Referral program mechanics designer: incentives, fraud checks, and viral loops

# Referral Program Design — Stacks (project management for agencies) **Product:** Stacks — PM tool priced at \$29/user/month **Target user:** Agency owners (5-50 employees) **Primary goal:** Reduce blended CAC from \$420 → \$280 within 6 months --- ## 1. Core Mechanic **Double-sided, credit-based referral.** Both parties get rewarded, both sides get workspace credits (not cash). Why double-sided: single-sided programs underperform by ~60% in B2B SaaS because the referrer has no immediate incentive to explain the product beyond altruism. Why credits (not cash): cash attracts fraud and low-quality referrals. Credits keep both parties inside the product, increasing activation and retention. --- ## 2. Reward Structure | Action | Referrer gets | Referee gets | |---|---|---| | Referee signs up (free trial) | Nothing yet | \$50 credit toward first paid month | | Referee converts to paid | \$100 credit (3.4 months free at 1 seat) | Already received \$50 | | Referee stays paid 3+ months | Extra \$50 credit | — | **Cap:** Max \$1,000 credits earned per referrer per year (prevents arbitrage). **Rationale:** - \$100 credit ≈ 34% of a single seat's annual value — aggressive but sustainable because LTV is ~\$1,800 - Staged payout (\$100 at conversion, \$50 at retention) filters out low-intent referrals - Credits = net-neutral cost because they're deducted from future revenue, not paid in cash --- ## 3. Attribution Rules - Cookie window: 60 days - Primary attribution: referral link with unique code - Secondary: manual code entry at signup - Qualified referral = referee creates workspace + invites 1 teammate + stays paid 14 days - Self-referrals blocked via IP + payment method match + email domain match --- ## 4. Anti-Fraud Safeguards 1. **Domain matching:** referrer and referee cannot share the same company email domain 2. **Payment method uniqueness:** same card cannot be used for 2 referral conversions 3. **Manual review trigger:** any account earning >\$500/month in referral credits is flagged 4. **Chargeback rule:** if referee churns within 30 days, referrer credit is clawed back 5. **Geographic check:** if referrer and referee IPs are within same /24 subnet, flag for review --- ## 5. Sharing UX **Channels offered in-product:** - Copy link (primary — biggest button) - Email (pre-filled template) - LinkedIn share - Slack share - Twitter/X share **Pre-filled email copy:** Subject: "The tool I finally stopped complaining about" Body: 3-sentence personal recommendation + referral link + "you'll get \$50 off your first month" **Referral landing page spec:** - Personalized headline: "[Referrer name] thinks Stacks will work for your agency" - Social proof: testimonial from referrer's industry segment - Clear \$50 credit badge above the fold - Single CTA: "Start 14-day trial" --- ## 6. Viral Coefficient Math - Activated users who share: 22% (industry benchmark for well-designed programs) - Shares per sharing user: 3.1 - Click-through on shared links: 18% - Signup rate on landing page: 24% - Trial-to-paid conversion: 28% **K-factor = 0.22 × 3.1 × 0.18 × 0.24 = 0.0295** (trial signups) **Paid K-factor = 0.0295 × 0.28 = 0.0083** (paid conversions) Modest but stackable with other channels. Expected contribution: 12-18% of new paid signups by month 6. --- ## 7. 30-Day Launch Plan **Week 1:** Build referral dashboard, credit issuance system, cookie tracking, fraud rules **Week 2:** Design sharing UX, landing page, email templates. QA with 10 beta users. **Week 3:** Soft launch to paid users only. Announce via in-app banner + dedicated email. **Week 4:** Monitor fraud signals, measure K-factor, iterate on low-performing share channel. **Success criteria at 30 days:** 8%+ of paid users share at least once, K-factor >0.005 (paid), <2% flagged for fraud review.