It’s been quite a while since I wrote anything on this blog, since mostly I’ve been ranting over on Twitter.

Given the current discourse, let’s look at single-vendor Free, Libre, Open Source projects for a moment. Since many others have also already voiced and written down their thoughts, my main goal here in voicing my own anecdotal perspective is not to point out something fundamentally new, but to highlight specific aspects.

Why Single-Vendor approaches?

Why do people attempt single-vendor (or vendor-owned, depending on how charitable you feel) Open Source projects? Often in combination with Contributor License Agreements (CLA) which allow them to further multi-license the code base and enable models such as Open Core or even fully proprietary versions?

Now, first I must acknowledge that there are other reasons - the “original” Free, Libre, Open Source software licenses such as the GPL don’t really protect the intended freedoms well in the context of hosted, SaaS offerings. The Tragedy of the Commons is, alas, real. As are ethical and moral considerations on who uses one’s software to what end. I do believe there are valid concerns that need to be addressed, but I will not cover them here.

Their driving goal is to capture the value as part of their company, their business, and their investor/IPO/M&A valuation. Free, Libre, Open Source is “just” a means of attracting attention, getting visibility more quickly, a hiring benefit, the occasional drive-by-contribution from an outsider.

And that works. They attract users, some contributors (excellent for recruiting), possibly can demonstrate some conversion to paying customers/partners, the potential for the project to be adopted more widely becomes visible. Excellent! This is what we all like to see.

Catch-22

The problem is that this must necessarily “fail” once the project succeeds.

I am broadly defining success as reasonably wide-spread adoption and utilization. At this point, multiple parties have a vested interest in the on-going success of the project, since they’ve build their own value on top or around - that might be a SaaS offering, consulting, training, using it in business contexts, or even just personal use.

The single-vendor needs a sufficiently large portion of these to pay them to sustain their own business; be that for support, maintenance, additional features, consulting, training, certification. And thus, they’re in direct competition with a number of their own supposedly-cherished community members! Uh-oh. And there is always a business incentive for competition to lower prices (or, in a fictional world, for higher quality).

Maintaining the competitive advantage

It’s either impossible or very costly (which in profit-maximizing capitalism are interchangeable terms) to maintain said massive competitive edge while on equal footing with the other players.

It does not matter whether this edge is technical differentiation (somebody else can build whatever add-on your open core withheld), or the experienced staff (you can never employ every single expert, and Open Source means others can become experts, and users necessarily build up expertise over time).

Hence, at some point, they will necessarily have to consider imbalancing the field. Modifying the license, trademarks, patent enforcement, withholding improvements, restricting use, increasing the barriers to collaboration, not accepting contributions so they can maintain control - we know the playbook.

The inevitable fork in the road

Alas, this is a slippery slope - there will come a point at which this creates a sufficient incentive for others to take the last commit before the vendor applied the disruptive license change and fork the project.

(Another outcome is that folks switch to or even start an entirely different project more to their liking, and the original mostly withers. Not typically desirable either, since it reduces the potential customer base, but hey, those remaining are obviously captive.)

Because for them - that project that’s the whole core of the single-vendor’s business? It’s just infrastructure. Their value lies elsewhere.

They would (and will) benefit from amortizing the cost of infrastructure maintenance, development, support across the community rather than paying someone else’s above-market-average profit margins.

Value-driven consequences

For the sake of this argument, I will ignore those community members that will walk away from the project because it now violates the spirit and values of Free, Libre, and Open Source projects. Said single-vendor would be negligent if they were dismissive of the impact this can and will have, however. (Still, they almost always do, because it’s not easy to relate to values that cannot be expressed in annual returns.) Feeling wronged, ethics, and, yes, spite are incredible motivators to create a more successful alternative. This will contribute significantly to the consequences - it will make it easier for said others to recruit and/or provide them with an alternative to switch to.

Decision time

A single-vendor project can succeed as a single-vendor product up to that breaking point. (Unless they fail the execution, or the project technically falls behind, but I am assuming a moderately successful project here.)

If their cost is mostly limited to paying the developers, technical writers, support engineers, and reasonable operational costs, this is a valid business proposition - the whole point of division of labor is, after all, to benefit from specialization. One can pay even rather fair and comfortable salaries. And since economy of scale comes into play as overhead is amortized across many, there is some growth potential in this.

Such models exist - as both dedicated organizations, or as departments handling Free, Libre, Open Source within larger businesses.

Why is this success a failure?

So, failure according to which metric then, you may ask?

Such a model cannot deliver, much less sustain, the above-market-average profit margins and exponential growth required by investors.

It may appear to for a time, but eventually, the market catches up and the regression to the mean starts. The commoditization is built into the DNA of Free, Libre, Open Source.

One trick is to exit before the investors catch on - or be bought by one of those companies for whom you’re just infrastructure and have your business model transformed. If that is too expensive or inconvenient, they will just - either individually or by coming together - swat your business model like an annoying fly.

Your customers and users together are always bigger than you are. That is the Power of Many. And to a global CSP, every single-vendor project company is unnoticeable roadkill.

TL;DR

There are fundamental conflicts between Free, Libre, Open Source models’s and (hyper-)growth-oriented capitalism’s definition of success and incentives.

The question is not if this will break, but only when.

Quo vadis?

It is my personal belief - nay, understanding - that we must find and utilize ways of sustainably creating and supporting Free, Libre, and Open Source projects, especially in the infrastructure space, without resorting to profit-maximizing venture capitalism and investors as the only option.

Supporting Free, Libre, Open Source commercially is feasible - but the model all but guarantees that such business models revert to sustainable approaches. (This may be a hopeful exaggeration.) Or fail painfully.

However, I am a proponent of taking this into account from the beginning, without going through all the sadness of massive commercial failure and enriching only a few while harming many.

Yes, we need incubators for testing out new ideas and innovations, with adequate incentives and rewards. It cannot be good for society if, as we currently do, much of this happens with somewhat stretched truths and by pretending unsustainable business models in fact are.

As investors begin to see and understand this, I further believe that single-vendor projects will receive lower - that is, more appropriate valuation and reduce the pressure to create even more of them. Those in search of billion-dollar pay-offs will go seek the next tulip mania. So if you’re currently caught up in this, brace for impact, and make sure you realistically assess what you can achieve for and from your current and future investors.

Relevant Free, Libre, and Open Source projects need to be owned collectively and in a collaborative fashion. They are public goods; this is especially true for infrastructure components. Organizational structures exist and are gaining traction - for-purpose organizations, associations, cooperatives, foundations, private-public partnerships, and even fully public-sector undertakings.

This will only be a small contribution to evolving our economy towards more sustainability.

As a society, and also as a software community, we need to reward and value collaboration, not divisiveness.

Let’s discuss!