Contributor Details 🙇🏽
This documentation was contributed by Jim Cousins as part of his work at The Graph Foundation. Thanks also to the following community members for peer review and other contributions to this article:
- Aderks, Graph Indexer @ Maple Nodes
- Chris Wessels, Graph Indexer @ Graphops
- Gary Morris, Graph Council Member and Indexer @ Staking Facilities / 0xFury
- Juan Manuel Rodriguez Defago, Developer @ Bootnode
- Konstantin Zaitcev @ P2P
- Oliver Z, Governance Lead @ The Graph Foundation
- Payne, Indexer Relations, The Graph Foundation, Indexer @ Stakesquid
In this guide:
- Subgraph Selection: How Many Subgraphs Should an Indexer Serve?
- Modelling Potential Revenue
- Modelling on-chain Operational Expenses
- Subgraph Deployment Strategies
- Weighting strategies for allocation
- Weighting strategies for query revenue
- Balancing Allocation Deployment versus Query Fee Revenue Share
Subgraph Selection: How Many Subgraphs Should an Indexer Serve?
Modelling Potential Revenue
- How much revenue can an Indexer generate at a specific number of subgraphs
This is an Indexer-specific question, depending on two revenue streams:
Projected revenue from Staking Rewards: Staking or Indexing rewards are heavily dependent upon the amount of GRT allocated within the network. The example tables below show how Indexer revenue can be impacted as a larger amount of the total GRT in the market enters the protocol and is allocated.
Examples showing how effective network inflation rates impact Indexer revenue
Projected revenue from Query Business: The Indexer's knowledge of the market's appetite for query business on subgraphs at a specific price or price range
- This is challenging to predict because the network does not openly share any information about the query demand within the network
- Indexers have tools such as Agora to help understand how much query revenue they can make with a certain query cost model, however good projections require large volumes of real data in order to be useful
From those two revenue streams, how much is an Indexer willing to sacrifice to transaction fees?
- Staking rewards will form the lion's share of the business in the early network stage, so initially they will be responsible for funding the transaction fees
- Query fees will build over time as the number of subgraphs on the network increases
- The most effective (and therefore profitable) Indexers will understand the query market and other query mechanisms long before the network query volume begins to match the revenue generated via staking rewards
- Is it necessarily a sacrifice? Can those costs be claimed as a business cost or expense?
- Entirely dependant on jurisdiction - consult your own Tax Advisor.
How does the total projected revenue impact overall subgraph selection?
- Early-stage subgraph selection outcomes are based largely on Indexer size (self stake and delegation), which drives total revenue potential