Organizing and Funding Software Production

One of the best characterizations of the difference between proprietary and open-source software (OSS) is Eric Raymond’s The Cathedral and the Bazaar¹. This contrasts the top-down coordination of a large centrally planned structure (the proprietary Cathedral) and an open bazaar where people interact freely bringing their contributions and needs (bugs and feature requests) together in a bottom-up process that tends to produce good software.

Proprietary software is characterized by relations of control. The proprietary software production company aims to limit those who can use their software by imposing certain conditions, such as paying for it (e.g. Microsoft products) and/or storing one’s data through a service they provide while granting them permission to use/sell it (e.g. Google/Facebook’s products).

When software is not Free Libre Open Source (FLOSS), there is typically some entity which generates revenue from controlling access to that software. Where that revenue is generated by selling licenses to use the software, the business model is quite recognizable as it amounts to selling units of a product. As such, the organization controlling that copyright also has a recognizable form - with departments for marketing and promoting the product, new product development, and protecting against the infringement of the copyright (through legal and/or technical means) which allows the business model to be sustained.

Technical means of enforcing copyright (e.g. Digital Rights Management) always weakens the software product itself. In the best cases, the user experience is compromised in some minor way (like reporting your actions back to a server), in the worst cases, the product becomes unusable in certain scenarios (such as a lack of internet access to periodically check in with designated servers, or those servers being taken offline).

A recently popularized alternative to charging a licensing fee to use the software has been to offer a service based on the software. Users do not download and install (all of) the software themselves, they use it by connecting to servers that are running the software. These servers are owned or leased by the company, which charges a subscription fee to access the service, and/or captures useful data about the users which generate value for the organization or can be sold to other organizations for this purpose. Services that benefit from network effects (i.e. the product benefits from more users, e.g. Facebook, Uber) have been growing quickly, often following a strategy of subsidizing the service until it can gain a dominant position in its market, before beginning to extract the value from those users and their data.

The point of this simplistic overview is to establish that the question of how software development and digital infrastructure is funded has become an important one for human society. Software is penetrating many walks of life, and the question of how its development and maintenance is funded is shaping the offerings which make it to the market.

In the proprietary software/data domain, the organizational forms are quite recognizable from the industrial era. In place of functions like obtaining and distributing physical raw materials, there are functions like protecting intellectual property. Whenever more revenue can be generated from more customers it makes sense to invest in marketing the product with a view to increasing its usage. With strong revenues coming in (or the prospect of strong revenues to come) it makes sense to hire staff (or contract out the work) to further develop the software or to develop new software products.

The question of whether to hire staff or contract with external entities to get work done was considered by Ronald Coaese in The Nature of the Firm ². In this article, Coaese considered why organizations form and hire employees when the production they engage in could be contracted out. An efficient market would allow for the organization to exist as a nexus of contracts, but in practice the transaction costs are prohibitive.

There are transaction costs associated with contracting out work, such as finding a reliable supplier and negotiating a fair price, then ensuring enforcement of the contract. At smaller scales, it tends to be more efficient to hire workers and organize their production within a firm, than to rely on the market to serve every need with individual transactions.

In place of transaction costs, the hiring of employees incurs the need to organize them, and associated costs. Coaese argued that the cost of organizing a workforce internally rises disproportionately with the number of transactions being organized - placing an upper limit on the size of firms beyond which it would be more efficient to revert to contractual price-based interactions.

The internet and ICT developments have significantly reduced many of the costs associated with organizing a workforce, and this has allowed organizations to reach new scales in terms of their geographic reach, complexity, and global significance. There are however reasons to believe that over time bloat, inefficiency and misallocation of resources become more prevalent within these large organizations, as suggested by Coase’s work.

One of the things that’s new about CBPP is that participants have a different motivational profile. They are engaging in the production effort for different reasons, with the desire to make and share things displacing the profit motive and need to sell labor as fundamental drivers.

References