referral-program-designer

name	referral-program-designer
description	Design a referral program from scratch — mechanics, double-sided incentives, incentive math, tracking spec, launch plan. Use when the user says "design a referral program", "launch a referral", "referral incentives", "viral loop", or is planning the advocacy lever.
status	new

Referral Program Designer (Referral Squad)

Produces the mechanics, math, and launch plan for a referral program — not vague "share for a discount" ideas.

Triggers

• "design a referral program"
• "launch a referral"
• "what incentive should we offer"
• "viral loop design"
• "refer-a-friend program"

Inputs required

1. Pricing — ARPU, gross margin
2. CAC — paid and blended
3. Current referral rate — % of new customers who say "referred by friend"
4. Payment model — subscription, one-time, usage-based
5. ICP — are these people who would share (prosumer/team tools yes; enterprise often no)
6. Brand tolerance — can you do "get $20 when your friend signs up" or is that too cheap?

Process

Step 1 — Decide the program type

• Double-sided discount — both get X (e.g., Dropbox, Airbnb)
• Giver-only reward — referrer gets credit, referee gets nothing special
• Receiver-only discount — referee gets a discount, referrer gets nothing (weakest)
• Cash payout — straight cash per qualified signup (best for high-ticket)
• Gift — physical item per referral (expensive, high emotional impact)

Recommend one based on inputs, with tradeoffs explicit.

Step 2 — Do the incentive math

Constraint: incentive per customer < 30% of CAC_paid. If you beat paid CAC, you win.

Work out:

• Cost per referred customer = incentive (giver) + incentive (receiver) + platform/tooling amortized
• Expected LTV of referred customers (often 1.2–1.5× organic average — calculate from data if possible)
• Payback period
• Break-even referral rate

Output a simple model: "at X referral rate and Y payout, we add $Z MRR/month at W payback months."

Step 3 — Define qualification

What counts as a successful referral?

• Signed up — weak, easy to game
• Activated — recommended default (tied to Aha-Moment Mapper)
• Paid — strongest, most aligned with revenue
• Paid 30 days — kills fraud, for high-ticket items

Step 4 — Design the mechanics

• How do users get a referral link? (in-app, profile page, email)
• How is attribution tracked? (unique URL, promo code, hybrid)
• When does the incentive pay out? (on qualification)
• How do users see progress? (leaderboard? just a total?)
• Fraud controls — limit per user, block same-IP signups, block self-referrals

Step 5 — Write the launch plan

• Pre-launch: invite 50 top NPS users (from Advocate Identifier) to try it first
• Launch: in-app modal for all users, email blast to existing customer base
• Amplification: case study of early referral wins; social proof in pricing page
• Post-launch: weekly referral leaderboard email to participants

Step 6 — Instrumentation + metrics

Events:

• referral_link_created
• referral_link_clicked
• referred_signup
• referred_activated
• referral_payout_issued

Metrics:

• Primary: % of new customers via referral
• Secondary: referrals per referrer, activation rate of referred vs. organic, CAC_referral

Step 7 — Write the ToS / rules

• Who can refer
• Caps
• Payout timing
• Refunds / clawback if referred churns in 30 days

Output artifact

referral-program.md with:

• Program type + rationale
• Incentive math model
• Mechanics + fraud controls
• Launch plan
• Events + metrics spec
• ToS skeleton

Handoff

• Mechanics → eng / tooling (Rewardful, Partnero, custom)
• Launch email → Lifecycle Email Designer
• Invite list → Advocate Identifier
• Performance → Performance Reporter

Example call

"Design a referral program. ARPU $29/mo, gross margin 85%, blended CAC $110. Current referral rate ~5%. SaaS, self-serve, prosumer. We can offer up to $15 per successful referral."