Referral programs are one of the most cost-effective growth levers in DeFi. Referred users cost less to acquire, activate at higher rates, and generate more fee revenue than users from paid channels.

Most DeFi teams treat referral rewards as a single decision ("how much do we pay per referral?") when it is five separate decisions: reward structure, reward type, vesting schedule, eligibility criteria, and measurement framework. Getting anyone wrong undermines the rest.

This guide covers:

• What a DeFi referral program is and the key elements that separate durable programs from exploitable ones

• How to protect your program from Sybil attacks, and why resistance needs to be built in from day one

• The 5 main referral reward structures used across DeFi in 2026, with real protocol examples and tradeoffs

• How to choose between tokens, stablecoins, and fee rebates, and why vesting mechanics matter

• Real-world examples from GMX, Hyperliquid, and Blast

• The onchain metrics and 5-metric dashboard that tell you whether growth is real or manufactured

What Is a DeFi Referral Program?

A referral program rewards existing users for bringing new wallets to an app. The referrer shares a unique link or code, the new user completes a qualifying action such as a swap, a deposit, or a liquidity provision, and both sides receive a reward. According to Harvard Business Review, referred customers have a 16%–25% higher lifetime value than non-referred customers, cost less to acquire, and generate more fee revenue than users from paid channels.

What makes DeFi referral programs structurally different from traditional ones is wallet-based identity. A wallet address costs nothing to create, which means any program that pays per referred wallet without verification is paying per wallet creation. That single fact separates a well-designed DeFi referral program from an exploitable one. 

Key Elements of a DeFi Referral Program

Three elements separate programs that generate lasting protocol growth from those that drain the rewards pool and produce no retention.

Mutual incentive on both sides. Programs that reward both the referrer and the referee consistently outperform single-sided structures. Dual-sided programmes generate 53% more referrals than single-sided equivalents and increase participation by 29%. When only the referrer benefits, referees have no financial reason to complete the qualifying action.

Eligibility criteria that filter for genuine users. Without wallet age requirements, attestation checks, and behaviour-based triggers, any open referral program is a Sybil target. Nansen's analysis of Linea's 2025 airdrop flagged nearly 40% of eligible addresses as Sybil. Referral programs share the same vulnerability. 

See what Sybil attacks are and how to prevent them for the full breakdown.

Measurement that goes beyond the share rate. Share rate and referral count measure funnel volume. The metrics that measure program health are qualified referral rate, 30-day retention of referred wallets, and revenue per referred wallet. A program optimised for share rate alone produces wallet counts that overstate growth and retention curves that collapse when rewards end.

Step-by-Step Guide to Launching a DeFi Referral Program

Step 1: Define the qualifying action

Decide what the referred wallet must do to trigger a reward. For a DEX, that is a completed swap above a minimum volume threshold. For a lending app, it is a first deposit. For a yield vault, it is a first liquidity provision. The qualifying action should require genuine product engagement, not just a wallet connect. 

See DeFi activation metrics for activation definitions by app type.

Step 2: Choose a reward structure

Match the structure to your protocol's stage. Early-stage apps benefit from two-sided activity-gated structures where farming risk is low, and the activation signal is high. Growth-stage apps with fee revenue can layer in revenue share for top referrers. The five structures covered in this guide map to different growth objectives and exploitation profiles.

Step 3: Set eligibility criteria

Require a minimum wallet age of 90 days, prior transaction history across at least one other chain, and at least one verifiable attestation before a referred wallet qualifies for reward distribution. These three filters remove the majority of farming wallets before any rewards are distributed.

Step 4: Configure vesting

Lock a portion of rewards for 30 to 90 days. Wallets that remain active long enough to claim the locked portion are a reliable signal of genuine engagement. Immediate full distribution maximises farming risk in equal measure to the participation rate.

Step 5: Set up measurement before launch

Configure UTM tracking on all referral links using a consistent naming convention. Verify that referral parameters are captured at the session level and joined to wallet activity before the program goes live. A program launched without attribution tracking cannot tell you which referrers drive retained wallets versus one-time visitors. 

