Open Core : How Did We Get Here?
14 Jul 2019 •OPEN-SOURCE
Open source is considered an exemplar of the ‘private-collective’ model of innovation,1 a compound model with elements from both the private investment & the collective action models.
This model was an attempt to rationalise and reason about the existence of the open source software industry, and answer the question: “why would thousands of top-notch programmers contribute, without apparent material incentives, to the provision of a public good?”.2
This essay revisits the assumptions of the private-collective model, in the cloud-compute era, to understand the surgent phenomena of the open core revenue model in the commercial open source software industry. This is of particular significance in view of the perceived siege of the open source model by cloud vendors.3
Private-Collective Model
A public good produced under the private-collective model has two classes of users: innovators and free-riders. Innovators are users of the the good, and active contributors to its sustained development. Free-riders are passive consumers of the good, with no contribution to the advancement of the good.
For the sustenance of an open source product, it is imperative that the innovator class has incentives beyond what is available to the free-rider class of users. It is a myth that the development of a public good can be sustained on the basis of pure altruism from the innovator class of users.
In case of individual innovators, the incentives, as postulated by the model, are:
- learning, which is an immediate incentive.
- signaling incentive, which is a form of delayed incentive.
- career concern incentive: in the form of future job offers, shares in commercial open source corporations, or future access to the venture capital market.
- visibility incentive: in the form of peer recognition.
For an organisation to participate as an innovator, a different set of incentives could be at play. Unlike the case of an individual contributor, the incentives are harder to classify, and much-less researched at scale. A majority of organisations have looked at the incentive structure of open-source from the same prism as an individual contributor. Open source provides organisations signaling incentives to attract talent, and visibility in the community (again, this helps in hiring).
Other delayed incentives traditionally identified for initiator organisations of OSS can take the form of complementary services.4 That is, the company expects to boost its profit on a complementary segment to the open source project (For example, technical support on the OSS product). This is popular as the Redhat business model. A caveat to this strategy is that the increase in profit in the proprietary complementary segment should offset the profit that would have been made in the primary segment, had it not been converted to open source.
Very few incentives beyond this have been identified and utilised by organisations, till recent times. This explains why organisations do not feel compelled to open source proprietary products on which the revenue of the company has a direct dependence.5
A recent wave of COSS (Commercial Open Source Software) organisations are trying to change this fact. A reason for this phenomenon is an hypothesis that has emerged in the industry that the open source process is a way for a small-scale organisation to use the diffusion networks associated with open source to take on a dominant player in the industry.6
The Cloud Era
The experimentation with open source as a business & product delivery model by COSS organisations is also becoming contemporaneous with the near-monopolistic rise of cloud vendors. In the private-collective model, the existing cloud vendors would be classified as free-riders. The zero-sum (winner-takes-all) nature of the cloud model7 means cloud vendors have little motivation to contribute back to the community, while capturing a disproportionate share of the value generated by an open source product.
The confluence of these two trends is the reason for the emergence of the open core model. The first generation of open core business models are simple tweaks to the open-source licenses designed to defend against cloud providers 8. And to be clear, the use of licenses to defend proprietorship is not much of an innovation. MySQL dual licensing has existed for a decade prior to the cloud era.
As with other innovations of business models happening in the COSS industry, open core is an emergent phenomena, and its efficacy and implications are yet to be understood in full measure. Its efficacy will be judged in terms of the degree of sustainability it brings to the COSS industry; while its implications with respect to the spirit of open source and free-sharing of software (and how far removed it is from that ideology) is up for debate.
P.S. A draft of this essay was published here, as part of the Collectiive initiative.
Footnotes
-
Hippel, Eric von, and Georg von Krogh. “Open source software and the “private-collective” innovation model: Issues for organisation science.” organisation science 14.2 (2003): 209-223. Link ↩
-
Lerner, Josh, and Jean Tirole. “Some simple economics of open source.” The journal of industrial economics 50.2 (2002): 197-234. Link ↩
-
Dix, Paul. “The Open Source Business Model is Under Siege.” (2017). Link ↩
-
The delayed incentives for a commercial open source organisation may not be limited to complementary services. The entire premise of the modern commercial open source industry is based on the type & form of delayed incentives an open source project can accrue. ↩
-
An argument against open sourcing products with direct impact on revenue is minimal product differentiation resulting in limited pricing power and corresponding lack of revenue. Link ↩
-
A prominent recent example of this phenomenon is Hashicorp Inc. ↩
-
A consumer of a cloud-vendor has every reason to use the cloud-vendor’s offering of a service (built on top of a open source product), rather than using it from a different organisation, even if that other organisation is the original initiator of the underlying open source product (and therefore, arguably, has more of an expertise in it). ↩
-
Examples include Redis, Gitlab, Elasticsearch among others. ↩