Open Source - the Darwinian approach to software

C/NET has a story this morning "Open Source, Open Wallet" on VCs investing in Open Source projects. David Skok of Matrix Partners, an investor in JBoss, warns "Too many of these companies (now forming) are being funded without a community," said David Skok, a venture capitalist at Matrix Partners. "If a community doesn't form and form fast, then they're going to burn through their venture capital, and they're going to be disasters."

Open Source projects must pass the Darwinian test (only the fittest (most interesting) survive) to attract developers and contributors to the project. Maybe 1% of all Open Source projects attract more than 30 developer/contributors. Then the user community applies its own Darwinian test to that 1% of projects, as only the most useful and supported projects attracts a significant user community. The Open Source projects you read about are the ones that passed the Darwinian tests. Yet, the media would have you believe that Open Source is a sweeping movement.

There are VERY few VC investment opportunities in Open Source. There will only be one or two winners in each space, and the larger spaces are already covered. Linux has already been covered by Red Hat, IBM, HP, Novell, and others. Apache is also supported by many companies including Covalent and Microsoft. The database space is already covered by MySQL, Postgres, Ingres, SleepyCat, BerkleyDB, and 4dbObjects. CNET has another story today on Ingres's entry into Open Source entitled "Spinoff player to compete in open source software"

The business model for Open Source is principally the support revenue stream, which even in the traditional software business, is very significant once the installed base grows. Some Open Source companies also use a hybrid licensing model. They give away a limited use free version, and sell commercial versions or upgrades. In either case the critical success factor for Open Source business models is to quickly obtain a LARGE user base.

The investment logic for VCs and entrepreneurs is that because there is no development cost, the business can be built on a MUCH smaller investment and the support revenue stream can be substantial after several years. Further, the cost of providing support is very low. JBoss, for example, typically only gets one or two calls per year from each contract. Remember, their customers are usually not end users; they are typically sophisticated IT guys. They get most of their questions answered on discussion boards and web sites. End users and consumers are different. The high cost of support for desktop end users would blow the Open Source business model out of the water.

Where is Open Source being successful and why?

1. Lower cost and lower complexity are the main reasons companies consider Linux or Open Source. Don’t underestimate the lower complexity factor.
2. Nearly 100% of Linux/Open Source users are also Microsoft customers. Very few enterprises will use strictly Open Source.
3. Linux is replacing UNIX market share for the most part, not Microsoft. Sun, HP, and IBM proprietary systems are being replaced by Linux.
4. Linux is largely used for web servers, print servers, and file servers, or in very specific, stable, single purpose applications where there is no technical innovation happening. There are exceptions of course, but based on my experience this is where OS is being used.
5. Customers do not choose Linux or Open Source because it has more features, better performance, higher reliability, better support, or even lower cost of ownership over the long term. In most cases they don't. They choose it because it is less complex to use and cheaper to buy for very specific applications. These same people choose Microsoft when it comes to mission critical business applications.

Open Source software is here to stay. It is appealing to the "Over served" market where a "good enough" alternative at a lower price is just what they want. This is the classic disruptive inflection point described in Clayton Christensen's book "The Innovators Dilemma" which I reviewed here.

The software business is an extremely competitive, rapidly changing, market. The early innovators and market leaders are often eclipsed by new innovations, business models, and market approaches. It is a market where The Next Big Thing can be the old thing, done in a new way.

Open Source may be one of those old things (existing applications) done in a new way (the code is free, you pay for upgrades and support) that changes the market. Is Open Source a disruptive business model that will change the market? Early results are gaining lots of attention but the ultimate impact may not be known for years. Stay tuned.

Comments

[Patrick Rich]

Agree on just about every point, Don, except the motivation behind VC interest. That is, and always has been, a statistically-based return on investment over time formula. Open Source today is an attractive investment because:

1. The general stock-buying public and institutions don't understands what it means, and related business models are still in the experimental stage.
2. Investment costs per company will tend to be lower.
3. Time to market will tend to be less.
4. It has buzz value.

These all eventually translate into the proverbial bubble (bubble size TBD) when hoards of Open Source labeled VC portfolio companies and attendant hype hit the stock market, spike, and then decline into oblivion.

Keep up the good fight. Too bad your opinions will likely be discounted because of the association with Big Bad Microsoft. As a confirmed capitalist, I admire the achievements.

[Dan]

"The high cost of support for desktop end users would blow the Open Source business model out of the water."

I get support from, for instance, www.Mepis.com for 10 dollars up front purchase. I usually get most of my support for all applications directly from the OS website (and their package download managers) which I think is amazing.
You might be referring to the "Power Tool" users who require something more sophisticated that OpenSource might not currently support. The current logic is that OpenSource is not equipped to handle enterprise solutions yet (Example: Mozilla Thunderbird) so this might be a contributor to the stall at the moment on the desktop.
Then the Closed to OpenSource crossovers can get expensive etc. so to me it can become an issue of 'all or none.'
Currently, www.Blender.org is on the right track especially for cross compatibility in complicated high end graphics environments because of less contract issues getting in the way on that level. Less code and easier plugin support for instance (terrain generator) I think it's pretty decent already and some OpenSource graphics apps (engines) www.ogre3d.org are working on plugins for a world editor built in to Blender.
Fun stuff.

I've also heard stuff about desktops and the Internet and one or the other as better. I think that we'll always need an OS on a portable device because you still have to download.

[Susan Wu]

You speak of Apache as if Apache is only one product =). In fact, there are numerous Apache projects, over 30+ at the top level alone.

I posted a response to Martin's article as well, to the gist of:

Simply speaking, any discussion around a bubble in valuation revolves around any outsized expectations that an open source business model will deliver greater profitability and greater resource allocation efficiency than a traditional model. The ‘hype’ is related to how large investors believe these potential gains to be and how quickly these gains can be realized. Just as companies in the late ’90’s realized that they could immediately increase their valuation multiples by appending their name with a ‘.com,’ chances are the words ‘open source’ also signal some expectation around potential profitability and growth.

However, just because a company employs open source in their business model doesn’t immediately lead to the conclusion that this company will be able to generate abnormal earnings (a higher Return on Equity than their cost of equity capital.) Valuation is simply an exercise in discounting the future expected cash flows the firm is expected to generate. ‘Open source’ can affect valuation in that the expected cash flows may be of a different composition and the perceived riskiness of these cash flows may differ than that of a traditional software model.

full post here:
http://feather.planetapache.org/?p=21