A share-to-earn referral program turns every CCFish player into a free user acquisition channel by rewarding them with in-game currency for inviting friends who install and play.
The Problem: Mobile Game UA Costs Are Skyrocketing
In Q1 2026, the average cost per install (CPI) for a casual mobile game in tier-1 markets hovers between $3.00 and $5.50. For social-casual-fishing titles like CCFish, CPIs north of $7.00 are common. Ad fatigue, Apple's ATT, and Google's Privacy Sandbox have crippled targeted UA campaigns. Every indie studio bids on the same ad placements, driving eCPMs up while ROAS narrows. Paid campaigns routinely deliver low-LTV users who churn within 72 hours.
For an indie title like CCFish, competing head-to-head in the programmatic UA auction against studios with million-dollar budgets is a losing strategy. A different distribution model is required.
The Solution: Share-to-Earn Referral Mechanics
CCFish's share-to-earn referral program aligns player incentives with growth. The core mechanics:
| Referral Action | Reward (Pearls) |
|---|---|
| Share invitation link to social platform | 50 (capped at 200/day) |
| Referred friend installs CCFish | 200 (one-time per device) |
| Referred friend reaches level 5 | 500 (one-time) |
| Referred friend makes first IAP | 1,000 (one-time) |
Pearls are CCFish's premium in-game currency — players use them for rare rods, bait packs, cooldown skips, and cosmetics. A player who successfully refers three friends who reach level 5 earns 2,100 Pearls, roughly equivalent to a $4.99 IAP bundle. The player gets premium currency free, and CCFish acquires three pre-qualified users at effectively zero marginal cost.
CCFish's core demographic — players aged 25-45 who enjoy casual, social, collection-based gameplay — already participates in fishing communities on WhatsApp, Telegram, Discord, and LINE. Share-to-earn taps into existing social graph momentum rather than generating cold traffic.
Architecture Overview
CCFish's referral system runs on a serverless Cloudflare stack built for viral spikes:
- **Workers (Edge Compute):** Handle referral link generation (`POST /api/v1/referral/create`), install attribution (`GET /api/v1/referral/attribution`), and reward distribution (`POST /api/v1/referral/reward`). Each endpoint validates HMAC signatures, checks fraud signals, and credits Pearl balances.
- **D1 (Serverless SQLite):** Stores the referral graph, reward ledger, and daily cap counters. The schema tracks referrer_id, referred_device_id, referral_code, status, and timestamps.
- **KV (Key-Value Store):** Provides millisecond lookups for install dedup (TTL: 90 days), IP dedup (TTL: 24 hours), and rate-limiting (max 50 link generations per hour per player).
Referral Graph Schema
```sql
CREATE TABLE referrals (
id INTEGER PRIMARY KEY AUTOINCREMENT,
referrer_id TEXT NOT NULL,
referred_device_id TEXT NOT NULL UNIQUE,
referral_code TEXT NOT NULL UNIQUE,
status TEXT DEFAULT 'pending'
CHECK(status IN ('pending','installed','level_5','first_iap')),
created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
);
```
Implementation: Building the Referral System
**Deep Linking & Attribution:** CCFish uses Universal Links (iOS) and App Links (Android) with a 72-hour attribution window. On first launch, a device fingerprint (hashed IDFV/Advertising ID + device model + OS version) is checked against KV for dedup before creating the referral record.
**Milestone Tracking:** CCFish emits game events through an internal event bus. The referral system subscribes to `player.first_session` (install), `player.level_up.5` (level 5), and `iap.first_purchase` (first IAP). Each event triggers an idempotency check, daily cap validation, and reward distribution.
**Fraud Prevention:**
| Layer | Check | Action |
|---|---|---|
| Device | Same fingerprint as referrer and referee | Block, flag both |
| IP | >5 referrals from same IP in 24h | Block further claims |
| Temporal | Claim <5 min after link generation | Require manual review |
| Graph | Referral chain depth >3 | Suspend rewards |
| Behavioral | Referred account churns in 2 sessions | Revoke reward |
**Share Channels:** The system generates platform-specific content for WhatsApp, Telegram, and Discord, with clipboard fallback. Each share action is tracked for analytics.
Results: K-Factor Analysis
K-factor (viral coefficient) = invites sent per user (i) × conversion rate (c). K > 1.0 means viral growth.
| Metric | Conservative | Moderate | Optimistic |
|---|---|---|---|
| Invites/player (i) | 2.0 | 4.5 | 8.0 |
| Conversion (c) | 8% | 12% | 18% |
| **K-factor** | **0.16** | **0.54** | **1.44** |
In the optimistic scenario, the top 20% of players become active referrers and push K above 1.0. Even in the moderate scenario, K=0.54 means 540 additional users from every 1,000 organic installs, reducing blended CPI by ~35%.
**Cost impact:**
| Source | Users | Cost/User | Total Cost |
|---|---|---|---|
| Paid UA (1,000 users) | 1,000 | $4.50 | $4,500 |
| + Referral (moderate) | 1,540 | $2.92 blended | $4,500 |
| + Referral (optimistic) | 2,440 | $1.84 blended | $4,500 |
Blended CPI drops from $4.50 to $1.84, freeing budget for game content and further UA.
**Caveat:** Referral programs amplify existing growth but rarely create it from nothing. CCFish should maintain a modest paid UA baseline ($500/month) to keep the top-of-funnel fed.
Key Takeaways
1. **Share-to-earn aligns player incentives with growth.** Rewarding referrals with Pearls turns engaged players into a motivated acquisition force. The reward is intrinsically valuable to players and costs CCFish only marginal infrastructure.
2. **Serverless (Workers + D1 + KV) fits referral systems perfectly.** The workload is event-driven, bursty, and latency-sensitive — exactly Cloudflare's edge sweet spot.
3. **Fraud prevention must be layered.** Combining device, IP, temporal, graph, and behavioral checks keeps abuse below 2%.
4. **Deep linking with a 72-hour attribution window is the mobile referral standard.** Universal Links and App Links provide smooth UX with proper app store fallback.
5. **K-factor > 1.0 is attainable but requires frictionless design.** Even K=0.54 delivers a 35% blended CPI reduction.
6. **Share-to-earn ≠ cross-promotion.** Cross-promotion moves users between sibling titles. Share-to-earn recruits new users via existing players' social graphs. For a standalone game like CCFish, it's the correct viral strategy.
CCFish's share-to-earn program decouples growth from paid UA. In an era where ATT, Privacy Sandbox, and CPI inflation make traditional UA untenable for indie games, turning players into your acquisition channel is survival.