See the onchain attribution guide for the full setup.

Real-World Examples of Successful DeFi Referral Programs

GMX

GMX's two-sided tiered structure is the most cited referral program in DeFi for a reason. Both referrer and referee receive a 5% trading fee discount at Tier 1. Tier 2 requires 15 active weekly users generating $5 million in weekly volume, unlocking 10% discounts for both sides. Tier 3 requires 30 active weekly users and $25 million in weekly volume, with affiliates earning 15% of fees in the market's collateral token. Within 9 months, the program contributed to $11 billion in trading volume and $11 million in protocol revenue, with $1.6 million paid out in referral rewards. The program works because tier advancement requires verified onchain activity, not wallet registrations. By September 2024, it had driven over $65 billion in total referred trading volume.

Hyperliquid

Hyperliquid runs a revenue share model where referrers earn 10% of referred users' fees on their first $1 billion in volume, with referred traders receiving a 4% fee discount on their first $25 million in volume. The structure selects for referrers who recruit genuinely active traders, because the referrer's income scales directly with the referred wallet's ongoing trading activity. Hyperliquid generated $844 million in protocol revenue in 2025, of which 96% came from perpetual trading fees, demonstrating what sustained fee-generating activity looks like at scale.

Blast

Blast's early access program gated expansion behind genuine capital commitment. Each account received 5 invite codes. Additional codes unlocked only once referred users deposited 5 ETH into the bridge. Referrers earned 16% of points from direct invitees and 8% from second-degree invitees. The activity-gated structure filtered out wallets with no intent to deposit while creating a measurable signal of genuine commitment at each stage of the funnel.

DeFi Referral Program Best Practices

Track qualified referral rate before scaling spend. A qualified referral rate below 50% means more than half the funnel is bots, Sybil clusters, or wallets that will never activate. Fix the eligibility criteria before increasing the rewards budget. 

See how to reduce CAC in DeFi for the full framework.

Separate referrer cohorts by acquisition channel. Referrers from organic community channels are retained differently from referrers recruited through paid campaigns. Use onchain attribution to track which referrer sources produce wallets with the highest 30-day retention and revenue per wallet, then concentrate incentives on those channels.

Monitor referred wallet retention after reward unlock. A cohort of referred wallets that stops transacting immediately after the locked rewards vest is a program functioning as an airdrop. Track this metric for every reward unlock event. If post-unlock churn is high, the reward structure is attracting farmers rather than users. 

See DeFi churn analytics for the measurement framework.

Use wallet intelligence to identify your best referrers. Power referrers are not always your most visible community members. Filter your referrer base by the retention and revenue metrics of the wallets they have brought in, not by the volume of referrals generated. Formo's wallet profiles surface net worth, DeFi positions, and lifecycle stage for every referred wallet, so referrer quality is visible at the individual level.

Review program performance quarterly, not annually. DeFi market conditions shift fast. Yield compression, competitor launches, and incentive program endings all affect referral program performance within weeks. DeFi growth experiments cover how to structure iterative program reviews around onchain outcome data rather than lagging dashboard metrics.

Why Sybil Resistance Is the Biggest Threat to Referral Programs

In DeFi, a user is a wallet address. Creating a new one costs nothing and takes under a minute. That single fact is the structural vulnerability every referral program inherits: one actor can operate hundreds of addresses simultaneously, routing rewards intended for distinct users back to themselves.

The scale of confirmed Sybil activity in onchain incentive programs is not marginal. Nansen's analysis of Linea's 2025 airdrop flagged 516,960 of 1,297,203 eligible addresses as Sybil, nearly 40% of all participants. Referral programs share the same eligibility mechanics as airdrops and inherit the same exposure.

