Most open source companies can be categorized into five main business models. In part one we reviewed the Ketchup Model and the Dual License Model. In this part, we review the Proprietary Crust Model, the Infrastructure Model, and the Adjacency Model.
The Proprietary Crust Model
The Proprietary Crust Model is a business model based on offering open source software with proprietary features or interfaces on top. In this model, the majority of the software is open source, but there are proprietary features only available to paying customers.
One common version of the Proprietary Crust model is an “open core” company. An open core company usually has a “community version” with limited or base functionality. The company also has plugins or an “enterprise version” with additional features. These features are usually specialized to enterprise needs, such as manageability, scalability, and policy compliance.
Another way that Proprietary Crust companies distinguish between open and proprietary components is by considering fixed vs. marginal costs. Shared infrastructure is fixed cost, so it makes sense to spread out the cost as widely as possible. A marginal cost component might only benefit one (or a small number) of users. These niche components may be more specialized, and thus valuable as proprietary add-ons.
Redis Labs is an example of a Proprietary Crust company. The Redis server itself is permissively licensed and available for anyone. Redis Labs has bundled the server with additional modules and connectors as part of an “Enterprise” version available for free or as a hosted service.
Many traditional-seeming companies are Proprietary Crust companies. For example, most of the system-level software underlying Mac OS X is open source. The GUI, the interfaces, and many of the top-level applications are strictly proprietary. Apple avoids copyleft licenses and presents its products as wholly proprietary, but a glance at Apple’s open source disclosures reveals that they are deeply invested in open source.
The Proprietary Crust Model can lead to confusion among customers who want to know what is open source and what is not. The most valuable parts of the software are usually proprietary, making it less inviting for others to use the open source portions alone. This can reduce ecosystem interest and slow growth. The proprietary components aren't accessible to the open source community, so they don't receive the same development and improvement. On the other hand, the open source community may create copies of proprietary addons with permissive licenses, reducing the market for the paid versions.
Benefits:
- Allows companies to monetize their open source software by offering proprietary features
- Provides companies with a way to differentiate their product in a crowded market
- Ties company revenues to easily-understood value
Drawbacks:
- Open source community members may create replacements for proprietary functionality
- Can limit the spread of the software, as proprietary users may not be as likely to contribute to the open source community
- Can create confusion for users, as they may not be sure which features are open source and which are proprietary
The Infrastructure Model
Companies using the Infrastructure Model provide value by managing and running the open source software for the customer. The customer pays for the convenience of having the software professionally managed and not having to manage it themselves.
One example of this model is Amazon Web Services (AWS), which provides many cloud computing services based on open source software. Customers pay AWS for the use of the software, but also for the infrastructure, security, and support provided by the company. WPEngine is another example, providing hosting for WordPress sites.
This is an increasingly common model for open source companies because it has good economics - SaaS subscriptions create very predictable revenue streams. It also keeps the company and the associated open source community mostly aligned.
One possible downside is that other companies may outcompete you using the same software. For example, Rackspace created OpenStack to compete with AWS and planned to make money from hosting and its “fanatical support.” Thirteen years later, many big companies are using and selling OpenStack services. But Rackspace struggled to make money and had to change their strategy.
Similarly, Docker pioneered its popular eponymous container format. But Docker had to shift its business model away from offering its own infrastructure services, as competitors were able to offer the same services using Docker's open source technology.
Benefits:
- Provides a way for companies to monetize their open source software without limiting its spread
- Predictable ongoing revenues from subscriptions
- Can be well-aligned with open source community development
Drawbacks:
- Other companies can compete using the same underlying open source software
- Creates incentives to make the software harder to install or run, limiting community use
- Requires development of a brand outside of the software brand
The Adjacency Model
Companies using the Adjacency Model create value through products or services that are "adjacent" to the open source software maintained or supported by the company. These adjacent products or services can be directly related, like support and consulting, or they can be indirect, like websites created using open source software like WordPress.
One common adjacency is hardware. Companies like Arista use Linux as the software base for their routers and switches. The open source software makes Arista’s hardware more valuable.
The logic behind the Adjacency Model is related to the concept of "commoditizing your complements," popularized by Joel Spolsky. In economics, “complements” are products that are usually used together, such as cereal and milk or printers and ink. Both products have a cost. If your company makes cereal, then it is good for you when milk becomes less expensive. Less money for milk means that there will be more money for cereal. Another increasingly important adjacency relates to machine learning. Companies can use open source software to generate data that can be mined for valuable insights that can be packaged and sold.
Benefits:
- The open source nature of the software can be a “moat” around the more profitable adjacent product
- Can be well-aligned with open source community development
- Provides a way for companies to monetize their open source software without limiting its spread
Drawbacks:
- Money spent on the open source project may not be perceived as directly adding value to the business
- Other companies can compete using the same underlying open source software
- May have lower margins than SaaS or software sales
Choosing a Model
For some companies, a single business model drives the whole business. For others, it makes sense to think about the business model at the level of individual products. Even companies that have majority-proprietary businesses can have successful standalone open source products with their own models.
When choosing a model, there is no need to be exclusive. Many companies have businesses that mix two or more of these models. For example, Cloudera has spent considerable resources on branding their particular distribution of Hadoop and libraries (a Ketchup Model product). They also provide customers a fully-managed data platform (an Infrastructure Model product), support and services subscriptions (an Adjacency Model Product).