Avi Press, Founder & CEO, Scarf
State of Open: The UK in 2022
Phase One: “The Open Source Journey”
Quantifying the value of open source software is vital yet challenging, given the absence of a conventional market. Traditional approaches are hindered by incomplete data and biases, requiring reliance on models and estimation. The value of open source software becomes tangible in markets where various goods and services are built upon it, spanning sectors like cars, healthcare, and productivity software. Utilising data from platforms like GitHub and Scarf, particularly through distribution analytics, helps identify users, including internal and operational applications not captured by public code analysis. Although assigning value to individual components within complex software ecosystems remains complex, improving visibility into global open source usage provides optimism for more accurate estimations. As understanding of open source utilisation advances, the quest to quantify its value becomes a progressively attainable goal.
How Much Value Does Open Source Provide, Exactly? – Thought Leadership
Avi Press, Founder and CEO, Scarf
Quantifying the value of any public good is a classically difficult problem. Attempts to model and estimate the value of any given public good are prone to various difficulties, from woefully incomplete information to psychological biases of buying versus selling.21 For more traditional goods, a better approach is typically to watch what the market does – if a good or service has an efficient market with well-aligned incentives, the good should be priced fairly, and your analysis is straightforward. Open-source software does not have one of those markets, but understanding the value supplied by digital public goods and in particular Open Source Software is of utmost importance for its continued success.
Without such a market in place, we must turn to models and estimation. Unfortunately, we collectively have inadequate information to even compute the majority of the value that Open Source Software delivers. If we had an omniscient catalogue of every piece of (Open Source and proprietary) software ever written, with every individual instance where that piece of software was used, the value created by that individual usage, and the percentage of that value that came from leveraging open-source software dependencies, then it would just be arithmetic!
A property of Open Source Software that can be helpful here is that its value is primarily realised and captured in other markets. From cars to healthcare to productivity software, goods and services in virtually every market are being built on top of Open Source Software. This implies that if we have data on the goods in other markets, and how the companies on the supply side of those markets use Open Source Software, we have another reasonable proxy to estimate the value we’re after.
This data is actually attainable! Static code analysis on platforms like GitHub and package distribution analytics from platforms like Scarf have given Open Source Software projects a clearer understanding of which companies, organisations, and even governments are using their work. Distribution analytics can play a particularly important role here, as a large portion of open-source usage is for proprietary, internal, or operational purposes, which are missed by public code analysis.
Of course, even a perfect understanding of which organisations use which pieces of Open Source Software is not enough on its own. When tens of thousands of software components combine in complex ways to power an organisation’s operations and products, assigning value to an individual piece or the entire set is a challenge on its own. However, it greatly reduces the scope of the problem and results in a more tractable system to model.
Estimating the value provided by Open Source Software remains an open and difficult task. However, we have a good reason to be optimistic in our efforts to reasonably do so – as we continue to improve our visibility into how Open Source is being used around the globe, we get closer to a clear answer.