The $70B Open Source Software Paradox


It’s no news that over the last 10-15 years, particularly noticeable in the last 5 years, the enterprise software industry has gradually transformed from complex, expensive, closed-source (proprietary) software stacks to extensible/configurable, low-cost, open source software (subscription).

This sea change in software innovation has manifested and been driven in a few interesting ways, of note:

  1. Rapid pace of innovation: Software developed in the open with a high-touch collaborative model means the pace of innovation is inordinately greater than proprietary models due to immediate and rapid feedback from users thanks to reduced friction in the traditional feedback loop (Closed-Source Software = push code to private repo > expensive, slow feedback from small group of users > improve > repeat … OSS = push code to public repo > free, immediate feedback from anyone globally > constantly improve > repeat)

  2. Free / highly-reduced TCO: Because OSS is free to use and innovation rates are dramatically higher, users have great incentives to at least try out the OSS alternatives to their proprietary tools. More often than not, they decide to move in an OSS direction because their license costs are eliminated by doing so. There are hidden costs with OSS to be sure, but even with the operational overhead and implementation curve, the TCO is much lower and the quality arguably higher = net positive win for the user/organization.

  3. Backing: a recent post by VC John Vrionis from Lightspeed Venture Partners (early investors in OSS successes Docker, Nicira, CloudBees, Datastax, MuleSoft and others) noted that over the past 5 years, the amount of money invested in commercial OSS businesses has jumped up by a factor of 10 as compared to the previous 5 years. This signals that many commercial OSS businesses are getting funded to grow into large emerging category leaders, which means more stable offerings and support services are being brought to enterprises, making OSS broadly viable to customers, replacing proprietary incumbents.

I could go on and on about the virtues of OSS. I am a huge believer. However, this post is entitled “The $70B+ OSS Paradox” for a reason…

Based on what I have observed broadly across a variety of perspectives, there seems to be a widely held view that OSS is not necessarily the best way to commercially introduce new and innovative technology in a rapidly evolving space (enterprise technology). As an example, many entrepreneurs have recently (during this industry transformation) preferred to build closed-source software companies since they believe holding onto proprietary IP, patents and other crown jewels will give them an unfair advantage and a degree of defensibility against their competition.

In my opinion, OSS has created more value for end users and organizations than all proprietary software ever written; the big problem (enter the paradox) is that OSS businesses conversely capture a tiny fraction of that value! On this note, I would encourage readers to check out the excellent business book “Zero to One” by Peter Thiel which elaborates on this concept of creating vs. capturing value and the many deep subtleties therein.

So… is building a commercial OSS business a good idea? I submit that not only is it a good idea, it is an excellent idea that will ultimately grow into being the standard way of building new enterprise software technology both for founders and consumers in the software industry at large. There are just too many positive, accretive network effects driving a long-term transition to the OSS model. It may take 10 or 20 years, but I believe transparency will win over opaqueness.

In a recent thought-provoking Twitter exchange sparked by Derek Collison on the difficulties of making money with OSS, Luke Kanies (founder and CEO of Puppet Labs) speculated that there couldn’t be more than 10 OSS businesses that have reached more than $100M in annual revenues. I found Luke’s comment to be my impetus for writing this post as many of these ideas have long been swirling in my head.

I am excited to be open sourcing the following sheet which was in some ways inspired by the BVP Cloud Index. This is an active attempt to track all of the commercial OSS startups ever founded that have reached or are now reaching $100M or more in annual revenue. Since this revenue number usually triggers IPOs and large exits, I feel it makes sense to build an index around these companies even though the data are private and somewhat fuzzy in many areas. To that end, I should say that this data is somewhat speculative – please, correct me where you think I am wrong!

Here is the “Commercial OSS Index”:

A few observations and predictions from this dataset:

O = observation, P = Prediction

O: OSS business models vary depending on how the project was created and which entity drove the evolution of the core project (see BDFL). However, open core is the predominant model the very successful OSS business have utilized.

P: There will soon (12-24 months) be general agreement around which OSS business models “work” and which ones don’t.

O: The mid-2000s saw a Cambrian explosion in OSS businesses. Most of the businesses have taken 10+ years to reach $100M in revenues.

P: Given the enterprise consumer inertia to adopting OSS in the mid-2000’s era, I predict that the ultimately successful batch of new OSS businesses founded in the last few years (since 2013) will take at least 30% less time to reach $100M. It is simply much easier to sell OSS to enterprises today than it was 5+ years ago, let alone 10.

O: The fact that the vast majority of these companies are still private is startling.

P: I would have predicted most would have exited/IPO’d by now. I predict that 30%+ of these companies will IPO or sell to an acquirer within the next 36 months.

The total value of the OSS businesses listed in my sheet = $50B+ minus Red Hat.


Now read this

Is open the new closed?

This post is nuanced and closely tailored to some of my thoughts on the increasingly popular question: “How to best monetize OSS in the enterprise?”. After listening to the most recent ~30min long a16z podcast, I wanted to take the... Continue →