Loop mechanics: from working referral to growth engine

Stop asking at signup or in a static settings tab. Wire your in-product share prompt to fire right after the user hits their first real value moment (Dropbox fired it at the storage limit; you fire it after the first completed core action). Add ONE trigger this week, plus a 48-72h backup nudge for people who didn't hit it.

In-product prompts at the aha moment convert at 8-15% vs 1-3% for email/static asks. Sharing is highest exactly when satisfaction peaks — you're harvesting intent instead of interrupting it. Same loop, 3-5x more shares with zero new traffic.

Over three 2-week sprints with equal traffic, test reward shape in order: (1) even double-sided, (2) referrer-weighted, (3) referred-weighted. Then test reward TYPE: cash/credit vs unlocking a premium feature. Pick the winner on viral coefficient AND referred-user retention, not raw signups.

Reward structure swings k-factor hard: single-sided sits at 0.3-0.6, even double-sided at 0.8-1.5. A free in-product unlock often matches a cash payout on conversion at a fraction of the cost, so you find the highest-k, highest-margin combo instead of guessing.

Split-test your share copy: 'Give a friend X' (friend-benefit) vs 'Earn X' (self-benefit), holding the actual reward identical. Ship the altruistic variant if it wins, and surface the friend's reward prominently on the invite/landing page.

Airbnb found friend-first framing beat referrer-first, and a $25/$25 even split drove referred-user conversion to ~52%. The lever is the words, not the money — and once altruistic framing wins it compounds on every future invite for free.

Replace flat 'one reward per invite' with an escalating ladder: a bigger unlock at 3 referrals, a premium tier at 5-10, plus a visible progress bar showing how close they are. Robinhood/queue-jump mechanics work the same way — people chase the next rung.

Tiers tap the achievement/progression instinct and turn your heaviest users into repeat referrers instead of one-and-done. A visible progress bar manufactures a reason to send the NEXT invite, lifting referrals-per-referrer rather than just referrer count.

Drop any 30-day claim delay. Credit the reward the moment the referred user activates (Uber model), and show a live 'Pending / Earned / Claimed' tracker on the referral screen so users see invites landing in real time.

Redemption and conversion friction silently kill working loops. Instant, auto-applied rewards plus a live tracker keep the dopamine attached to the action, so referrers keep sharing — multiplying the same intent you already generate.

Generate a unique, good-looking share image/card per user (their stat, rank, league crest, or result) baked into the invite link so every share is native UGC, not a bare URL. Pair it with one-click prefilled copy for the channels your users actually post on.

Visual UGC is ~85% more cost-effective than influencer content and posts with it earn ~70% more engagement, so each share buys more clicks. The asset turns your referrers into a free content-distribution loop on top of the reward loop.

Set rewards to 'count' only after the referred user ACTIVATES (not signs up), hold rewards pending until genuine engagement, add device fingerprinting / IP + disposable-email checks, cap referrals per period, and add claw-backs for charge-back/abuse patterns.

Referral fraud is a ~$1B/yr problem and 1 in 4 merchants got hit in 2024; the moment a loop works, fraud rings find it. Activation-gated, claw-back-protected rewards keep your unit economics real so a rising k-factor compounds profit instead of leaking it.

signups grew 100K → 4M in 15 months (~3,900%); referrals drove ~2.8M invites/month at peak

Pay the reward in your own product, not cash — Dropbox gave free storage to BOTH sides, so every referral deepened the habit and cost almost nothing. Reward in the thing that makes people stickier.

topped the US App Store and hit ~1M+ daily users within months, purely on invites

Doing the most unscalable thing imaginable — making the invite BE the product — was the whole strategy. Gas only works once your school's friends join, so growth and usage were the same action.

built a ~1M-person pre-launch waitlist; 10K signups day one, 50K week one, from a single referral page

Gamify the WAIT, not just the app. A pre-launch leaderboard ('skip the line by inviting friends') plus a loot-box reward (free stock) turned curiosity into a viral queue with zero product shipped.

scaled ~921K → 21.6M users in a year and 53M+ downloads, largely friend-invite driven

Make the invite intrinsic, not a bolt-on. BeReal is dead solo — your friends MUST be on it — so onboarding pushes contact invites because the product is literally broken without them.