Open Source by Default
outlining the unique advantages of OSS
Open source has been a popular movement in software for almost 20 years, but its business models, licenses, and perception have changed. To get basic definitions out of the way, open source software (OSS) is publicly accessible, meaning its source code is available for anyone to modify, view and distribute. The goal of the movement is to make software development more transparent and to allow any developer to contribute to a project.
With a new wave of open source businesses like GitLab, Hashicorp, and Confluent going public in the past few years, there are more examples of what the open source model can look like at scale. As these businesses reach 100s of millions in revenue, open source has quickly moved from the fringes to center stage in the software world. And like any other category of technology, more capital floods in as these new models get derisked.
This momentum does not seem to be slowing down either. Looking towards the future, IT leaders are expecting more of their software to be open source and less proprietary according to Red Hat’s State of Enterprise Open Source.
The same survey found that 77% of IT leaders claimed to have a more positive perception of enterprise open source than they did a year ago. It’s hard to get at the truth with surveys alone, but the sheer number of OSS businesses with meaningful revenue suggests open source is becoming the default in many purchasing decisions. It’s worth unpacking why.
WHY OPEN SOURCE
Why is it that developers, ML engineers, and other software users are choosing open source products instead of closed source options? The answer, of course, varies depending on the user and category, but I think it primarily comes down to 5 major advantages:
Community-driven development
Try-before-you-buy
Extensibility & Customization
Vendor independence
Security
I’ll caveat up front that these benefits are not exclusive to open source. For example, freemium products like Slack also let you try-before-you-buy via free trials or a free tier. Similarly, companies like Salesforce offer extensibility through their APIs, SDKs, and developer platform. Open source, however, is a more serious commitment to these principles as these are inherently true from the onset. For these reasons, choosing open source is partly philosophical (what is the best approach to software development) and partly utilitarian (what is the best product out there).
Community-driven Development
Even though “community” was one of those terms that lost meaning during the 2020-2022 tech boom, it really plays a major role in open source. Community-driven development refers to the way in which open source projects prioritize and build features. A community of users can openly flag issues, help build integrations or fix bugs, and contribute to the overall direction of the project on community calls. This type of bottom-up decision-making results in user-aligned decisions and a tighter loop around product development.
One often overlooked aspect of open source is its ability to harness the talents of a global developer base, not just those accessible through a company's sales efforts. Even though the majority of revenue for companies like Hashicorp and MongoDB comes from the US, the community is far more diverse.
Try-before-you-buy
The ability to try before you buy is one of the major draws of open source software. With OSS, users can download source code and run the software without any commitment or cost (outside of hosting). This allows users and businesses to evaluate the software in a real-world environment before even interacting with a commercial organization. Open source fits into the broader shift towards product-led growth where product usage drives software buying as opposed to just sales and marketing. Companies may add closed source features around collaboration or security, but it typically doesn’t inhibit users from trying the product out. This benefit carries weight in any category where the end user is also responsible for choosing the tools they prefer, which we’re seeing increasingly across software today.
Extensibility & Customization
One of the most significant advantages that you seldom see outside of open source is extensibility. Most software wraps a UI around code which makes it easier to use, but very challenging to customize or integrate with. As a result, users are forced into specific ways of working without being able to tailor the code to their own needs. For example, if a user needs a specific integration, they simply have to wait until it is prioritized by the vendor. The most straightforward example of this is in the ETL market. Unlike closed source vendors such as Fivetran or Matillion, Airbyte believes ETL needs to be open source to properly address the long-tail of integrations. The community is able to build whatever connectors they want independent of Airbyte. It’s a similar argument to why developers may want to use Appsmith vs a closed source internal tool builder as well. This benefit is most important in areas where there is a long tail of integrations or there is significant customer-to-customer variation in how the product is used.
Vendor Independence
Another important philosophical tenant of open source is its immunity to vendor lock-in. Preset founder, Maxime, paints the example of BI where M&A has shaken up the industry arguably for the worse. He notes that products get shut down (see Chart.io acquired and promptly shut down by Atlassian) or in other cases, consolidation results in aggressive pricing and limited innovation. With open source, you have multiple alternatives even if one company is removed from the ecosystem. For most large categories, alternatives exist from Amazon’s infinite open source offerings to Aiven to hosting the code yourself.
Security
This one is admittedly less clear cut as I’ll often see arguments made on both sides. One argument is that closed source is better because the general public and bad actors do not have visibility into the code. The security principle here requires you to put your trust into your software vendor. The argument in favor of open source is that it’s much better to expose the code so that a large community of developers can discover, audit, and fix vulnerabilities. The case for open source becoming more secure is stronger now with companies like Snyk helping developers find and remediate vulnerabilities in their open source dependencies. The security and privacy angle is particularly important in sectors like healthcare and financial services where the regulations and data privacy rules are particularly stringent.
A majority of users believe that open source software is usually better than proprietary software (OSS survey).
WHEN OPEN SOURCE WINS
Despite my excitement around open source, I don’t view it as a silver bullet to every product problem. There are certain product categories (namely within infrastructure) where I think open source shines and others where its advantages are less valuable.
Databases have become increasingly open source. This appears to be true across database types, from document stores like MongoDB to OLAP databases like DuckDB. The database is one of the cornerstones of any infrastructure stack, meaning developers have less appetite to be tied to just one vendor. They also may need to connect their database to various internal applications which requires manipulating the code directly. Lastly, databases are so widespread that they attract a large enough user base to get the network effects around the community.
Frameworks like Flutter for web development or Pytorch for ML also create many of the most popular projects. Often, these frameworks are extremely horizontal in nature and thus, attract a very wide audience of users who contribute to the project. While these frameworks are very popular, they are largely maintained by Facebook and Google rather than new startups. Projects from Jina AI or Haystack are examples of emerging search frameworks built by startups that are growing very quickly given their wide range of end use cases.
I’m sure there’s an exception to every rule, but I am generally less excited about OSS at the pure application layer. For example, productivity software is typically adopted by non-technical business users and relies on viral growth, where open source actually becomes more of a friction than a benefit. Another example is vertical SaaS, which doesn’t require extensibility in the same way that more horizontal, long tail problems do.
INVESTING IN OPEN SOURCE
From an investor lens, I view open source as an important methodology for how software can be built and distributed. While it’s not required in every category, we think it is becoming the default in many of the largest infrastructure markets. It also corroborates the idea that “code” is not defensible on its own. In other words, many of the advantages of open source are more of a wedge that helps attract early users and customers, but every business ultimately needs to cultivates its own modern software moats.
Every open source startup is different, but we typically look for the following progression across the rounds that we invest in.
Seed: open source project has some early adoption and is loved by its early users
Series A: open source project continues to grow with initial data points on willingness-to-pay for a paid product but sales motion is still unrefined
Series B: scaling beyond founder-led sales with early evidence of how open source product-led motion gets married with an enterprise sales motion at scale
If you’re building an open source business or generally interested in chatting about open source, feel free to shoot me a DM or email.