Open source software (OSS) is often available at no cost. However, in recent years, we have seen a significant number of high-growth companies emerge with business models that incorporate OSS. These organizations are often categorized as Commercial Open Source Software (COSS) companies. They deliver value to customers by offering commercial versions of open source software, which has interesting implications when viewed within the context of higher education.
At the recent OSPO++ ‘Open Source Innovation in Universities’ event, Andrew Wichmann, Senior IP & Licensing Manager for Digital Technology at Johns Hopkins Technology Ventures(JHTV), discussed the rise in interest and investment in COSS technologies, as well as the business models emerging to support it.
Commercial opportunities for university open source
Wichmann explained that while the mission of most universities is to disseminate knowledge; the mission of the technology transfer office (TTO) at Johns Hopkins, “… is to translate technology out of the university and into the world”. He also noted that fulfillment of the TTO’s mission typically involves a commercial transaction.
In many ways, these two purposes go hand in hand, yet the conversation can become complicated when it comes to the commercialization of technology - particularly open source. Wichmann pointed out that to translate technology, it must be sustained, meaning it must work and be maintained in the world.
“Commercialization sustains open source,” Wichmann said simply. “There are other ways of sustaining open source … However, selling a product or service helps to offset the cost of building and maintaining a software product.”
From the perspective of JHTV, the overarching objectives of both TTOs and Open Source Program Offices (OSPOs) are aligned with commercialization of open source software and therefore should support it where possible.
Growing interest in COSS
Evidence shows that the market agrees. Since the growth of the open source community and founding of RedHat in the early ‘90s, OSS has gained a solid foothold in commercial technology. While the early days concentrated the commercial value of open source within RedHat, the past two decades have seen that change dramatically. Exponentially greater volume and value of COSS exists today; and it is spread across multiple private companies and organizations.
“There has been an increasing number of COSS companies or companies that are built upon or relate to some sort of open source project,” Wichmann explained. “One [reason] is that they’re very capital-efficient. You can find an existing project that’s free, find a customer base, adapt it for that use case, and then sell that as a product. This can all happen very rapidly.”
As OSS-based companies emerged, venture capital firms looked to support them. OSS Capital remains the only early-stage VC that exclusively invests in COSS companies, but many of the major VCs like Y Combinator, A16z and others have made significant investments as well. This points to increasing interest from investors in building companies around open source projects.
Emerging models for commercialization
Developing in parallel with investment are new models for COSS companies.
Services and support
The original and most straightforward model is services and support. In this model, the work and research create the open source project, but the commercial ‘product’ is the knowledge and experience that went into the project, offered for sale in the form of services or support of the software. This may take the form of providing commercial grade versions of the software, specialized implementations, warranties for the project, offering support and maintenance.
A key part of this product offering is trust and reliability. Companies will pay to have an expert on the software to help them leverage and maintain it.
Wichmann asserted that this type of commercialization was happening in universities from the beginning of open source, largely without the involvement of TTOs. For example, faculty who built and published an open source project as part of their research may consult with, work part-time for or offer additional services to companies who leverage the software.
“Because they built the project, the faculty know all the ins and outs of it,” he said. “You see this all the time in machine learning because [researchers] know how to tune it for specific use cases or how to help a company build it out into a version that’s going to work for their purposes.”
The next model to emerge was Software-as-a-Service (SaaS). This is similar to the original services model in that the software itself is open and free but it is packaged with a specific service offering that is the commercial product. For example, the company could sell storage, computing or a data management function that delivers value for customers. Wichmann noted that SaaS products increasingly have proprietary features or ‘shells’ around them that enhance the value of using the software.
While possible, it is highly unlikely this model will be used in higher education because the majority of universities are not-for-profit.
The more likely scenario for a university would be to license a project to an outside entity. TTOs and OSPOs often advise on licensing code for startups that can be created specifically for commercial use of an open source project. This is becoming a more popular route for entrepreneurial researchers and graduates.
A third model gaining popularity is the dual licensing approach. This model uses a typical open source license for the OSS project code plus a proprietary license to cover the commercial additions or exploitations of that original project. When the project is published, the researcher will use, for example, a copyleft license for the public project. Alongside that, they will offer a proprietary license for commercial use.
The dual licensing approach is common and straightforward for universities. If a faculty member has ideas for commercializing a project or an interest in receiving revenue from its commercialization; an external company can license the project under proprietary commercial terms.
Finally, the fourth model to emerge, “Open Core”, can be viewed as a blend of the other models.
“The basic idea is to build a public, open, active contributor base for an open source project and that the fundamental foundation of the project will remain open,” he explained. “But you build proprietary pieces of it depending on different criteria that you might select. And you can create broader or narrower versions of that. You can make it more or less proprietary.”
Unique to this model is the idea that decisions around what proprietary elements will be built and sold are guided by who will pay. For example, executives might pay for features while developers won’t. The future development of the project depends on what customer groups are identified and what return can be generated from them.
Additional opportunities for universities
At the end of his talk, Wichmann mentioned open source consortiums as an area for increased participation by universities. Such consortiums are typically industry-driven, with a specific company or industry sponsoring an open source research area with the end purpose of making the foundational work open source. “Rather than selling code or selling any sort of proprietary use rights or anything like that, you’re banking on time as being where the value is created.”
An open source consortium functions like an incubator where research and development is sponsored for a length of time, during which the participants and sponsors can track progress and identify commercial opportunities to pursue. Wichmann concluded, “When the project becomes publicized and available for use, the participating members have had time to develop their product lines and are first in the market with their solutions.”
While the reasons for making open source software vary widely within universities, it is clear that the growing interest in commercializing open source is increasingly influencing outcomes and next steps – largely fuelling the need for OSPOs in higher education and effective technology transfer processes.
View the complete OSPO++ event talk with Andrew Wichmann here.
This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0)