The damage runs deeper than the reward budget:

• Financial drain: At 40% Sybil concentration, the majority of reward spend transfers to farmers rather than the community, the program was built to grow

• False growth signals: Sybil wallets produce metrics that look identical to real growth. Wallet counts rise, referral volume increases, and TVL figures move. The divergence only becomes visible at the retention layer, after budget decisions have already been made on inflated numbers

• Fairness erosion: Legitimate referrers notice when a small number of actors extract a disproportionate share of the rewards pool. Genuine program participation shrinks as a result

• Reputation damage: A program that is publicly gamed signals to the broader community that the app can be exploited. Subsequent programs attract more farmers and fewer genuine users

The goal of Sybil resistance is to raise the cost of exploitation above the value of the reward, making farming economically irrational. Combining 2 or 3 of the following layers achieves adequate resistance for most apps.

How to Protect Your Referral Program

Layer 1: Behaviour-Based Reward Triggers

Tie rewards to actions that require sustained engagement: 30-day liquidity provision, a minimum swap volume threshold over 14 days. A 2025 paper on Sybil address detection published on arXiv found that temporal features achieve greater than 0.9 precision, recall, F1, and AUC in Sybil detection across a dataset of 193,701 addresses. Time-locked activity patterns are the most powerful fraud signal available.

GMX's tier system applies this logic directly: new referral codes start at Tier 1 (5% rebates) and only advance to Tier 2 after the referrer demonstrates 15 active weekly users generating $5 million in weekly volume.

Layer 2: Wallet Quality Scoring

Require referred wallets to have a minimum age (90 days is a reasonable baseline) and prior transaction history across at least one other chain before becoming eligible for reward distribution. The same arXiv paper found that 97.4% of confirmed Sybil wallets had wallet lifecycles under one year. This filter alone removes the majority of farming wallets.

Score wallets on verifiable signals and assign reward tiers accordingly. Formo's wallet labels automate this scoring by assigning each referred wallet a set of labels derived from three sources:

• Onchain activity - protocols used, tokens held, transaction patterns across chains

• Attestations - verified credentials including Coinbase Verified Account, Binance Account Bound Token (confirming KYC completion), and Human Passport unique humanity scores (0-100)

• App behaviour - session count and engagement patterns on your app, surfaced as lifecycle labels (New User, Returning User, Power User)

A referred wallet holding a Coinbase Verified Account attestation and a Human Passport score above 50 is verifiably human and KYC-linked. A wallet with zero attestations, one session, and no cross-chain history scores low regardless of wallet age.

Wallets matching multiple positive labels qualify for full reward distribution. Single-label or zero-label wallets enter a holding period pending further activity.

Layer 3: Progressive Reward Unlocks

Distribute rewards in stages rather than as a lump sum. Each unlock is triggered by a new onchain milestone. A wallet that completes the first milestone and disappears before the second is a farming signal; a wallet that completes all milestones over 60 days is a strong retention signal.

Token-gated forms are a practical entry-point filter. Requiring referred wallets to pass a verification step (wallet balance check, social proof, onchain attestation) before entering the referral funnel catches bots at the top of the funnel. Formo's token-gated forms support wallet balance verification, X/Discord/Farcaster social verification, and onchain attestations across EVM and Solana.

DeFi Referral Reward Structures Explained

The right structure depends on your protocol's stage, token economics, and what you are optimising for: wallet count, TVL, trading volume, or long-term retention. Each model below solves a different problem and carries a different exploitation profile.

1. Flat One-Sided Reward

The referrer earns a fixed reward when a referred wallet completes a qualifying action. The referee receives nothing.

• Best for: Protocols with an established community whose members will refer to product quality alone

• Real example: Lido's referral program paid 0.75% in LDO tokens per ETH staked through a referral link. The rate was reduced to 0.50% in May 2022, and the DAO voted to retire both programs on 15 April 2023, citing reward pool depletion and sustainability concerns.

• Watch for: With no financial incentive to complete the qualifying action, referees in one-sided programs consistently underperform on activation. Impact.com's State of Referral Marketing Report (2024) found that rewarding the referee produces higher conversion rates, yet 96% of single-sided programs reward only the referrer

2. Two-Sided Reward

Both the referrer and the referee receive a reward when the referee completes a qualifying action. This is the dominant model in DeFi growth programs.

