GitLab Inc. is a for profit company that balances the need to improve the open source code of GitLab with the need to add source-available features in order to generate income. We have an open core business model and generate almost all our revenue with subscriptions to paid tiers. We recognize that we need to balance the need to generate income and with the needs of the open source project.
We have tried different business models, and many didn’t work. As a company, we realized we needed recurring revenue to continue our mission, and we introduced source available code that is proprietary. Initially there was a worry we would stop working on the open source code but the community saw we were able to accelerate the work on open-source code too.
We promise that:
Our stewardship promise applies only to the GitLab open source codebase, not to GitLab, Inc services such as GitLab Software as a Service (SaaS). For GitLab SaaS, we may add limits around usage (e.g., storage, compute, traffic) or proxies for usage (e.g., number of users in a group) as we seek to optimize costs and revenues. Our stewardship promise allows other companies and organizations to provide a SaaS offering based on the GitLab open source codebase.
To determine what is open source and what not we ask ourselves: Who cares the most about the feature.
If the likely buyer is an individual contributor the feature will be open source, otherwise it will be source-available (proprietary).
There aren't any features that are only useful to managers, directors, and executives. So for every source-available feature there will be an individual contributor who cares about it. We're not saying that there aren't any individual contributors that care about the feature, just that we think that other buyers are relatively more likely to care about it. The more of GitLab that you use the more likely it is that you benefit from a higher tier. Even a single person using GitLab might be best off using our highest tier.
It is hard to get the tier right, and if we put something in a tier that is too high we won't hesitate to open-source it or move it to a lower tier. We listen to our community in order to find what we feel is the right balance, and we iterate and make changes based on their feedback. At the same time, the premium product needs to hold value, and we believe we provide that.
All stages of the DevOps lifecycle have at least some features available in GitLab Core. There are companies using GitLab Core with more than 10,000 users.
If people ask us why a certain feature is paid we might reply with a link to this section of the handbook. We do not mean to imply you don't need the feature. Feel free to make the argument for moving it to another tier, we're listening.
If the wider community contributes a feature that is currently source-available we use the process linked in Contributing an existing feature to open-source it.
Sometimes people suggest having features in EE for a limited time. An example of a limited time release strategy is the Business Source License that keeps features proprietary for 3 years.
At GitLab we want to give everyone access to most of the features (and all the essential ones) at the date they are announced. We want to give people the option to both run and contribute to an open source edition that is maintained and that includes the most recent security fixes.
From time to time we do open source a feature that was previously proprietary. We do this when we realize we made a mistake applying our criteria, for example when we learned that a branded homepage was an essential feature or when we brought GitLab Pages to the Community Edition.
Our plan is to become the most popular tool for people’s own git hosting service; we’ve managed that so far. Secondarily, we want to get to be the one with the most revenue. Thirdly, we want to become the most popular tool for hosting private repos. Once we’ve reached that, we want to be the most popular tool for hosting public repos. And, lastly, we want to be the number one tool for people to host not just code but books, tech papers, visual models, movies, etc. More info on this is on our strategy page.
GitLab Inc. has an open core business model that includes source-available code and selling subscriptions. This benefits the open source part of GitLab in the following ways:
There are valid criticisms of Open Core. At Open Core Summit in 2019, Deb Bryant of Red Hat highlighted the following four:
Participation Is Constrained. Open Core companies may limit participation to keep premium features from being contributed. - We actively work to grow participation since our missions is that Everyone can Contribute. Also, we do not say no by-default to having existing paid features contributed to our open source project.
Vendor Lock-In. Premium features make it more difficult to switch workflows. - GitLab the product plays well with others. As we outline,
Many other applications integrate with GitLab, and we are open to adding new integrations to our technology partners page. New integrations with GitLab can vary in richness and complexity; from a simple webhook, and all the way to a Project Service. GitLab welcomes and supports new integrations to be created to extend collaborations with other products. GitLab plays well with others by providing APIs for nearly anything you can do within GitLab. GitLab can be a provider of authentication for external applications. GitLab is open source so people are very welcome to add anything that they are missing.
Community Is Devalued. Community is seen as a marketing tool instead of as participants working to make the product better together. - We care deeply about our community and depend on all GitLabbers to help us improve our category and stage maturity. We have Merge Request Coaches who help contributors get their merge requests to meet the contribution acceptance criteria, and wider community contributions per release is a GitLab KPI.
Deployment Constrained. Customers are afraid of confusing or unsanctioned licenses entering a business environment. - We don't use a limiting license such as SSPL. As we highlight in our docs, GitLab Community Edition (CE) is licensed under the terms of the MIT License, which is an official Open Source license as defined by the Open Source Initative.
We never move existing features already in GitLab FOSS into a paid tier. This applies regardless of whether the feature was created by GitLab team members or the wider community of contributors. On occasion, the reverse does happen where we open source a previously source-available feature.
On occasion, we might change the product in a way that bundles functionality together or refactor functionality that was in different tiers. Sometimes the new bundle will go to the Core/Free tier and sometimes to the paid tiers.
When someone contributes an existing source-available feature to open-source code base, we have a hard decision to make. We encourage contributors to focus on new features not already existing, so that both codebases of GitLab benefit from the feature, and we can avoid any difficult decisions.
We encourage contributors visit our direction page to see features that are welcomed and also review CONTRIBUTING.md.
When someone contributes an existing feature to open-source it, we weigh a number of factors to decide in accepting it, including:
We'll weigh all factors and you can judge our stewardship of the open source codebase based on the outcome. As of Dec 4, 2018, we had only two cases: One had low code quality and the other one copied the source-available code down to the last space. If you find these or other examples please link them here so people can get an idea of the outcome.
When someone contributes a not yet existing feature on the issue tracker that has a paid tier label, and it has met the contribution acceptance criteria, we will accept it under whatever license (open-source or source-available) they prefer, provided that GitLab Inc. has not already started working on the feature. (The contribution should not contain any already existing source-available features in it.) We encourage contributors to @-mention the relevant product manager earlier in the development process (in the issue or merge request) to ensure GitLab team-members are not already working on the feature in order to avoid conflicts.
Some companies try to support "20% time" for contributing to open source projects or other self-directed work. While this has been discussed several times within GitLab, we intentionally do not implement something like this for the following reasons:
Instead of optimizing locally across many different projects, we are able to create much more value and impact for the open source community by singularly focusing on our vision. Doing so creates a virtuous cycle that enables us to:
Make More Improvements Upstream. GitLab is open core built entirely with open source. We rely on thousands of dependencies and always try to contribute our improvements upstream.
Build More Open Source As Part Of GitLab. Shipping as much open source as possible is critical to our business model because it increases the velocity of our flywheels. The faster they spin, the more we can contribute back.
Contribute At Scale. GitLab is a single application for the entire DevOps lifecycle. By freely providing the top tier of GitLab to open source projects, we are enabling them to be more efficient, secure, and productive.
The combination of the above leads to global optimization on a much larger scale than anything we could contribute with a small portion of our time working on things that don't directly move our vision forward.