Overcollateralized borrowers deposit funds into Sky and receive DAI stablecoin.

Sky is a non-custodial stablecoin protocol built on Ethereum. Sky (previously MakerDAO) was founded by Rune Christensen and launched in 2015.

SKY tokenholders own and govern the protocol. They currently take all of the total interest and PSM fees paid by borrowers (revenue). Sky’s investors include a16z, Polychain, IOSG Ventures, etc.

governance

github

twitter

discord

etherscan

website

Stablecoin issuers

Sky | Fees and revenues breakdown

# Sky Active users (monthly) Methodology

## 1. Overview

Sky Monthly Active Users (MAU) measures the number of unique users who interact with Sky's vault system within any 30-day rolling window. This metric tracks users across different activity types (borrowing, repaying, depositing, withdrawing, and trading) and provides dimensional breakdowns by chain, user type, and product. The calculation uses HyperLogLog sketching for efficient cardinality estimation across the rolling window.

## 2. Data sources

- **Smart contract function calls**: Sky's Vat contract `frob` function calls from Ethereum, which represent all vault modification events including debt changes and collateral adjustments
- **Smart contract function calls**: Sky's Vat contract `grab` and `fold` function calls from Ethereum, used for collateral type identification and metadata
- **Blockchain transactions**: Ethereum transaction records providing user address identification through the `from_address` field
- **Contract metadata**: Collateral type definitions and classifications, particularly distinguishing PSM (Peg Stability Module) collateral from traditional vault collateral

## 3. Step-by-step calculation

### Component A: Collateral type classification

- **Purpose**: Identifies and categorizes different collateral types in Sky's system
- **Logic**: Extracts unique collateral identifiers from Sky Vat contract function calls (`frob`, `grab`, `fold`). Converts hex-encoded collateral IDs to human-readable names by removing hex prefixes and null bytes, then decoding to UTF-8 strings. Creates a lookup table distinguishing PSM collateral types (used for trading) from traditional vault collateral (used for borrowing/lending)
- **Key fields**: `collateral_id` (hex identifier), `collateral_name` (decoded string), `chain` (always ethereum)

### Component B: User activity extraction

- **Purpose**: Identifies user interactions with Sky vaults and categorizes activity types
- **Logic**: Processes Sky Vat contract `frob` function calls to identify vault modifications. Categorizes users based on parameter changes: users with positive `dart` (debt changes) become Borrowers, negative `dart` become Repayers, positive `dink` (collateral changes) become Depositors, negative `dink` become Withdrawers. Users interacting with PSM collateral types are classified as Traders regardless of parameter changes. Excludes PSM collateral from traditional borrowing/lending categories to avoid double-counting trading activity. Joins with Ethereum transaction data to identify user addresses through the transaction `from_address` field
- **Key fields**: `user_type` (Borrower/Repayer/Depositor/Withdrawer/Trader), `user_address`, `collateral_id`, `product_id` (usds), `block_timestamp`

### Component C: Daily active users calculation

- **Purpose**: Aggregates unique users per day across all activity types
- **Logic**: Groups user activity data by day using timestamp truncation to daily boundaries. Counts distinct user addresses per day, maintaining dimensional breakdowns by chain (ethereum), user type, and product. Creates the foundation dataset for rolling window calculations
- **Key fields**: `timestamp` (daily), `user_dau` (distinct user address), `user_type`, `product_id`, `chain`

### Component D: Rolling window preparation

- **Purpose**: Sets up the infrastructure for 30-day rolling MAU calculation
- **Logic**: Applies HyperLogLog initialization to daily user counts for efficient cardinality estimation. Creates a complete datespine covering all days from the earliest data to present, ensuring no gaps in the time series. For each date, generates a 30-day lookback window by creating join relationships with the previous 29 days plus the current day
- **Key fields**: `hll` (HyperLogLog sketch), `timestamp`, `joinday` (dates in rolling window)

### Final aggregation

- **Combining components**: Joins the rolling window datespine with HyperLogLog-enabled daily user data, creating 30-day windows for each date. For each window, merges HyperLogLog sketches across all days to calculate the distinct count of users active within that 30-day period
- **Currency conversion**: Not applicable - this is a user count metric
- **Time aggregation**: Uses daily timestamps as the base unit, with each day representing the end date of a 30-day rolling window. Monthly partitioning is used for data management efficiency
- **Final calculation**: Extracts the distinct user count from merged HyperLogLog sketches using `hll_count.extract(hll_count.merge_partial(hll))`. Creates multiple dimensional breakdowns: project-level (all dimensions combined), chain-level, user type-level, and product-level. Combines all breakdowns into a unified result set with appropriate null values for non-applicable dimensions

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

Unique revenue-generating users over a rolling 30-day window.

SkySKY

Active users (monthly)

Active users (monthly)

Competitive landscape

Chain breakdown