• Best for: Protocols in active growth mode where acquisition cost is justified by TVL or fee revenue from activated wallets

• Real example: GMX's two-sided structure gives both referrer and referee a 5% trading fee discount at Tier 1. The referee gets an immediate financial reason to trade; the referrer earns ongoing rebates. Within 9 months, the program contributed to $11 billion in trading volume and $11 million in protocol revenue, with $1.6 million paid out in referral rewards

• The data: Dual-sided programs generate 53% more referrals than single-sided equivalents and increase program participation by 29% 

• Watch for: Higher reward cost per acquisition. Eligibility criteria need to block self-referral across wallet addresses from day one

3. Tiered Referral Program

Rewards scale with the referrer's cumulative performance. Tiers can be based on referral count, referred TVL, or referred trading volume.

• Best for: Protocols targeting KOLs and power users who can drive volume at scale

• Real example: GMX V2 runs 3 tiers. Tier 1 requires no minimum and gives a 5% fee discount to both sides. Tier 2 requires 15 active weekly users and $5 million in weekly volume, unlocking 10% discounts for both sides. Tier 3 requires 30 active weekly users and $25 million in weekly volume; affiliates earn 15% of fees in the market's collateral token. dYdX's affiliate program starts all affiliates at 30% of taker fees, scales to 40% for referrers generating over $1 million in 30-day volume, and reaches 50% for those generating over $10 million, with a single $10,000 cap per referred wallet per 30-day period

• The data: Tiered incentive structures generate 27% more referrals than flat-reward programs 

• Watch for: As referrers approach tier thresholds, the incentive shifts toward recruiting any wallet to hit the number. GMX builds this in by requiring active weekly users toward tier advancement, with wallet registrations alone insufficient to advance tiers. Require onchain verification of referred wallet activity before counting a referral toward tier advancement

4. Revenue Share (Recurring)

The referrer earns a percentage of fees or yield generated by referred wallets on an ongoing basis. This is the structure with the strongest long-term alignment between referrer and protocol.

• Best for: DEXs, lending protocols, and perpetuals platforms where fee revenue per active wallet is predictable and meaningful

• Real example: dYdX runs a single volume-scaled affiliate tier: all affiliates start at 30% of taker fees, rising to 40% once they refer over $1 million in 30-day trading volume, and 50% once they refer over $10 million. Payouts are capped at $10,000 per referred wallet per 30-day period and are paid in USDC. Hyperliquid pays referrers 10% of referred users' fees on their first $1 billion in volume, with referred traders receiving a 4% fee discount on their first $25 million in volume

• Watch for: Revenue share selects for referrers who recruit high-volume traders. It also rewards wash trading when volume goes unverified. Track the referred volume against open interest and liquidation rates to identify artificial inflation

5. Activity-Gated Milestone Reward

Rewards unlock progressively as the referred wallet reaches onchain milestones: first swap, first liquidity provision, and a 30-day active streak. No single action triggers full reward release.

• Best for: Protocols focused on long-term retention where farming is a real concern

• Real example: Blast's early access program gave each account 5 invite codes. Additional codes unlocked only once referred users deposited 5 ETH into the bridge, gating program expansion behind genuine capital commitment. Referrers also earned 16% of points from direct invitees and 8% from second-degree invitees.

• Watch for: Complexity reduces conversion. Give users real-time visibility of their milestone progress to maintain mid-funnel momentum.

How to Choose the Right Referral Program Structure for Your DeFi App

The right referral structure depends on where your DeFi app or protocol sits in its growth cycle and what you are optimising for.

• Pre-launch or early traction: Start with a two-sided activity-gated program. Farming risk is low, activation signal is high, and the milestone completion data tells you exactly where your onboarding funnel breaks before you scale spend. Blast's approach applied this logic: referred users had to deposit 5 ETH to unlock additional invite codes, gating program expansion behind genuine capital commitment

• Growth stage with fee revenue: Run a two-sided program as the base and layer in a revenue share component for your top referrers. This selects for referrers who actively recruit high-volume wallets. GMX's tiered structure is the clearest example: the base tier is accessible to anyone, but Tier 3 (30 weekly users, $25 million in weekly volume) selects for professional affiliates who generate real trading activity

