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

PancakeSwap is a non-custodial exchange protocol built on Binance Smart Chain and other blockchains (originally a fork of Uniswap). PancakeSwap was founded by an anonymous team and launched in 2020.

CAKE tokenholders own and govern the protocol. They currently take a cut of the total trading fees paid by traders (revenue).

governance

github

twitter

discord

etherscan

website

Exchanges (DEX)

PancakeSwap

# PancakeSwap Average fees per user (AFPU) Methodology

## 1. Overview

Average fees per user (AFPU) measures the average daily fees generated per unique active user on PancakeSwap. The average revenue generated per active user is reflected by this metric. The calculation is performed by dividing total daily fees by the number of active users (daily) to provide insight into user monetization and platform economics.

## 2. Data sources

- **Blockchain events**: `Swap` events from PancakeSwap V2 `Pair` contracts, `Swap` and `Flash` events from PancakeSwap V3 `Pool` contracts, `PairCreated` events from PancakeSwap V2 Factory contracts, and `PoolCreated` events from PancakeSwap V3 Factory contracts, covering Arbitrum, Base, BSC, Ethereum, and 2 other chains
- **External data**: End-of-day price feeds for assets, asset metadata (decimals and symbols), and a hardcoded list of denylisted trading pairs

## 3. Step-by-step calculation

### Component A: Trading pair identification

- **Purpose**: Establish the universe of valid PancakeSwap trading pairs and pools.
- **Logic**: Trading pair addresses are extracted from Factory contract creation events. For PancakeSwap V2, `PairCreated` events containing `pair`, `token0`, and `token1` fields are parsed with a fixed 0.25% fee tier. For PancakeSwap V3, `PoolCreated` events containing `pool`, `token0`, `token1`, and `fee` fields are parsed with variable fee tiers (0.01%, 0.05%, 0.25%, or 1%). These events are joined with asset metadata to obtain decimals and symbols for pair naming. Denylist filtering is applied to exclude pairs with limited price feed support.
- **Key fields**: Pool/pair `address`, `token0_address`, `token1_address`, `token0_decimals`, `token1_decimals`, `fee_tier`, `account_name`

### Component B: Trading activity extraction

- **Purpose**: Capture all swap transactions on PancakeSwap.
- **Logic**: `Swap` events are parsed from identified trading pairs. For PancakeSwap V2, `amount0In`, `amount0Out`, `amount1In`, and `amount1Out` fields are extracted, and net asset amounts are calculated as `amount0In - amount0Out` and `amount1In - amount1Out`. For PancakeSwap V3, `amount0` and `amount1` are extracted directly, along with `protocolFeesToken0` and `protocolFeesToken1` for revenue tracking. These events are joined with transaction data to obtain user addresses (`from_address`). Denylisted pairs are filtered out, and incremental processing logic is applied.
- **Key fields**: `block_timestamp`, `transaction_hash`, `account_address`, `user_address`, `token0_amount`, `token1_amount`, `token0_revenue`, `token1_revenue`

### Component C: Flash loan activity extraction

- **Purpose**: Capture flash loan transactions on PancakeSwap V3.
- **Logic**: `Flash` events are parsed from PancakeSwap V3 pools, and `amount0` and `amount1` fields representing flash loan amounts are extracted. These events are joined with transaction data to obtain user addresses. Denylisted pairs are filtered out, and incremental processing is applied.
- **Key fields**: `block_timestamp`, `transaction_hash`, `account_address`, `user_address`, `token0_amount`, `token1_amount`

### Component D: Trading volume calculation

- **Purpose**: Convert raw trading amounts to USD trading volume.
- **Logic**: Trading activity is aggregated daily by summing the absolute values of asset amounts. Amounts are normalized by decimal precision. Corrected amounts are converted to USD using end-of-day price feeds for each asset. Anti-manipulation logic is applied by taking the lesser of the asset0 and asset1 USD values as the final trading volume. For PancakeSwap V3, protocol fees are separately tracked using the same normalization and USD conversion process.
- **Key fields**: Daily aggregated `trading_volume`, `revenue` (PancakeSwap V3 only)

### Component E: Flash loan volume calculation

- **Purpose**: Convert raw flash loan amounts to USD volume.
- **Logic**: Flash loan activity is aggregated daily by summing the absolute values of asset amounts. Amounts are normalized by decimal precision and converted to USD using end-of-day price feeds. Both asset0 and asset1 USD values are summed to obtain the total flash loan volume.
- **Key fields**: Daily aggregated `flash_loan_volume`

### Component F: Fee calculation

- **Purpose**: Calculate total fees generated from trading and flash loan activity.
- **Logic**: Trading volume is multiplied by the corresponding fee tier percentage for each pool (0.25% for PancakeSwap V2, variable for PancakeSwap V3 ranging from 0.01% to 1%). Flash loan volume is multiplied by the corresponding fee tier percentage. Trading fees and flash loan fees are combined using a union operation to obtain total daily fees.
- **Key fields**: Daily aggregated `fees`

### Component G: Active users (daily) counting

- **Purpose**: Count unique active users (daily) across trading and flash loan activities.
- **Logic**: Distinct user addresses are extracted per day from both trading and flash loan fact tables using a union distinct operation. All users are classified as 'Trader' type. Unique users are aggregated daily across the dataset. Dimensional breakdowns are created by project, chain, user type, and product to enable flexible aggregation.
- **Key fields**: Daily aggregated `user_dau` counts

### Final aggregation

- **Combining components**: Daily fees data is joined with active users (daily) data using the `calculate_arpu` macro, which performs the division of fees by user count.
- **Currency conversion**: All amounts are converted to USD using end-of-day price feeds at the volume calculation stage.
- **Time aggregation**: Data is aggregated to daily boundaries using timestamp truncation to the day level.
- **Final calculation**: AFPU is calculated as total daily fees divided by active users (daily), with proper handling of zero-user days and missing data through the `finalize_metric_afpu` macro which applies fill strategies for complete time series.

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

PancakeSwapCAKE

Average fees per user (AFPU)

Average fees per user (AFPU)

Competitive landscape