In-Depth Behavioral Profiles

To move beyond superficial usage statistics and uncover deeper patterns of engagement on Aerodrome, we designed three original behavioral metrics tailored to evaluate user activity across time, intensity, and economic footprint. These metrics are computed over a rolling 60-day window, with a new snapshot generated at the end of each calendar month.

The rationale behind the 60-day window is strategic: in the context of a permissionless DeFi protocol, a two-month horizon offers sufficient time to distinguish between fleeting activity and meaningful participation. It balances short-term responsiveness with enough historical depth to assess user impact on protocol health. In other words, it captures both recency and continuity—two essential dimensions for understanding real behavioral commitment.

All three metrics are computed monthly using a rolling 60-day window. As noted in Tab 1, all swap data is analyzed at the user level, and pool-level swaps are not included in this dashboard. To be eligible for any of the behavioral analyses, users must have conducted at least 2 swaps on at least 2 distinct days during the corresponding 60-day window. This minimal threshold ensures we only evaluate users with recurring, non-random engagement, without applying stricter filters such as total volume or wallet tenure.

1. TWAD – Timing-Weighted Activity Density

TWAD captures how consistently and recently a user has engaged with the protocol. Each active day is weighted using the formula √(60 − distance_to_end), giving more credit to recent activity. This sum is then normalized by the theoretical maximum of 1.967—the ideal score for a user active across all 60 days.

Formula:
TWAD = SUM(√(60 − distance_to_end)) / 1.967
(The denominator represents the maximum theoretical activity weight, corresponding to a user who is active on all 60 days.)

Behavioral Segmentation:

A- Very Sparse (< 0.05), B- Sparse (0.05 – 0.20), C- Moderately Active (0.20 – 0.40), D- Consistent (0.40 – 0.70), E- Highly Active (≥ 0.70)

2. TAD – Transactions per Active Day

TAD measures behavioral intensity by calculating the average number of swaps per active day. This metric helps distinguish users who may appear moderately active overall but interact in short, concentrated bursts.

Formula:
TAD = Total Swaps / Active Days

Behavioral Segmentation:

A- Occasional Swapper (< 1.5), B- Light Swapper (1.5 – 2.49), C- Steady Swapper (2.5 – 3.49), D- Frequent Swapper (3.5 – 4.49), E- Power Swapper (≥ 4.5)

3. MSS – Median Swap Size

MSS evaluates the median USD value of each user's swaps during the 60-day window. By using the median rather than the average, this metric captures typical economic behavior and filters out one-off whale trades or noise.

Formula:
MSS = Median(USD Value per Swap)

Behavioral Segmentation:

A- Micro Swapper (< $15), B- Small Swapper ($15 – $99.99), C- Mid Swapper ($100 – $399.99), D- Large Swapper ($400 – $1499.99), E- Strategic Swapper (≥ $1500)

Unified Scoring and Normalization (Tab 4 - Composite Behavior Scores)

Together, these metrics form the behavioral foundation of Aerodrome’s intelligence-driven analysis. By profiling users across timing, intensity, and transaction value, they offer a multi-dimensional view of how individuals contribute to protocol performance—enabling more accurate targeting, incentive design, and retention strategies in future growth cycles.

In Tab 4 (Composite Behavior Scores), we normalize all metrics onto a [0, 1] scale to produce a unified scoring model. TWAD is inherently bounded between 0 and 1 by its mathematical design, requiring no further adjustments. However, TAD and MSS exhibit long-tailed distributions with extreme outliers. To ensure robustness and fairness, we applied statistical caps:

TAD is capped at 4.0, informed by percentile analysis and label alignment. This threshold preserves the behavioral integrity of the “Power Swapper” category without letting a small number of hyperactive wallets skew the scores.
MSS is capped at $1500, derived from distribution percentiles, boxplot upper fences, and behavioral label transitions. This ensures that economically significant users are rewarded without overrepresenting one-time whales.

These capped values are then scaled to fit the [0, 1] range and aggregated using a weighted scoring formula. The final score is computed for all eligible users in the latest 60-day snapshot, offering a holistic view of user quality based on three orthogonal behavioral dimensions.

Dashboard	Creator	Published
Aerodrome \| User-Level Growth May 25, 2025	Zanyar_98	May 25, 2025
iAgent AGNT and Node Holders Analysis Apr 15, 2025	optimus_prime	Apr 15, 2025
violent-crimson Dec 12, 2024	shaunoff	Dec 12, 2024