In this post, we'll dive into the enormously distinct differences between creating and capturing OSS value.
It’s fairly obvious that I’m a huge fan of M.C. Escher’s math-inspired drawings and art in general. In this surprisingly not-so-famous lithograph, we see order amidst chaos. This juxtaposition is a perfect analogy for what we’ll be digging into here in Part II of “Creating OSS Value != Capturing OSS Value”:
- Creating OSS value is fairly straightforward to understand and in the “order” category: Simply design, develop, build and open source a software project for the world and if it solves a real enough problem/pain point for enough users, anyone on the internet can frictionlessly download it, use it for free and derive unbounded value. Idealistic, maybe, but this is the general idea.
- Creating value is a function of a) how effectively a given problem is solved, b) how much a given workflow/process is improved along various dimensions, c) how powerfully the project unlocks never-before-possible capabilities or d) all of a, b & c simultaneously or any combo++.
- Value creation allows humans to do things they could never do before — harkening back Steve Job’s famous quote: “Computers are like a bicycle for our minds”. OSS is similarly powerful: It is an amplifying, democratizing technology lever that allows us to magnify our abilities in novel ways.
- One major consideration when releasing a new OSS project is to understand and be intentional about whether it falls into one of two broad categories:
- Are you A) Creating a new kind of software category or B) Democratizing previously expensive, unattainable-to-the-masses tech?
- The implications of these two paths are quite significant and I’ve written previously about this at a high level in response to an a16z podcast between Martin Casado (a16z GP and founder/CTO of Nicira — creators of the wildly successful Open vSwitch project) and James Watters (Pivotal SVP Strategy, the Cloud Foundry company).
- **New category creation: **Creating an entire new category often means either combining previously distinct layers in the stack (bundling) or disaggregating vertically integrated products/services (unbundling). This follows Jim Barksdale’s ethos, former Netscape CEO.
- Examples of “new” category creation: Nicira (major OSS creators of Open vSwitch and associated products) driving the SDN space, GitHub driving Git with a new kind of social network for code, Docker driving the “containerization” movement/category. Examples here can certainly span both categories (new cat. + commoditization/democ.).
- Commoditization and/or Democratization: This is the far more common path that many OSS projects take. It essentially boils down to disrupting incumbents. Talend did this to Informatica/Ab Initio, etc. Cloudera/HortonWorks/MapR did this to legacy data warehouse vendors and proprietary data storage systems like Teradata (arguably they also created a new kind of category), MuleSoft did this to TIBCO, etc.
- Value can be created directly and indirectly:
- Direct value creation: There are many “direct” recipients to the value that OSS provides:
- Developers have immediate access to libraries/frameworks/tools without needing to go through procurement processes.
- End user companies can access new technology without signing a contract with a vendor.
- Indirect value creation: Indirect recipients to the value OSS provides also abound:
- Product builders can accelerate their roadmaps by picking up OSS code and integrating/applying it without re-inventing the wheel, testing cycles and R&D contributions can be “crowdsourced”, feedback cycles immediately “access” the whole of the internet vs. a small employed group of developers in a single company, etc.
- Governments and organizations of all kinds also benefit from many of the same things as product builders.
- Capturing OSS value is a completely different matter. This is definitely in the “chaotic” stone/bread (we’ll never know) circular background of the Escher lithograph. 😄
- I have a theory that it may be possible to model a mathematical constant that represents the amount of value OSS creates relative to the value it is possible to capture. In conversations with large commercial OSS company/project founders, it seems clear that if there are 10,000 users of a given OSS project, the potential to “convert” those users to paying customers of some kind inside a company is usually capped at 3–5% (on the high end) or .01–1% (on the low end). This model is highly prone to entropy given how challenging it is to even know how many users of an OSS project there are to begin with!
- If we consider a corollary example, fully proprietary software likely captures 10X+ (conservatively) of the value that it creates. I’ll write more on this in future posts.
- Many questions arise here that must be contextually assessed on a case by case basis:
- Do you charge for support?
- What about training services?
- When and how does it make sense to offer those things?
- Can a product be built around the OSS project to commercialize it somehow with a closed/proprietary license? (open core)
- What about donations/Patreon? Is that scalable? A tip-hat approach isn’t exactly a fair way of aligning the value an OSS project creates relative to the project creators inherent contributions.
At the recent OSS leadership summit earlier this year, I presented a version ($25B+ in value has been added since then) of COSSI and various thoughts here on creating and capturing value with OSS. There are many blogs to be written from this highly dense deck.
- Capturing value with OSS also happens directly and indirectly.
- Approaches for directly capturing value:
- We’ve previously investigated the many business models that work quite well. These are all valid ways to “directly” capture OSS value. By directly, I mean that the OSS project is the predominant emphasis underlying the value proposition, UX and product building block of the company/entity selling IP to customers.
- Approaches for indirectly capturing value:
- If COSSI shows that the top open source companies have directly captured $100B+ in value (if I were to loosen my $100M revenue criteria slightly, easily excluding Red Hat), then companies indirectly capturing value span many hundred-billion / trillion-dollar categories. The opportunity is simply much, much larger here.
- Categories: SaaS ($300B+), Cloud Computing Platforms ($500B-$1T+), Banking/Finance ($10T+)…+ more.
- In the indirect category, here are two interesting examples:
- Cloud providers: i.e. AWS EC2. Without the OSS Xen hypervisor, would it have been possible to build and launch EC2? Sure. Would the cloud computing category have materialized as fast? Maybe. Maybe not! No one knows, but what is true: Xen powers EC2 and is an OSS project. EC2 has “indirectly” captured an enormous amount of value (many tens of billions, far above Red Hat’s valuation alone).
- SaaS: All these vendors heavily use OSS to build their proprietary products. While this $300B+ industry is growing, here’s a prediction — I believe commercial OSS (direct value capture dynamics) will exceed SaaS by 2X within the next 15 years.
- What kind of properties are involved in building products that use OSS but do not expose it or emphasize it as part of the value proposition? This is a topic for a future post, to be sure.
This post just scratches the surface of many of these deeply nuanced and complex ideas and concepts. However, hopefully this serves as as starting point basis for thinking about how OSS value can be best delineated. 🙏