12 Apr 2017

Monetizing Open Source

The a16z Podcast for 10 April 2017 was on Monetizing Open Source (Or, All Enterprise Software). It was a conversation with James Watters, Martin Casado, and Sonal Chokshi, asking the question of: how do software companies build a viable business model on top of open source?

These are my notes, with my comments in italic.

Example of a large successful business built on open source: Pivotal is USD270m in license (software sales) not support (professional services, which doesn’t give much margin).

Key elements:

  • Know how to cater to enterprises, rather than users (developers). You do need to get users excited. But commercial and community strategy are not the same thing.

  • Get a major trend that affects enterprises. E.g., ORACLE caught the relational market, which changes how enterprises can build applications. Pivotal caught the web-services wave.

  • Commoditization as a strategy (e.g., jboss, mysql, android, linux) has not historically been a big business model. It is a rare event. You need something new. Lead a new trend, and a big trend leads to a big business.

  • Cloud is affecting open source: Pivotal don’t sell you a tarball and support contract, they do continuous delivery of their software and cloud API. This gives you leverage beyond support. I don’t really understand this part, but Noel commnented: “E.g. if you expose an API they can integrate with your system, and once integrated it’s harder to leave..”

  • Bringing polish to open source (client facing vs developer facing). Pivotal made the UI closed source. Gives differentiation, which is important for “packaging for procurement”. Procurement is trying to compare things for pricing, and they are very good at that for labour cost. Having a small piece of differentiation is important for navigating the politics of procurement. When you get to procurement someone wants you to win (your champion), but you’re in a political domain.

  • Avoid “dog pile”. E.g., MongoDB were a sole supplier so controlled pricing to some degree because there wasn’t another MongoDB procurement can go to. Unique market position is important, because a larger company can quote for something that sounds similar and will kill a deal, even if they have no real capability.

  • Up-selling support only is hard. They have your product running, they only want professional services by phone. Better to sell a vision, train, and work with customer. You want a full-packaged differentiated approach, a unique offering, not a support company for a technology.

  • Why do companies buy licenses for open source? Because the software company, if successful, is an extended part of the enterprise team. They want you to be part of their strategy because of the expertise you have gained (hard won). You want to get to USD400k+ to have a good dedicate relationship with the account, so you can put good people with them, to grow the relationship. This allows you to be a better advisor, which the enterprise will value for the “talent on the ground”. Communicate the CEO/CTO not director or technical levels for “strategic account control” which you only get via deeper engagement. Enterprise software is expensive because of the trust model that you will give outcomes, not just support, for the enterprise.

  • Why now? Cloud/microservices is giving companies the chance to be the “leader of a thought”. Open source, and buying open source, has become a “normative behaviour”. Open Source is a requirement or close to, to avoid lock-in. They want the right to use the software, without ever-rising prices.

  • Advice for new businesses: Trade off between existing trend that has been validated, low risk vs. selling vision of change over the next few years.

Probably worth re-reading The New Kingmakers.