Why a littte dot-com knowledge is a dangerous thing.
Priceline's Jay Walker got it first. He gave it to Idealab's Bill Gross, who passed it on to Divine InterVentures' Andrew Filipowski. Then it spread far and wide, and pretty much everyone caught it -- Jim Clark, Jeff Arnold, the partners of the Internet Capital Group. Before you knew it, entrepreneurs and executives at in-house Internet ventures, consulting firms, private equity funds and incubators were all stricken.
What they have is the dreaded Second System Syndrome, a business ailment first diagnosed by Frederick P. Brooks Jr. in his classic 1975 book The Mythical Man-Month. The book captures Brooks' insights as project manager for IBM's 360 operating system, the first great mainframe systems product. But his insights are invaluable well beyond the world of big software projects.
"An architect's first work is apt to be spare and clean," Brooks wrote. "He knows he doesn't know what he's doing, so he does it carefully and with great restraint." But once the first system is done, the architect becomes overconfident, unaware of how much he still doesn't know. The second system is usually overengineered, full of unnecessary "frills and embellishments" -- in short, a mess. It is, Brooks wrote, "the most dangerous system a man ever designs."
Many of the architects of the Internet Economy exhibit acute symptoms of Second System Syndrome. Worse, they seem to suffer from a particularly virulent strain. They are so eager to start building the second system -- or product or company or fund -- that they never get the first one working. While they are off trying to repeat their "success" and teach others how to do it, the first project languishes, misfires and runs out of gas.
Second System Syndrome, I believe, was the principal cause of last year's Internet meltdown -- the real Y2K bug. So understanding how the disease spreads and how it can be cured can go a long way toward getting us back on course for sustainable and successful e-business.
MAKING THE ROUNDS
Jay Walker has the most serious case of the disease I've ever seen. Walker, you'll remember, came up with the truly great idea of selling airline seats based on what consumers are willing to pay for them. That innovation offered considerable promise in fixing the airline industry's ruinous post-deregulation "yield management" system -- a problem big enough to keep his company busy for years.
But instead of refining the nascent Priceline model, Walker extended it to commodities as varied as home mortgages and groceries -- markets that were not as ripe for repair. Priceline wasted hundreds of millions of dollars on the expansion and now faces a doubtful future. At the end of December, Walker quit the board and turned his attention to what was left of his other company, Walker Digital -- a "business solution invention" company--which aims to come up with more breakthrough ideas and patent them.
Brooks wrote 25 years ago that a good architect has to finish three or four successful systems before he can begin to "identify those parts of his experience that are particular and not generalizable." Without that experience he still doesn't know what part of his success is solid engineering and what part is dumb luck -- which is what makes the second system so dangerous. You can't generalize from a sample of one.
The keyword here is "finish." Priceline hadn't been in business long enough to prove its model could be generalized to other products. And having one good idea does not mean you can have hundreds more on demand. Even before he'd finished testing his first idea, Walker tried to abstract it -- not once, but twice. Brooks could have predicted the result.
Generalizing too soon is a mistake Walker shares with the founders of many Internet incubators. These are businesses that create other businesses, and so are particularly vulnerable to the Second System Syndrome -- especially when they are run by entrepreneurs or consultants who have never even built a first company. [See "Jurassic Business Park," April 10, 2000.]
As a consultant to incubators, I don't want to suggest that they can't do great things, just that they ought to be run by true Internet business experts. Not one member of the senior staff of Divine InterVentures, for example, has built a successful Internet venture. That includes founder Andrew Filipowski, whose previous company, Platinum Technology, sold mainframe application software and Y2K solutions -- far from relevant experience for building consumer-oriented Web companies. How did Filipowski expect to launch multiple Internet ventures simultaneously without having completed even one?
In Divine's case, the market woke up quickly. The company completed a lack-luster IPO in July and has been stuck in a tar pit of bad news ever since. Today, Divine's stock is worth about half of the company's reported assets.
SPREADING THE VIRUS
Second System Syndrome ultimately led investors last year to bail out of unfinished companies. It was those very investors, however, whose enthusiasm helped spread the disease. How? Their money gave entrepreneurs and others a false sense that their first ventures had succeeded. A successful IPO sure looked like a victory, and it seduced the founders into thinking they had hit upon a new formula for building companies: Articulate a good idea, build the shell for a business to exploit it, get going on the next one.
The willingness of public investors to buy into early-stage ventures had other unfortunate side effects. It panicked otherwise rational businesspeople -- including brick-and-mortar types -- into believing that time-to-market had not simply compressed, but had collapsed altogether. The markets encouraged them to launch their next ventures before the party was over, and before they were close to proving their first ones.
What no one seems to have noticed is that over the last two years the concept of "Internet time" mutated in a dangerous way. It's true that Moore's Law accelerated the speed with which successful technology ventures could be launched. But the mania for Internet IPOs collapsed development time to almost nothing. Now, hopefully, we are back on solid footing. If the venture process is going to work this time, however, we need to cure everyone who caught Second System Syndrome in the interim.
THE ANTIDOTE
The solution Brooks outlined in The Mythical Man-Month is not particularly practical. He warned project managers to not accept systems architects with fewer than two finished systems under their belts. But some project managers must do so, otherwise no one would ever build more than one system. I offer a simple solution that is equally applicable to entrepreneurs, venture capitalists, consultants and heads of e-business at traditional companies:
1. Finish your first venture before you start the second one.
2. Don't even think about generalizing your experience into a process until you've finished at least two ventures.
The solution is simple but not easy. Finishing is hard. It means, in the case of a business or project, that you have a freestanding, self-sufficient and profitable operation; or, if the idea didn't play out as expected, a noble failure that has been definitively terminated. No more declaring victory and walking away after the first positive press coverage, or after the initial partners are signed up, or even after the the first day of trading following an IPO.
A current example: For all the self-congratulatory ink that has been spilled about Covisint, the auto-parts supplier exchange, there has yet to be a single transaction processed. Agreeing on a name is a big deal for such a delicate alliance, but it is hardly the stuff of business legends. It might be good PR to call yourself the winner before you have hired a CEO (Covisint still can't find one), but excessive rhetoric is contagious -- and Second System Syndrome is what it spreads.
For a consumer-oriented venture, success is not a function of how many millions of users you sign up. Instead, it is measured by your ability to serve them on a consistent, sustainable basis. That is what makes McDonald's a $43 billion company, not the "concept" of fast food. Once you take your eye off that ball -- perhaps to expand, perhaps to launch other "related" ventures -- you begin a short, nearly invisible and nearly irreversible slide into failure.
Finishing the first venture does not mean, by the way, pursuing only one idea at a time. Given the probability that even your best idea may not lead to an industry-altering killer app, it is important at all times to maintain a portfolio of ideas and to test out as many of them as you can. Just don't launch them all as full-fledged businesses.
Or get other people to do it for you.
Larry Downes (ldownes@killer-apps.com) is on e-business consultant and a co-author of Unleashing the Killer App: Digital Strategies for Market Dominance (Harvard Business School Press).
COPYRIGHT 2001 Standard Media International
COPYRIGHT 2001 Gale Group