Thoughts on Open Source

January 17th, 2019

This post was prompted by Ben Thompson’s discussion on the latest open source angst. His post is not the first, nor is it the more comprehensive, but it best captures the heart of the issue for me.

Ben offers a comparison of open source’s woes with the music industry’s woes in the early 2000s. The music industry wasn’t selling music: they were selling convenience via cassettes, CDs, etc. When the internet came along and dropped the cost of distribution to zero, that convenience vanished and along with it, that revenue.

The same has effectively happened to software. Just like the music industry, we used to sell CDs of software like Office or Netscape for hundreds of dollars, but webapps and the browser have effectively decimated those. We now spend $100 to acquire a customer that generates $9/month (or even less with ads). Minus the breakout hits, consumer software is more a distribution and cash flow game.

For a long while, B2B software still had advantageous margins. Marketing was expensive and sales cycles last years. So companies started using open source to reap marketing returns from your R&D investment. Building (or co-opting) a community around the software was much cheaper than traditional marketing. Once customers were big enough, you could sell them commerical licenses or services at low prices and still have great margins because acquisition was so cheap. In particular, having a cloud-hosted service offering was popular because unlike support or ad-ons, it let the product team focus on a single offering, was operationally simple, and was recurring. MongoDB, Sentry, Citus, Confluent, the list goes on.

IaaS has begun to challenge this approach. Whereas before, the Internet had dropped the cost of distributing software, IaaS has dropped the cost of distributing services. Commodity services, like operating systems and databases, are particularly vulnerable to this.

For a comparison, M10 instance with Mongo’s cloud offering Atlas is $0.77/hour or $554.40/month. For a comparable DocumentDB 15.4GB memory instance is $0.277/hour, plus $8.00/month for 80 GB, for a total cost of $279.44/month. Amazon is here is clearly undercutting Mongo and while I doubt many MongoDB’s current customers are going to cut and run, AWS has reduced total potential value of the market and thereby the potential value of MongoDB the company.

Open source remains great at creating new markets. It can help spread an idea faster and more cheaply than proprietary tools. But now, it’s shifting to become more of a marketing channel than a business model. But once an idea or a market has been proven out by open source software, how are businesses expected to maintain the lovely profits and valuation that SaaS once provided?

I obviously don’t have any answers, but let’s look at how the music industry has adapted. Music performers now have vastly different business models. Album sales and streaming represent a fraction of revenue. Instead, performances (concerts or sponsorships) are the majority of music industry revenue. Music videos and celebrity gossip are the new means of marketing and capturing attention and musicians are better than ever at controlling these narratives (compare Taylor Swift to Michael Jackson). The music is less a profit center (though it’s still very profitable) and more of a marketing tool for creating icons and brands for people to buy into.

My guess is that OSS businesses will need to become much better at building brands and cultures around their software, so markets “accept no substitutes” towards IaaS clones. It shouldn’t be that hard: developers care about security, performance, stability, simplicity, design. Stripe, for instance, isn’t a better payment processor than its competitors, but their product is just nicer. And while their business will never be the majority of the market, but I doubt they care: owning the best 30% of the market is better than owning the other 70%. MongoDB, on the other hand, has a relatively shitty developer brand, having cheesed their way to the top. OSS is bound to keep growing in importance and no matter how the economics shift, there’s bound to be opportunity if you’re willing to adapt.