We talked to founder Roland Vogl to learn more about Merico in February 2019…
Tell us about your company. What problem are you trying to solve?
Merico is a code analytics company. We are developing algorithms to measure developers’ contributions to software projects.
Our technology offers a new approach for company managers to evaluate software engineers’ performance in an objective, transparent and efficient way. It can also help job seekers showcase their contributions to open-source projects, receive badges and list them on their resumes.
Additionally, we are building a solution for open-source projects to assign project income to developers based on their contributions. This model solves two critical problems for open source projects. First, developers currently do not fully participate in the value they contribute to the project. At this point, our solution is the only existing mechanism to distribute the income of an open source project to all its contributors. Second, developers currently do not get any downstream income for the code they created, which may be generate licensing income over a long period of time. The salary or the so-called bounties they were paid when they generated the code, are merely one-time payments. We feel that their intellectual contributions to a project can merit continuous, long-term returns, and we are providing them with a mechanism to receive royalties for their individual contributions to a project.
How did your company get started?
Our company originated from a moonlighting research project on quantifying the development value of code contributions. Our co-founder and CEO Jinglei was a Researcher at Microsoft Research when he met our co-founder and CTO Hezheng, a PhD student at UC Berkeley, at the wedding of a common friend (all of them are alumni of the Tsinghua University, one of China’s top-ranked university). Both Jinglei and Hezheng are very passionate about letting coder’s individual contributions stand out and enable them to capture the value of their code. To realize this ambitious vision, the our platform will have to provide an environment where participating developers from around the world, with different technical skills and cultural background, can trust each other. To facilitate this trust creation, we realized that we can leverage so called smart contracts on the blockchain. This in turn involves many practical legal issues. Consequently, Jinglei and Hezheng reached out to Roland, Executive Director of the Stanford Program in Law, Science and Technology and CodeX – The Stanford Center for Legal Informatics. Roland joined Jinglei and Hezheng, and the team worked together on this idea for over a year before deciding to start a company.
Why is your team well positioned to solve the problem you’re tackling?
Our work is based on insights gained from a rigorous multi-year research effort and on a team that can implement our big idea in an easy to use platform. We are seasoned researchers and experts with complementary skills in software systems, software engineering and legal informatics. We published the research underlying our company in a short paper entitled “Towards Quantifying the Development Value of Code Contributions” (https://per.pub/A1Pd7xaM.pdf) in the 26th ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering (ESEC/FSE ’18), a premier academic conference on software engineering.
What’s the biggest lesson you’ve learned along the way?
Can algorithms assess the value of individual software developer’s code contributions? We had a thesis that this can be done, but quite frankly, had our own doubts about the feasibility of this endeavor. But through our experiments we confirmed that algorithms can do exactly that with a performance level that is comparable to humans assessing developers’ code contributions. We surveyed open-source developers by asking them to compare the importance of code commits in pairs, which are either authored by themselves or others. If we view code authors’ comparisons as the ground truth, human accuracy is less than 75%. Our current algorithms can already achieve over 73%. Keep in mind that the human assessment is always influenced by subjective criteria and personal interests. We cannot guarantee our algorithms cover every corner case, but in the long run, we believe our algorithms can provide a more accurate assessment of the value of a developer’s contribution to software project.
What resources do you rely on to help your company grow?
We’re running a globally distributed team of developers. We use Shoobx to help us manage important HR and other important corporate processes.
We use other platforms to manage our work tasks and internal communication.
Aside from that, we rely on a network of very smart investors and advisors who are equally excited about our company’s mission, and who help us tremendously with their guidance and industry insight.