Building a fundable open source company

Editor's Notes

• #5: There are lots of different ways to directly make money from open source, and a few indirect ones as well.  You could make a compelling argument that developing and contributing to open source projects, if done well, is likely to increase your changes of getting hired and earning a higher salary.  The direct ways to make money from open source involve starting a company and figuring out how to get a subset of your users to pay you Purposefully want to stay out of the philosophy of open source and focus on how to build a company that sells a product or service to generate revenues
• #6: Bootstrap - get to CFBE quickly Not mutually exclusive. Atlassian is a great example. Founded in 2002 and the founders bootstrapped the company with $10,000 on a credit card.  In 2010, they raised $60M from Accel and 5 years later in 2015 went public at a $5.9B valuation Each has pros and cons. Bootstrapped companies grow slower because it typically takes capital to invest into growth
• #7: Using over twenty-five years of data from the Census Bureau, Puri and Zarutskie (2011) find that relative to the non-VC control group, VC-backed companies are more likely to go public, more likely to be acquired and less likely to fail. This basic pattern has generally been found in various subsamples over time.
• #9: So the first step in getting your company funded is to understand how you will make money.  Let’s start by looking at the biggest successes in the enterprise of open source software.  I would submit that there have been three projects that have fundamentally changed the way software is built and delivered and introduced open source to the world.  These are Linux, Apache, and mySQL
• #10: Apache remains a community project under the Apache Software Foundation, though its dominance is being threatened by nginx, which does have a commercial entity behind it and we’ll talk about how it makes money later on
• #11: Red Hat was created to provide enterprise support for Linux in 1993 and went public in 1999.  Today, it has revenues $2B per year and it worth north of $13B.  It’s by far the biggest company created from an open source project.  
• #12: MySQL also had a large financial outcome.  It was founded in 1995, sold to Sun in 2008 for $1B to headlines line this one and was later picked up by Oracle in $7.4B when it acquired Sun 2010
• #13: In the case of Linux and Red Hat, many CIOs needed that assurance of a viable business to be on the other end of support calls before they were willing to put their mission-critical applications on a new operating system.  They needed someone to say “this is going to work and you can call us to fix it if it doesn’t”. Then there was indemnification. Lot of problems with this business model today (1) software matures, linux example (2) more & more becoming software companies
• #14: Every company is turning into a software company as software extends beyond the online and becomes pervasive in every part of our physical world.  We can already see this happening with companies like Nike, Capital One and JPMC, and GE building large in-house software development organizations in what were historically non-software companies.  The result is that there’s a higher ability to consume open source software without the need for an insurance blanket from another entity to provide support and services
• #15: 3rd reason, is that it’s difficult for this business model to capture enough value to sustain ongoing innovation
• #16: So if the support and service model won’t work, what comes next? What are viable business models for open source today?
• #17: One reason that open source has been so quick to takeoff is that every project that gains critical mass develops a core asset that is its community.  It’s the strength of the community that tends to determine the eventual success of the project.  The challenge in attempting to monetize open source is that you must thread the needle of providing enough value to the community for it to continue to thrive while attempting to extract some of that value through pricing.  The field is littered with failed attempts at doing this.  A number of business models have emerged since the original support and service model. 
• #18: The first and by far most popular one is the open core model.  The open core model involves developing a substantial portion of the project as free and open-source software and building a community around that.  Other features are then built as proprietary software that typically cater to more of the enterprise use cases.   Carve out functionality that those with higher willingness to pay will want. Examples LDAP integration and more broadly integration into common enterprise systems, security, management, provisioning, monitoring, availability
• #19: Here are a few examples of open core projects.  Hadoop as the open-source project and Cloudera as the commercial entity.  There are different dimensions for segregating free vs commercial offerings and these include security, management, monitoring.  Offering an on-premises version of the software is another way of  charging in an open core model. Pause on Docker. Good example of the challenges and tradeoff of the open core model.
• #20: Old – monoliths New – microservers, increased willingness to consume external services result is developer services trend
• #21: The SaaS model is becoming more and more popular with micro services.  Outsourcing the undifferentiated heavy lifting is a common theme these days, made popular by Amazon. Twilio was the first service I can remember that developers wanted to outsource. Who wants to build communications infrastructure? Since teams were already used to relying on other internal teams for abstracting away services, why not outsource that function when it not part of your core IP.
• #23: Related but different usage of open source… Reason for going the open source route is...don’t represent differentiated IP. Instead undifferentiated infrastructure At Netflix...Cassandra 0.7, AWS in 2009/2010 Necessities, strong desire to put out in the community to (1) reduce chance we end up alone, railroad to town (2) reduce ongoing tax and maintenance Other benefits – code quality Wouldn’t call these open source companies...product they sell isn’t OSS. OSS here is a component rather than product itself
• #25: Market: re there clearly identifyable users and buyer for the product, and do they have a willingness to pay. Is the market sufficiently big (# buyers x average selling price) Comp: what is unique about what you are doing and why hasn’t it been done before? What alternatives exists to solve the same problem for your buyers? Will there be any barriers to prevent others from copying you?