SourceCred makes the labor of individuals more visible and rewardable as they work together in a project or community. Taking inspiration from Google’s PageRank algorithm, SourceCred uses both objective and subjective measures to accurately capture and update an individual contributor’s impact on a project.
In traditional organizations, contributors to a project are at the mercy of their manager or boss to determine their value. This system of value determination occurs at set intervals and isn’t flexible or transparent. When contributors add more value than anticipated, all of the extra value accrues to their manager. In DAO structures, it can be challenging to keep track of the contributions of all members and to agree on the relative value of different contributions. This only becomes more difficult as a DAO scales.
SourceCred attempts to solve this problem via a compensation algorithm that relies on both subjective and objective measurements. Coordinape and SourceCred are similar tools, but Coordinape relies solely on subjective measurements.
SourceCred maps all participants and contributions in a graph, illustrating how contributions and participants in a project are connected to each other. Each contribution a participant makes is represented as a separate “node” in this graph.
This data is used to create a Cred Score for each node using a modified PageRank algorithm. The DAO decides which contributions will "mint" Cred, making them a source of new Cred.
Participants earn Cred by contributing to a project. A participant's Cred score reflects how valuable their contributions were. As an example, consider a Discourse forum post. The post is itself a contribution, and it generates Cred for its author, but it also generates Cred for any post to which it replies and/or for any posts or users that it references.
A key tenet for SourceCred is to let communities decide what they value. This is done through “weights” and these are established when setting up SourceCred. There are two kinds of weights:
Participants are rewarded with digital currency, called Grain, based on their Cred score. Grain is redeemable for USDC, a stable coin worth $1.
SourceCred uses Google’s PageRank algorithm as inspiration. It allows human contributors to be ranked based on objective measurements (counting # of commits, lines of code, etc) and subjective criteria (peer reviews, ratings system, etc).
Cred Scores retroactively update if a user’s contribution to a project continues to grow in impact over time (this is also true for negative Cred).
Setting up a SourceCred instance is fairly technical. Users will need to be comfortable using the terminal, installing packages from NPM, and be familiar with the basics of Git and GitHub.
For now, communities will need to be using at least one of the available plug-in platforms (Discord, GitHub, and/or Discourse) in order to get meaningful use out of the tool.
SourceCred is currently most effective with small communities with like-minded members. As more communities adopt and use SourceCred, the core team at SourceCred plans to build more robust anti-gaming tools.
Every SourceCred instance needs to leverage data about contributions. This information comes from the SourceCred plugins which “scrape” data from a platform to use in calculating Cred scores. Some users are uncomfortable with this practice.
If you have some technical skills and are using currently supported platforms (GitHub, Discourse, and Discord), take a look at How To Set Up SourceCred. Note that for Discord, you‘ll need the permissions to add a bot to the server.
If you’re interested in SourceCred but lack technical chops, you can still get involved on Discord/Discourse forums to learn more, ask questions, and contribute to the project.