Liquidity providers deposit funds into HMX and receive trading fees from traders (supply-side fees).

HMX is a non-custodial exchange for perpetual contracts built on Arbitrum. HMX was founded by Shu Wei, Rockman, and Taoree and was launched in 2022.

HMX tokenholders own and govern the protocol. They currently take a cut of the total trading fees paid by traders (revenue). HMX's investors include Coral DeFi, Compound Capital Partners, Cryptomind, etc.

github

twitter

discord

etherscan

website

Derivative exchanges

# HMX Average revenue per user (ARPU) Methodology

## 1. Overview

Average Revenue Per User (ARPU) for HMX measures the daily revenue generated per active trader on the HMX perpetual trading platform. This metric answers how much revenue the protocol generates on average from each daily active user, calculated by dividing daily protocol revenue by daily active users. The calculation focuses specifically on users who actively trade (open or close positions) and revenue from three fee types: borrowing fees, liquidation fees, and trading fees.

## 2. Data sources

The calculation uses the following raw data sources:
- **Arbitrum blockchain event logs**: Parsed `LogIncreasePosition` and `LogDecreasePosition` events from the HMX TradeService contract to identify active traders
- **Arbitrum blockchain event logs**: Parsed `LogSettleBorrowingFeeValue`, `LogSettleLiquidationFeeValue`, and `LogSettleTradingFeeValue` events from the HMX TradeHelper contract containing fee amounts
- **Protocol configuration data**: HMX-specific revenue percentage splits that determine what portion of each fee type goes to the protocol versus liquidity providers
- **Time-based dimension data**: Historical changes to revenue percentage splits with specific timestamps when fee structures changed

## 3. Step-by-step calculation

### Component A: Daily active users calculation
- **Purpose**: Count unique daily active users who trade on HMX perpetual trading platform
- **Logic**: Extract user addresses from Arbitrum `LogIncreasePosition` and `LogDecreasePosition` events emitted by the HMX TradeService contract. These events capture when users open new positions or modify existing positions. Block timestamps are truncated to daily boundaries and users are classified as "Trader" type. The calculation uses HyperLogLog aggregation to efficiently count distinct users across multiple dimensions (chain, user type, product). Users are aggregated at project level, chain level, user type level, and product level to support different breakdowns.
- **Key fields**: `primaryAccount` (user address), `block_timestamp` (for daily grouping), transaction metadata for deduplication

### Component B: Fee collection and processing  
- **Purpose**: Extract and process three types of fees from HMX perpetual trading
- **Logic**: Parse three separate fee types from Arbitrum TradeHelper contract events. Borrowing fees come from `LogSettleBorrowingFeeValue` events with `feeUsd` field, liquidation fees from `LogSettleLiquidationFeeValue` events with `usd` field, and trading fees from `LogSettleTradingFeeValue` events with `feeUsd` field. All fee amounts are denominated in USD with 30 decimal places and are converted to standard decimal format by dividing by 10^30. Block timestamps are truncated to daily boundaries for aggregation.
- **Key fields**: `feeUsd` and `usd` (fee amounts), `address` (account address), `block_timestamp` (for daily grouping)

### Component C: Revenue percentage application
- **Purpose**: Apply protocol-specific revenue splits to convert total fees into protocol revenue
- **Logic**: Use time-based revenue percentages to determine what portion of each fee type becomes protocol revenue versus liquidity provider fees. The percentages are: borrowing fees 10% (constant), liquidation fees 100% (constant), and trading fees that changed from 10% to 35% on 2023-08-10. Fee data is joined with the revenue percentage dimension table based on daily timestamps to apply the correct percentage for each time period. Revenue is calculated by multiplying fee amounts by their respective revenue percentages.
- **Key fields**: `borrowing_revenue_percentage`, `liquidation_revenue_percentage`, `trading_revenue_percentage`, daily timestamp for matching

### Final aggregation
- **Combining components**: Daily revenue is calculated by summing all three fee types (borrowing, liquidation, trading) after applying revenue percentages, aggregated by day, chain, business line, and other dimensions. Daily active users are counted using HyperLogLog aggregation across the same time boundaries.
- **Currency conversion**: Fee amounts are already USD-denominated in the source events, requiring only decimal adjustment from 30-decimal to standard format
- **Time aggregation**: Both revenue and user metrics use daily boundaries created by truncating block timestamps to day level, ensuring consistent time alignment between numerator and denominator
- **Final calculation**: ARPU is calculated by dividing daily revenue by daily active users using the `calculate_arpu` macro, with zero-filling applied for days with no activity to maintain continuous time series. The result represents average USD revenue generated per daily active trader.

*This methodology was automatically generated by Token Terminal's methodology analyzer.*

HMXHMX

Average revenue per user (ARPU)

Average revenue per user (ARPU)

Competitive landscape