There is no Open Source bubble

I've heard rumblings from some quarters about a return to the salad days of the '90's for the IT industry, this time with Open Source and Web 2.0 inflating the bubble. Setting aside the question of whether the dot-com bubble was something to which we'd actually want to return (it isn't), the possibility of a repeat is pure fantasy.

The rocket fuel at the core of the dot-com bubble had two main constituents: first, the availability of the IPO as an exit. A public offering made the VC a 10X return on his investment, made investment banks mountains of cash for nothing, and allowed institutional investors and fund managers to pad their return rates and bump up their bonuses and credibility.

Second, the fear of the telcos that they were becoming obsolete. In the early days of the Web the telco, the enterprise and Wall Street didn't take it or firms based on Web technologies seriously. The telcos in particular allowed ISPs to spring up using their phone lines, IP-based network equipment and Web-based businesses to develop, all without lifting a finger. Probably because that finger was bloated on all the free cash on which the telcos were gorging.

Then, seemingly overnight, the enterprise woke up to the fact that thousands of web servers were springing up internally, matching the grassroots growth of Web-based companies like Yahoo!, Lycos, and Amazon. The IT analysts like Gartner and Forrester smelled enterprise IT spending and started talking up the Web, VCs began funding an increasing number of companies with ".com" at the end of their names, and The Street started taking the whole phenomenon seriously.

The Street's interest was really the boost that got the bubble truly under way, because it was their efforts to promote the IPOs they were underwriting to their big institutional investors as a quid pro quo that got the telcos annoyed. Here were nothing little software companies like Netscape with market caps that rivaled the firms whose phone lines upon which their business model depended. The telcos struck back with the ridiculous "ASP" story around 1997, claiming they were going to sell enterprise applications as services to the biggest companies on the planet. As soon as they could figure out their billing systems. And build the infrastructure to host the applications. And figure out what an enterprise application was, and where to find one. But anyway! The important part was the "build the infrastructure" stuff, because it meant a massive surge in telco spending on IP-based equipment and Web technologies attracting the VCs and investment banks. Sure, there were detours down deadends like ATM and WAP, but it was still money. The quote that is seared into my mind from this period was from the CEO of a systems management company who, during a workshop on future scenario planning, when the opinion was expressed that the telcos would all fail at their ASP plans, said, "sure, but they're going to spend a lot of money failing." Which to me sounded like a business plan focused on polishing the marble floors of the ballroom in the Titanic.

The other crucial ingredient brought by The Street and their promotion of the building pipeline of IPOs was new money the equity markets, supplied by a new class of investor: Joe and Sally Lunchpail. The massive influx of capital into the equity markets triggered by the migration from pension funds to 401(k)'s and mutual funds was encouraged to play the Las Vegas version of The Street. High risk, high gain lottery tickets that got all the publicity like Amazon and Netscape, where an individual could make 40% gains or lose everything. Of course, the banks didn't fully disclose the way underwriting works - the bank gets to pre-allocate big chunks of the offering, handing out candy to the Big Boy customers to lure in more of their money. That meant that by the time Joe and Sally got a chance to buy shares the price was already inflated. The Lunchpails paid a premium and by the end of their buying wave the Big Boys would start to take profits, sucking the momentum out of the overvalued stock. In the technical argot this is called a Ponzi scheme, more colloquially "bait-and-switch."

This long-winded analysis of the dot-com bubble is really just a setup to explain why Open Source and Web 2.0 won't create another cycle of overvaluations and overnight millionaires. In case you haven't noticed, there isn't an IPO market, so that exit is effectively closed to startups. The Lunchpails licked their wounds after the 2001 bust, then started sinking their money into their homes and real estate in general. That's the current bubble that's about to burst. With middle class families' "savings" tied up in increasingly illiquid assets, there is little possibility of pools of capital flooding the equity markets and goosing the growth rate.

With no 5-year path to a 10X exit, the VCs are not nearly as excited about startups these days. In fact being a VC is much less glamourous in the 'naughties than it was in the 'nineties. Frankly, from the VCs I have spoken with it sounds like it kind of sucks. While The Street would like the VCs to find the golden egg that would hoover more pocket change out of the heartland and into their bank accounts, they still have lots of M&A activity, consumer spending, real estate and energy prices to drive their income.

So what is behind the buzz around Open Source and Web 2.0 startups, then? It is a little more than a nostalgic wish to return to "the halcyon '90s, when Pets.com ruled the world." There is an exit that allows a lucky few to realize some significant gains, but the emphasis is on "few." Both Open Source and the Web 2.0 obsession with quick, simple results threaten the incumbents in the IT industry - IBM, Microsoft, SAP, and Oracle. These giants are slugging it out as the industry consolidates around three big platforms - .Net, J2EE and SAP itself - and the last thing they want is to have to fight over a diminishing pie. But that's exactly the net effect on the market of the commoditization effects of Open Source and the 'do more quicker with less' approach of Web 2.0. Enabling the big guys' best customers to spend less on IT and get the same results is the best way to ensure an exit through acquisition. The buyer can be the incumbent you are hurting the most, the competitor most interested in hurting the incumbent, or a 2nd or 3rd tier IT company seeking a way to be more attractive to one of the big platforms and therefore ensure either their continued existence, or future acquisition for a bigger price tag. That was the force behind Red Hat's acquisition of JBoss, IBM's acquisition of Gluecode, Oracle's near acquisition of MySQL, and many more acquisitions to come.

If it sounds like I'm describing a world where little fish are gobbled up by successively larger fish, well, that's what a consolidating market looks like. Two important points, to circle back to our topic: first, that exit by acquisition nets 2-4 times your invesment, maybe 6X tops, but not the astronomical returns that a 'nineties IPO reaped. This is bad news for the typical VC portfolio management model of "one grand slam, two singles and seven strike outs," which depends on those 10X grand slams to even out the losses from the rest. Second, being a consolidating market, there is a finite and rather close horizon for this acquisition market and the exit it represents. So instead of a throwing a lit cigarette into a National Park, this flurry of activity is more like holding a lighter to a pile of trash in an barrel.

Of course, this analysis is not directly addressing the new face of the IT industry: the Google/Yahoo/eBay struggle. For the moment I think Cringely has written most of what needs to be said on that topic, and for those interested I can think of no better essays to read.