• Established protocol with an active community: Tiered structures work well here. Your existing users have enough protocol context to recruit quality wallets, and the tier mechanics give them a reason to keep referring after the first conversion. dYdX's sliding affiliate program, where commission scales from 30% to 50% based on 30-day referred volume, is designed precisely for this stage

Structure, reward type, vesting schedule, eligibility criteria, and measurement all need to point at the same outcome: wallets that stay, make transactions, and generate fee revenue. The programs that compound over time reward genuine protocol usage. GMX reached $11 million in revenue from $1.6 million in referral spend within 9 months; by September 2024, the program had driven over $65 billion in total referred trading volume.

For teams building or auditing their DeFi growth stack, the DeFi incentives and growth playbook covers the broader context of when token rewards accelerate growth and when they undermine it.

The DeFi liquidity bootstrapping guide covers how to attract capital without creating mercenary TVL.

For retention strategy after your referral program is live, the churn analytics guide covers how to detect and act on pre-churn signals before wallets go dormant.

Key Referral Reward Types: Tokens, Stablecoins, or Fee Rebates

The structure determines the mechanics. The reward type determines the quality of users you attract. These are two separate decisions.

Native Tokens

The default choice because they are cheap to distribute and, in theory, align referrers with the protocol's long-term success. In practice, they attract farmers who sell immediately after claiming, creating selling pressure at the moment you want price stability. Token rewards work best when paired with a vesting schedule that makes immediate selling unprofitable.

Stablecoins

Costs more per referral but attracts a different user profile: one who values the reward for its cash value rather than speculative upside. dYdX pays affiliates in USDC, selecting for referrers motivated by consistent, predictable income rather than token speculation.

Fee Rebates

Carry the strongest alignment for fee-generating apps. A referrer earning a share of the fees their referred traders generate has a direct incentive to bring in genuinely active users. GMX's collateral-token rebates and Hyperliquid's fee-share model follow this logic: the referrer's income scales with the referred trader's ongoing activity.

Vesting and Lock-Up Periods

Immediate reward distribution maximises participation rates and farming risk in equal measure. Locking a portion of rewards for 30 to 90 days filters out wallets that exist only for extraction. Wallets that remain active long enough to claim the locked portion are a reliable signal of genuine engagement.

The Arbitrum case: Independent onchain analysis by X-explore found that 279,328 same-person addresses, forming 60,000+ coordinated clusters, controlled an estimated 47.96% of the total ARB airdrop supply. Eligibility criteria based on shallow onchain activity, with no time-locked behaviour requirements, was the direct cause. A referral program that distributes tokens immediately to unverified wallets runs the same risk.

How to Measure Referral Program Performance

Share rate and referral count tell you how many wallets entered the funnel. They say nothing about whether those wallets generated lasting value. The metrics that matter fall into 3 layers.

Layer 1: User Acquisition Quality

Referral CAC is only meaningful when compared against your other acquisition channels. Onchain attribution makes this comparison possible: Formo captures both the offchain referral link (UTM and ref parameters) and the onchain qualifying action (swap, deposit) in a single funnel, so you can see cost-per-activated-wallet by channel side by side.

Layer 2: User Activation

These answer whether referred wallets are doing anything meaningful after they arrive.

• Activity depth per referred wallet (30-day): Wallets that complete only 1 swap or deposit are a farming signal. Active users place multiple orders or add liquidity within the first 30 days

• Referred TVL: For lending and liquidity protocols, the total value locked by referred wallets is a more direct growth signal than wallet count. A program that adds 500 wallets but $0 in TVL has produced no capital growth for the protocol

• Milestone completion rate: In activity-gated programs, the percentage of referred wallets reaching each milestone reveals exactly where the funnel breaks. If 70% complete milestone one but only 20% reach milestone two, the second milestone is either too complex or too distant in time

• K-factor (viral coefficient): The number of new users each referred user subsequently refers. A K-factor above 1.0 indicates self-sustaining growth. Most DeFi programs operate between 0.3 and 0.7; anything above 0.8 is strong performance

Layer 3: Long-Term Retention

These answer whether the growth holds when rewards stop.

