Referral mechanics that actually get people sharing
Most customers say they'd recommend you; only a fraction do. That gap is mechanics, not loyalty: the reward isn't worth the effort, the ask lands at the wrong moment, or sharing takes too many taps.
AI can design the incentive and write the share copy. A referral tool (ReferralCandy, UpPromote) mints the unique links, tracks redemptions, and holds payouts.
Fastest path: one prompt, end to end
🤖 AI prompt — paste into ChatGPT / Claude
You are a retention marketer. Design a two-sided referral program from MY unit economics only.
My numbers:
- Average order value (AOV): [$__]
- Gross margin %: [__%]
- Repeat purchase rate: [__%]
- Current customer acquisition cost (CAC): [$__]
Do this:
1. Compute my gross-margin dollars per order (AOV x margin %) and state it.
2. Propose a give-X-get-Y reward where the TOTAL cost of both sides is comfortably below my CAC. Give the advocate store credit/points; give the new friend a flat-dollar discount. Show the math that both rewards together cost less than CAC.
3. Recommend the reward TYPE for each side and one sentence why (credit keeps the advocate; a dollar discount removes first-order hesitation).
4. Write the pre-filled share message the advocate would send (1-2 sentences, warm, no hype).
5. Give the timing: exactly when to trigger the ask relative to delivery.
Output ONE markdown table | Element | My recommendation | Why | plus the ready-to-send share message in a quote block.
Do not guess or estimate any number. Use only the figures I gave you; if CAC or margin is blank, leave the sizing blank and tell me to fill it in. If you cite a referral benchmark (rate, conversion), open a real source such as referralcandy.com and name it next to the number. If you can't browse, say so and leave benchmark cells blank.
Or do it in 5 steps
- Build a two-sided (give-get) reward. Both the referrer and the new customer get something, which removes the awkwardness of handing a friend a coupon that only benefits you. The classic give-$20-get-$20 is standard because it's easy to explain and feels fair.
- Split the reward type. Give the advocate store credit or points (keeps them coming back); give the new friend a flat-dollar discount (removes first-purchase hesitation).
- Size it off your CAC. If paid acquisition costs $50, a $15/$15 give-get is still a bargain, and a referred customer usually has higher lifetime value than a cold one. Done: both rewards together cost less than your CAC.
- Place the share prompt where people already are. Account page, order-confirmation screen, not a buried settings menu. Pre-fill the message and link, and use native share sheets (SMS, WhatsApp, email) instead of forcing a copy-paste.
- Time the ask right after delivery, not at checkout. The advocate needs the product in hand and a good feeling first; waiting even 24 hours past that moment measurably lowers share rates. A common sequence: delivery confirmation → referral ask → review request a few days later.
Launch it in a tool (ReferralCandy or UpPromote) so links, redemptions, and payouts are tracked automatically.
Worked example: referral cost vs. CAC
Say your CAC is $50 and AOV is $80 at 50% margin ($40 margin/order).
| Line | Amount |
|---|
| Advocate reward (store credit) | $15 |
| New-friend discount | $15 off first order |
| Total program cost per referred customer | $30 |
| Your normal CAC | $50 |
| Saving per referred customer | $20, and the reward stays inside your store |
The $30 program cost is below your $50 CAC, and the credit pulls the advocate back for another order. So it pays for itself even before you count the referred customer's repeat purchases.
Expect a ramp while your base seeds its network. The first referred sales take time to build, so review at 90 days rather than two weeks.
Review the numbers quarterly.