After subsidies end, verifying user retention is not about total addresses or TVL; it's about how many users actively return to use the protocol's core functions in months 3–6 after incentives stop. Verification requires three steps: choosing the right window, breaking down retention types, and cross-checking data authenticity.
Step 1: Set the Verification Window – Months 3 and 6 After Incentives End
Data from the first month after rewards end has limited value because users need time to react and there may be delayed withdrawals.
What to do: Identify two key observation points – the end of month 3 and the end of month 6.
How to do it: After the incentive program ends (set as T+0), record data at T+90 and T+180. Compound's analysis shows that users inactive for over 6 months are considered "churned," while those returning within 6 months are defined as "retained."
Completion criteria: Ability to compare changes in TVL and active addresses across T+0, T+90, and T+180.
Step 2: Distinguish Real Retention from "Zombie" Retention
Not all users who haven't withdrawn funds are genuinely retained. You need to look at two types of behavior.
Case A: Capital Retention vs. Behavioral Retention. Capital retention means users didn't remove LP positions or deposits—maybe due to laziness or high gas fees, not trust. Behavioral retention means users are still actively trading, borrowing, staking, etc. Only the latter counts as true retention.
Case B: High-Value Retention vs. Low-Quality Retention. Data shows retained users typically have frequent, regular usage habits and higher average deposit amounts than churned users. Addresses with tiny deposits and no activity, even if they haven't withdrawn, add limited real value.
Completion criteria: Ability to calculate "average deposit amount of retained users vs. all users." If the former is significantly higher, the capital quality of retained users is stronger.
Step 3: Cross-Check On-Chain Data with Protocol Revenue
On-chain data showing users are still there doesn't mean the protocol is creating value. You need to look at user numbers and revenue together.
What to do: Compare protocol revenue before and after incentives end.
How to do it: The Revenue-to-Incentives Ratio is a key sustainability metric. If revenue doesn't fall proportionally after incentives stop, the subsidies truly converted into real demand. Also, compare TVL quality with trading volume: if TVL stays stable but volume plummets, users left money but stopped using protocol functions—this is "storage," not retention.
Completion criteria: Can you answer, "After incentives ended, did protocol revenue establish a new baseline higher than before?"
How to Verify Completion
If three months after incentives end, the number of monthly active addresses for the protocol's core functions (trading, lending, staking) stays above 30% of the peak incentive period, and the average deposit of retained users is clearly higher than that of churned users, it indicates subsidies served as a "filter" rather than "renting" users.
Next Steps
If verification shows poor retention, watch whether the protocol implemented a tapered end instead of a cliff—gradually decreasing incentives usually keep more users. For yourself, if you're still providing liquidity, use this framework to check the pool you're in: if its trading volume continuously drops to zero after incentives end, it's smarter to exit early rather than wait for higher slippage losses.
