the funnel before changing the offer
Track eligible customers, exposure, advocates who share, friend visits, qualifying purchases, referral approvals, and issued rewards. Keep dates and counting rules consistent. A customer can share repeatedly, so event counts and unique-person counts are not interchangeable.
- Participation rate
- Unique advocates who shared ÷ unique eligible customers exposed. If unique exposure is unavailable, state the denominator you actually use rather than calling email sends impressions.
- Friend conversion rate
- Unique purchasing referred friends ÷ unique referred visitors, using the same cohort and observation window. Keep repeat orders separate.
- Approval and issuance
- Report approved referrals ÷ resolved referrals, and issued rewards separately. Pending referrals need their own aging view; they are not failed conversions.
- Cost and quality
- Track reward and discount cost per qualified new customer, returns, repeat purchase, and contribution margin where your data supports them. Reconcile referrals to orders before drawing conclusions.
Attributed orders do not establish incrementality. Use a credible holdout or controlled comparison when the question is whether the campaign caused additional purchases. Predefine the primary metric and avoid ending a test the first time a result looks favorable.
A practical monthly review
- Check tracking, coupon delivery, pending referrals, and data freshness first.
- Identify the largest meaningful funnel constraint. Low exposure needs a placement fix; high friend visits with weak purchase conversion needs a different investigation.
- Choose one material change and record the hypothesis, comparison group, and guardrails.
- Allow enough time for purchases, approvals, and returns to mature. Monthly review does not mean every test must finish in a month.
- Keep, revise, scale, or stop. Record the decision and what the team learned.
Copy-ready campaign brief
Complete this with marketing, ecommerce, analytics, and support before build work begins.
Campaign and owner: [name, accountable owner, backup]
Objective and hypothesis: [behavior to change, reason, primary KPI]
Audience: [source, eligibility, exclusions, overlap priority]
Experience: [trigger, placement, creative, sharing message, friend landing page]
Offer: [friend benefit, advocate reward, qualifying purchase, caps, expiration]
Controls: [tracking setup, approval timing, returns, fraud review, escalation]
Evaluation: [baseline or holdout, counting rules, observation window, cost guardrail]
Decision: [review date, launch approver, pause criteria, next action]
Before launch
- Test an eligible customer and an excluded customer.
- Complete the advocate-to-friend journey on desktop and mobile.
- Confirm checkout recognizes the coupon and enforces the intended restrictions.
- Verify a qualifying purchase, an unqualified purchase, and a returned order.
- Check that support can explain pending, approved, and unrewarded referrals.
- Confirm reporting separates the test from existing campaigns.
Implementing the brief in Talkable
Talkable’s custom properties documentation explains how supplied customer data can be used in join, incentive, and placement criteria. Those fields must be available in your integration; writing them into a brief does not create them.
The segments documentation covers traffic-source reporting and purchase segments. Use these to plan reporting labels, then confirm the actual values arriving on test records.
Review referral tracking methods and referral approval with your implementation owner. Tracking, referral validity, and incentive qualification are separate parts of the process.
Frequently asked questions
Does every campaign need an end date?
No. An evergreen campaign needs a review date and an owner. A limited-time offer also needs clear customer-facing expiration terms.
What should we test first?
Fix broken tracking or redemption first. Then test the largest funnel constraint that you can change without introducing several competing explanations.
Should every customer receive the same offer?
Not necessarily. Start with a manageable number of experiences and introduce segmentation only when it changes timing, eligibility, creative, placement, or incentive.
Is referred revenue enough to judge success?
No. Check costs, customer quality, returns, and the limits of attribution. Incrementality requires a comparison designed to answer that question.