OSS Will Kill Your Startup
Do I have your attention? Good.
I’m gonna blow up everything you think you know about OSS and your wrong-headed decision to base your startup on your own OSS-licensed software and a freemium model.
First, some background.
I invented hosted source control in 1999 three months before the next two services emerged. Started a company, raised capital, had employees - and a 401(k) that we fully matched - before any of the current providers in this market were founded.
I employed one of the first-ever freemium models - the name itself was Freepository.
That company failed (for the details, buy my book 145 MPH) but I kept the service running, moving the data center into my house. It grew to host more than 3 billion lines of code for over 400 thousand members across the globe. NASA, Microsoft, HP and thousands of other companies had code on freepository. While running the service, I also did infrastructure work for Wily, Walmart.com, StubHub, Oracle, Juniper and many others.
I know this space very well.
Some assumptions :
- startup refers to a company providing services implemented by founder-written software
- OSS refers to software licensed such that others can use it without paying fees
- as founder, you have limited (or even zero) operating capital when you start
- IANAL, though I am criminally handsome
Starting a company begins with a single action:
Deciding what to do and how to make money doing it.
In the case of a tech startup, the flexibility you have to explore new ideas without spending a lot of money is almost limitless. Assuming you have a laptop and an internet connection, you can download - for free - virtually all the tools you’ll need to design whatever it is you can imagine.
The irony of using free tools isn’t lost on me.
Once you have something ready to test, you can - again, free of charge with some conditions - provision a new virtual instance of a fully functioning server (or set of servers) and then run your new service.
You’re now a real business. Well, minus the paying customers. And without effective marketing. And perhaps most importantly, without any legal counsel.
If you’re like a lot of founders (I could make up a percentage here, but why bother… it’s a lot and you know it), you’ve already made the mistake that has the very real potential to kill your startup.
You chose to release your software using one of the permissive OSS licenses. You’d read that this was the way to ensure lots of people use your service, gaining popularity so that you can later monetize… blah, blah, blah.
What you’ve really done is this:
you’ve dropped a thing into the wild under terms that lets anyone use & dissect that thing free of charge. If the thing becomes popular, you’ve provided - free of charge - a shit-tonne of market research & validation to every one of your potential competitors.
If you’ve used one of the most popular of the OSS licenses, the MIT License, then your competitors can use your own software to provide a competing service. Any customers they obtain are taking revenue from you. And if they make improvements or bugfixes to it, they are under no obligation to release those back to you or anyone.
That freemium model that looked so good on Day 1 is now going to drive revenue to your competitor - who probably controls a very large platform - while quickly starving your startup of any oxygen.
Let that sit with you for a moment.
You back? Ok, let’s continue.
This isn’t hyperbole. It happens every day and billions in revenue is captured by companies that use software written by author-founders who provide the original service(s), now in competition with firms that are simply harvesting - or strip-mining - the profits.
Some of you will by now be practically screaming “but [company x] doesn’t do that! They give back to the community…” yada yada yada.
You can talk your book all day long. I understand that you want to believe that you & your company are on the right side of this. This may be legally permissible, but it is sketchy as hell.
Once a company has released their work under a permissive OSS license, there’s essentially no way to prevent competitors from using that original work to siphon off revenue as the original service becomes popular.
There are only two dimensions along which you can prevent your fantastic new service, which is becoming popular, from being directly copied:
- Don’t release any of the code (duh)
- Trademark the name
If you trademark the name Flubber to describe the flubberistic service you created, when Flubber becomes popular another company cannot begin offering their own Flubber without an actionable infringement. I trademarked freepository in 2004 for this reason. That’s enough to keep the big players from lifting your brand name.
They will, at a minimum, have to rewrite & reimplement part of your stuff and then offer it under a similar name like Open Flubber, implying that your (closed) Flubber is somehow deficient.
This is happening today at AWS. It’s been written about publicly at the New York Times & Medium and has been covered by several industry analysts, including myself.
So what should a startup do?
I am not your lawyer. This is not legal advice.
Get a good corporate attorney who understands software license & trademark law, then follow her advice.
In no particular order:
Don’t release any of your software under any flavor of OSS license. Ever. This is a forever decision that is impossible to undo.
Don’t place your work in a public repo on github. This is de-facto releasing it with no license. It will be copied.
If you decide you must release parts of your work under some form of OSS license, ensure that you protect your project name with trademark. This isn’t as simple as writing “Trademark 2020 Flubber Corp” in the header. There’s a process to make it legal. Ask your attorney.
Monitor and enforce your trademark. Lack of enforcement is part of a legal concept called abandonment. If you want the protection of the trademark, you must enforce it.
Releasing your software under an OSS license while hoping that a freemium model works for you is no substitute for marketing.
Don’t make a tactical licensing decision without fully understanding the long-term strategic impact it will have on the very existence of your company.- jbminn