Crypto Exchange Referral Program Development

Crypto Exchange Referral Program Development

A referral program is one of the most effective user acquisition channels for a crypto exchange. The mechanics are simple: user A brings user B, receives a % of B's trading fees. But implementation has several non-trivial aspects: multi-level structures, real-time commission calculation, fraud prevention, and transparent dashboards.

Referral Program Structure

Single-level vs Multi-level

Single-level: referrer gets % only from direct referrals. Simple to implement.

Multi-level: referrer gets % from referrals of referrals (tree structure). Binance uses two-level: 20% from tier-1, 10% from tier-2.

CREATE TABLE referrals (
    referrer_id UUID NOT NULL,
    referee_id UUID NOT NULL UNIQUE,  -- each user has one referrer
    created_at TIMESTAMPTZ DEFAULT NOW(),
    PRIMARY KEY (referrer_id, referee_id)
);

-- For multi-level: recursive CTE to get tree
WITH RECURSIVE referral_tree AS (
    -- Base case: direct referrals
    SELECT referrer_id, referee_id, 1 AS level
    FROM referrals
    WHERE referrer_id = $user_id

    UNION ALL

    -- Recursion: referrals of referrals
    SELECT r.referrer_id, r.referee_id, rt.level + 1
    FROM referrals r
    JOIN referral_tree rt ON r.referrer_id = rt.referee_id
    WHERE rt.level < 3  -- Limit depth
)
SELECT * FROM referral_tree;

Referral Links and Codes

A referral code is a unique string (6-8 characters) associated with a user. A user can have multiple codes for different campaigns and different rates.

import secrets
import string

def generate_referral_code(user_id: str) -> str:
    alphabet = string.ascii_uppercase + string.digits
    code = ''.join(secrets.choice(alphabet) for _ in range(8))
    # Check uniqueness in database
    while self.db.code_exists(code):
        code = ''.join(secrets.choice(alphabet) for _ in range(8))
    return code

# Referral link: https://exchange.com/register?ref=ABCD1234
# UTM parameters for analytics: ?ref=ABCD1234&utm_source=referral&utm_medium=link

Commission Calculation and Accrual

Accrual Formula

On each trade:

def calculate_referral_reward(trade: Trade, referrer: User) -> Decimal:
    # Base referral program rate
    base_rate = Decimal('0.20')  # 20% of collected fee

    # Bonus for referral network volume
    volume_bonus = get_volume_tier_bonus(referrer.referral_network_volume_30d)

    # Rate can be customized for partners
    effective_rate = referrer.custom_referral_rate or (base_rate + volume_bonus)

    # Reward size = trader's fee × referral rate
    taker_fee = trade.fee_amount  # already calculated at execution
    reward = taker_fee * effective_rate

    return reward.quantize(Decimal('0.00000001'))  # 8 decimal places

Asynchronous Accrual

Referral accruals should not block trade execution. Pattern: trade completes, event is published to Kafka, referral service consumes asynchronously:

# referral_worker.py
async def process_trade_event(self, trade_event: dict):
    trade_id = trade_event['trade_id']
    user_id = trade_event['user_id']
    fee_amount = Decimal(trade_event['fee_amount'])
    fee_currency = trade_event['fee_currency']

    # Find referrer of the user
    referrer = await self.db.get_referrer(user_id)
    if not referrer:
        return

    reward = self.calculate_reward(referrer, fee_amount)
    if reward <= 0:
        return

    # Atomically: create record + credit balance
    async with self.db.transaction():
        await self.db.create_referral_earning(
            referrer_id=referrer.id,
            referee_id=user_id,
            trade_id=trade_id,
            amount=reward,
            currency=fee_currency
        )
        await self.db.credit_balance(referrer.id, reward, fee_currency)

Anti-Fraud

Typical Abuse Patterns

Self-referral: user registers a second account using their own link, trades on it, and gets referral % from themselves. Detection: check device fingerprint, IP, and payment methods at registration and KYC.

Wash trading: two accounts trade against each other, generating volume for referral payouts. Detection: analyze counterparties — if % of trades between two accounts is anomalously high, it's suspicious.

Referral farm: mass registration of accounts from one source. Solution: CAPTCHA, SMS verification, require KYC for referral payouts.

def check_self_referral_risk(referrer_id: str, new_user: NewUser) -> RiskScore:
    referrer = db.get_user(referrer_id)
    score = 0

    # Same IP at registration
    if referrer.registration_ip == new_user.registration_ip:
        score += 50

    # Same device fingerprint
    if referrer.device_fingerprint == new_user.device_fingerprint:
        score += 40

    # Same phone or email domain
    if referrer.phone_prefix == new_user.phone_prefix and same_carrier(referrer, new_user):
        score += 20

    return RiskScore(score, 'block' if score >= 70 else 'flag')

Partner Program (Subaffiliates)

Large exchanges highlight "partners" — KOLs, influencers, crypto channels — with elevated rates and separate management:

Partners get custom dashboard with detailed analytics on their referral network, API access to data for embedding in their own products.

Dashboard and Analytics

The referral dashboard should show:

• Number of referrals (total, active last 30 days)
• Accumulated earnings by period
• Table of top referrals by volume (without revealing email/name)
• Real-time earnings counter for today
• Payment history with CSV export option

interface ReferralDashboard {
  stats: {
    totalReferrals: number;
    activeReferrals30d: number;
    totalEarned: Record<string, string>; // currency → amount
    earnedToday: Record<string, string>;
    pendingPayout: Record<string, string>;
  };
  topReferees: {
    rank: number;
    userId: string; // masked: "user***1234"
    volume30d: string;
    yourEarnings30d: string;
  }[];
  earningsHistory: {
    date: string;
    amount: string;
    currency: string;
    tradeId: string;
  }[];
}

Transparency of accruals is key to trust. Users should see each accrual tied to a specific referee's trade.

Payouts and Tax Implications

Referral payouts are credited in real-time to trading balance. Withdrawal is a standard exchange withdrawal process. Some exchanges add a minimum threshold for referral income withdrawal.

In some jurisdictions (US, EU), referral payouts qualify as taxable income. The exchange must maintain payment history and, when thresholds are exceeded (e.g., $600/year in the US), provide tax forms. This requires collecting TIN/tax identification numbers from high-earning partners.