• 30/60/90-day wallet retention: Referred wallets should retain at rates comparable to organic users. A DeFi program where referred wallets retain at significantly lower rates than organic users signals that the structure or eligibility criteria need revisiting

• Revenue per referred wallet (RPW): Total fee revenue generated by referred wallets divided by the number of referred wallets. This is the clearest single signal of program ROI. GMX's $11 million in revenue from $1.6 million in referral rewards translates to an RPW that justified scaling the program; by September 2024, it had driven over $65 billion in total referred trading volume

• Referred wallet churn after reward unlock: A cohort of referred wallets that stops completing transactions immediately after the locked rewards vest is a program functioning as an airdrop. Track this metric for every reward unlock event

Key Metrics For a Referral Program Dashboard

A referral program can show strong share rates and referral counts while generating no lasting value. These 5 metrics answer the question that matters: are the wallets you acquired worth keeping?

1. Qualified Referral Rate

Formula: Referred wallets passing eligibility criteria / Total referred wallets

The first filter on program health. A qualified referral rate below 50% means more than half your funnel is bots, Sybil clusters, or wallets that will never activate. Fix eligibility criteria before scaling spend.

2. Referral Conversion Rate

Formula: Referred wallets activated / Total referred wallets

Measures whether the referee reward is compelling enough to drive action. A low conversion rate with a high share rate means referrers are sharing, but referees are not following through; the referee reward needs revisiting.

3. Referral CAC vs. Other Channels

Formula: Total program cost / New activated wallets

Referral CAC is only meaningful in comparison. Onchain attribution makes this comparison possible by capturing both the offchain referral link and the onchain qualifying action in a single funnel, so cost-per-activated-wallet is visible by channel side by side.

4. 30-Day Retention of Referred Wallets

Formula: Referred wallets active at day 30 / Total referred wallets activated

Referred wallets should retain at rates comparable to organic users. A large gap between referred and organic retention is a direct signal that the program is attracting farmers rather than genuine users. Formo's retention chart shows cohort-level return rates by week, filterable by referral source.

5. Revenue Per Referred Wallet (RPW)

Formula: Total fee revenue from referred wallets / Number of referred wallets

The clearest single signal of program ROI. GMX's $11 million in revenue from $1.6 million in referral rewards is the benchmark: every dollar in referral spend returned roughly $6.9 in protocol revenue. Track RPW by referrer cohort to identify which referrers are driving genuine fee-generating activity.

For a broader view of the metrics that drive DeFi growth beyond referrals, the DeFi KPIs and metrics guide covers the full measurement stack.

How to Analyze Your DeFi Referral Program with Formo

Formo connects the 2 halves of the referral funnel that most analytics tools treat separately: the offchain referral link and the onchain qualifying action. It helps you answer the following key questions:

• How is my referral program doing?

• What’s the ROI of my referral program?

• Which referrals are driving the most volume?

• Which channels and campaigns are driving the most revenue?

• Are referrals retaining at a higher rate than normal users?

When a wallet lands via a referral link, Formo captures the ref parameter and UTM data automatically at the session level. When that wallet connects and completes a transaction, Formo links the onchain action back to the originating referral source. The result is a complete view of the referred wallet journey, from first click to 90-day retention, in a single dashboard.

• Onchain attribution: Captures referral code parameters and campaign data automatically, then ties them to wallet connects and transactions. The Overview page shows wallet connects, transactions, and volume broken down by referral source, side by side with other acquisition channels

• Funnel analysis: Build a referral funnel (landing page visit → wallet connect → first transaction → 30-day activity) and filter by referral source to see where referred wallets drop off, with both sequential and any-order funnel modes

• Wallet intelligence: Every referred wallet that connects gets a wallet profile with net worth, DeFi positions, token holdings, cross-chain transaction history, and lifecycle stage (New, Returning, Power User, At Risk, Churned). Filter the Users page by referral source to compare wallet quality across referrers

• Retention tracking: Formo's retention chart shows cohort-level return rates by week. Create a segment of referred wallets and compare their retention curve against organic users to detect farming behaviour before reward unlock

• Segments and alerts: Build a segment of referred wallets that have hit each milestone and set an alert for high-value referred wallets (net worth above $50,000 or Power User lifecycle stage)

• Ask Formo: Query referral data in natural language. "Show me wallet connects and transactions grouped by ref parameter for the last 30 days" returns a ranked table of referrer performance. The Explorer supports custom SQL for deeper analysis, including RPW and churn rate after reward unlock

Get started for free or learn more about onchain attribution.

