secondary title

Why Community Matters

Open source projects are fundamentally driven by community input, even if most contributors never code the final project. Many open source projects are maintained by a small core of developers, and when it comes to actually writing the code and guiding the project, the community still plays a fundamental role in the project through bug reports, product roadmap feedback, and feature request submissions.

We believe that the community provides the greatest value to open source projects by helping them spread information about the project organically, improving the reputation of the product, and ensuring that the project grows according to the needs of its users.

secondary title

which metrics are important

Accurately measuring success for something as qualitative as a community is difficult, especially given all the different stakeholders and relevant metrics.

For our Bessemer investment target, we are most concerned about the number of community users and contributors, because they are the group with the most insight into the scale of the community. That's why we rarely pay attention to numbers like Github Stars, which, like other vanity metrics, tend to spike with big press releases, and because they don't reflect ongoing engagement, they have some value. degree of deception.

Users and contributors represent groups that are actively involved in and rely on the project, but given that most projects rely on telemetry by their users Limited, so true active users are difficult to measure.

Contributors, on the other hand, represent only a small fraction of users, but this group is easier to measure. These users tend to engage more deeply with the project by investing time in providing feedback in the form of issue comments, or occasionally contributing code to the project. We don't use the number of contributors to measure a project's development prowess, instead we use it as a proxy indicator of how much adoption the project is getting.

secondary title

What does optimal engagement look like

When discussing the largest open source project communities, it's impossible not to bump into projects from the tech giants. The top five open source projects of all time are from Google, Microsoft, and Facebook. These projects play a fundamental role in modern software development and cloud infrastructure, and they have built incredibly large communities, each with more than 1,000 contributors per month. Interestingly, each project has become one or two of the most popular technologies in their respective categories, including VSCode (Note: an open source free cross-platform code editor, widely used IDE), Kubernetes (a container orchestration program open sourced by Google engine, which supports automated deployment, massive scalability, application container management) and React Native and Flutter (two of the most popular cross-platform front-end frameworks).

However, once we remove projects backed by big tech companies (which are less relevant to independent open source projects) and instead focus on the top 10 independent projects related to affiliated commercial startups, the graph looks would look very different:

In terms of engagement, the Moby project (the underlying project behind Docker), and more recently Gatsby, are significantly higher than the average for the rest of the projects on this chart. The average number of monthly contributors for the rest of the projects in the table is distributed between 200 and 600, which is definitely the top level in its category. Of the top 10,000 projects we analyzed, less than 5% had more than 250 contributors per month, and only 2% consistently had more than 250 contributors for 6 months or more.

Measuring contributor activity is far from a science, but in our analysis we found some trends:

• Variations in Maturity - Most projects experience strong growth in their early years, and after a peak in community activity, participation levels off as the project reaches full maturity. With this in mind, we take a project's age and maturity into account when evaluating engagement versus benchmarking.

• Differences by type - some types of projects (e.g. databases) attract far fewer contributors than others (e.g. front-end frameworks) which have larger teams of developers able to contribute to them contribute. Therefore, in order to compare the differences between projects, it is important to adjust for these variables by segmenting them by software category, which is discussed in more depth below.

• Skewed - Contributor activity is heavily skewed towards the largest projects. While the largest projects often have more than 1000 contributors per month, very few projects have fewer than 100 contributors per month. Given our bias toward larger projects, we focus more on new projects meeting specific thresholds that are consistent with their phase and type.

There are also countless nuances to a project that can affect whether it can attract a large number of contributors, including how well organized the project is, how welcoming it is to new contributors, and how broad the community is. For analysis purposes, we focus on one of the biggest factors: how many developers are the project's potential users? For example, we know that far more developers are capable of contributing to a front-end project like Font Awesome than there are developers contributing expertise to ElasticSearch. Therefore, the number of contributors we expect for database projects will be significantly lower than the number of front-end projects.

The graph above shows project engagement over time, illustrating the difference in size between popular front-end projects like React Native and Font Awesome, and popular database and analytics engines like ElasticSearch and Apache Spark.

While this is far from a perfect evaluation system, we have attempted to develop a benchmark set that accounts for project type differences by combining a benchmark set of the top 10 to 20 projects in four major technology types (frontend, backend, devop, and database) benchmarks to help us compare projects to relevant technologists. The chart above shows the range of monthly contributor engagement in these top projects.

Here are some conclusions we drew from our analysis:

In order to make these benchmarks more widely applicable, we can look at some current projects, including Cypress and Netdata, which we invested in, to understand the results of some popular projects compared with relevant benchmarks today.

A top-notch benchmark can be a useful guide, but few projects achieve benchmark-scale engagement. Frankly, to a certain extent, more participation does not necessarily indicate that a project or company will be commercially successful.

If some emerging open source projects can sustain more than 100 contributors per month, they will be in the top field of the industry. If the number of monthly contributors exceeds 250, they will be close to the achievements of the most active projects in history. In fact, only 6% of the top 10,000 projects maintain 250 contributors per month for 6 months or more.

secondary title

The future of software is community-driven

We are in a stage where software development is increasingly driven by the community, and the line between commercial software and open source software is blurring. We hope that more and more companies will open source their core technologies, because open source can give the community Bring common interests with the company and focus on how to achieve profitability. As Bessemer's investments in these types of companies continue to increase, our research on open source project metrics will continue as we have more data sources to include in our analysis and better refine our metrics.