HashiCorp’s announcement last week that it would shift away from traditional open source licenses to BSL is the latest nail in the coffin of traditional open source. It was absolutely the right business decision and one that mirrors the moves made earlier by Mongo, Elastic, CockroachDB, and others. These shifts are informative to any startup that relies on or is considering open source strategies. Here’s why:

1. Today, companies built on open source apply it for distribution advantage and NOT for development leverage

Many classic open source projects relied on vibrant communities of developers, and while this is still the case for great technologies that make a big difference in the world (see Rust), it is safe to say that successful businesses are going to be based on captive projects where one team—or even one person—controls the project and crafts the vast majority of its code.

If this is true, you may be asking, what’s the point of open source at all? Well, it comes down to initial distribution advantage: users, especially individual developers, want ease of adoption. And open source licenses deliver this free use and the potential for independent deployment and enhancement that many developers care about.

 

2. Successful open source projects will be copied 

Stretching back years, AWS and other major vendors have taken open source projects, offered them up as hosted products on their own platforms, and then hoovered more than half of sector economics. The history of AWS EMR vs. Cloudera and Hortonworks, AWS Elasticsearch vs. Elastic, etc. is stark.

The natural response of open source companies threatened by this kind of competition has been to shift strategies to restrict competitive use of source code. This came in the form of various licensing strategies that either delayed open licensing or restricted authorized use to prohibit hosted solutions. HashiCorp’s BSL is the most recent example, but the road was paved a long time ago by the first vendors who took similar defensive steps (again, Mongo, Elastic, etc.). This is where the business success deviates from true open source, as it means that a user can’t “do anything” with the source code, including in many cases build new products and services. In fact, these licenses are not recognized as “open source” by OSI.

 

3. An optimal open source strategy approximates a freemium one 

The first goal of any company is to get escape velocity in number of active users, and when venture capital is available, converting that product use to revenue can come later. Open source—especially traditional licensing like Apache 2.0—is a great way to do this. But longer term, it’s almost assured that successful adoption and revenue growth will demand the same kind of defensive move we just saw from HashiCorp. And this is when an open source-based revenue strategy becomes less distinguishable from a closed source, freemium one…

Again, hats off to the teams like HashiCorp that have figured this out while at the same time delivering incredible products. And to be explicit, none of this commentary is meant to question the power of open source or the reality that it can be a compelling model for early company growth. Rather, it is a call to action for us all to consider how to execute on optimal open source strategies, even at the earliest stages of growth.