Final Takeaway

A DeFi referral program compounds when every decision points at the same outcome: wallets that stay, complete transactions, and generate fee revenue. Reward structure, eligibility criteria, vesting schedule, and measurement framework are not independent choices. A well-structured program with no Sybil resistance drains the rewards pool. A Sybil-resistant program with no measurement framework cannot tell you which referrers are worth paying more for. The programs that generate durable growth, such as GMX, Hyperliquid, and Blast each demonstrate in different ways, are the ones that tie every mechanic back to verified onchain behaviour rather than wallet counts.

Frequently Asked Questions

What is a DeFi referral program?

A DeFi referral program rewards existing users for bringing new wallets to an app. The referrer shares a unique link or code, the new user completes a qualifying action such as a swap, a deposit, or a liquidity provision, and both sides receive a reward. What separates DeFi referral programs from traditional ones is wallet-based identity: a wallet address costs nothing to create, which means any program that pays per referred wallet without verification is paying per wallet creation rather than per genuine user.

What is the best referral reward structure for a DeFi app?

Two-sided structures work best for growth-stage apps where activation cost is justified by TVL or fee revenue. Activity-gated milestone structures suit retention-focused apps. Tiered and revenue-share models fit fee-generating DEXs and perpetual platforms with predictable volume per active wallet. Match the structure to your stage and what you are optimising for.

How do you measure whether a referral program is working?

Track 5 metrics: qualified referral rate, referral conversion rate, referral CAC against other acquisition channels, 30-day retention of referred wallets, and revenue per referred wallet. Share rate and referral count show funnel volume. The 5 metrics above show whether that volume is generating lasting value for the app.

How do you stop Sybil farming in referral programs?

Combine at least 2 layers: behaviour-based reward triggers tied to sustained activity, wallet age and cross-chain history requirements, onchain attestation scoring (Coinbase Verified, Human Passport), smart contract-level self-referral prevention, and progressive reward unlocks. Time-locked activity patterns achieve above 0.9 precision in Sybil detection according to the 2025 arXiv paper.

Should referral rewards be paid in tokens, stablecoins, or fee rebates?

Each attracts a different referrer profile. Native tokens are cheapest but attract yield farmers who sell on receipt. Stablecoins attract income-motivated referrers who value predictable payouts. Fee rebates suit fee-generating apps because the referrer's income scales with the referred user's ongoing trading activity, not a one-time event.

Why do tiered referral programs attract low-quality wallets?

Referrers approaching a tier threshold are incentivised to recruit any wallet to hit the number. The fix is counting only verified onchain activity toward tier advancement. GMX's tier system applies this directly: tier progression requires active weekly users generating minimum weekly volume thresholds, with wallet registrations alone insufficient to advance tiers.

What onchain metrics should I track for a referral program?

Start with referral CAC, referral conversion rate, and qualified referral rate to assess acquisition quality. Then track 30/60/90-day retention of referred wallets and revenue per referred wallet to assess value. For activity-gated programs, milestone completion rate shows exactly where the funnel breaks between each reward unlock.

Which DeFi referral programs have generated the most verified results?

GMX's tiered two-sided structure drove over $65 billion in total referred trading volume by September 2024, generating $11 million in protocol revenue from $1.6 million in referral spend. Hyperliquid's revenue share model generated $844 million in protocol revenue in 2025, with 96% coming from perpetual trading fees. Blast's activity-gated program gated expansion behind a 5 ETH deposit requirement, filtering out low-intent wallets before they entered the rewards pool.