Last updated: August 23, 2002 12:14 pm

After the bubble: five lessons from dotcom mania

A year ago, I was a calm observer of the dotcom bubble. Since then, I have been immersed in the web frenzy, helping to edit the FT’s websites. So the thoughts that follow come from someone who has lived inside the web craze, but escaped the rollercoaster economics of a start-up.

There are five important lessons. The first is that much internet technology does not work. Few websites work smoothly; very few frequently updated websites avoid intermittent disasters. Some of this is because the vital applications - such as the middleware, which links together big databases and the web front end - have been rushed to market to capitalise on the web boom, and still suffer endemic bugs.

A bigger part of the problem, though, can be laid at the door of the website owners themselves. Under great commercial pressure to create and update their sites, they have slung together systems using the electronic equivalent of string, toilet roll cores and sticky-backed plastic. They have lashed them to existing databases never designed for such tasks. They have switched their existing networks and support systems from traditional batch operations - with peaks of activity once or twice a day - to serving continuously updated 24-hour websites. And once they have them working, they have promptly piled on new features, changed the design, or added millions of users.

The burden on the technology teams has been intolerable. As a result, the internet does not offer anything like the reliability associated with electronic systems such as bank cash machines or the telephone network.

But there is some good news. The underlying internet machinery - its protocols and backbone - have turned out to be able to withstand the insatiable demands. And website reliability does get better - though, infuriatingly, even the best-managed sites are capable of redesign or re-engineering bungles.

Technology problems have helped kill off some websites. They will also damage the prospects of the weakest of the internet infrastructure companies - a sub-sector of the market that has not yet been hurt quite as much as its clients. But in time the technology will become more reliable and websites will use it better. This is one lesson that the industry has already taken to heart.

The second lesson has to be learnt afresh in every bubble: most new business models do not work. Traditional business models dominate the economy because they have stood the test of time. Every generation or so, a genuinely new business model arrives, and takes its place among the tried-and-tested veterans: hire purchase, franchising, network commercial television. But most apparently new models have been tried before.

Take the radical breakthrough in advertising that the web supposedly allows. Advertisers can now track every click, and pay media owners only for the ones that deliver customers or at least warm prospects. Actually, anyone watching late-night TV is intimately familiar with the pay-per-click model, in its long-established form of telephone selling of compilation record albums and implausible kitchen equipment. Advertisers pay the TV station only for orders or inquiries that the advertisement generates.

It works, but it is marginal - because it is not in the interests of TV stations to devote any but the least valuable airtime to such a proposition. The clear winners already starting to emerge among the web media will pursue a similar approach as soon as the bargaining power swings their way.

Some of the new internet business models may yet attain lasting success. The jury is still out on consumer-to-consumer auction sites, like Ebay; and we do not yet know whether business exchanges like the auto industry’s Covisint will prove an innovation or only another way for monopoly buyers to squeeze their suppliers. But we do know that one promising idea - Priceline.com’s Dutch auctions for perishable commodities such as airline seats - seems to have limited appeal. Consumers like the predictability of posted prices.

Some dotcoms that touted new business models turned out to be using time-honoured techniques in a new wrapper. The claim of innovation was only a fancy way of saying: “We’re a new entrant with a distinguishing competence.” Alas, the iron law of business is that any incumbent that is half-awake is hard to beat, unless the incomer’s fancy new competence is a killer. That does not mean that none of the new entrants can succeed; but it does mean that the qualities required to prosper owe more to tradition than to mould-breaking.

How did they ever get financed? Well, that leads to the third lesson: nothing destroys value like underpriced capital. The dotcom bubble will be studied not so much for its technical or entrepreneurial innovation, but as a capital market anomaly. In the closing years of a long bull market, the suppliers of capital - much richer on paper than a decade before and more willing to take a flyer - poured money into venture capital funds and then into the premature initial public offerings that validated those venture capital investments. This virtuous circle brought in more venture money and more successful IPOs, rewarding investment bankers, lawyers and other facilitators handsomely in the process.

This surge of money produced the biggest mispricing of capital since the Japanese equity warrant boom of the 1980s. When capital is nearly free, the need to discriminate between new ideas disappears. Clearly, the dotcom boom was a disaster.

Or perhaps not. Because this is the fourth lesson of the boom: financing start-ups in the public markets is a terrible idea, except that it is better than the alternatives. Since the commercial banks established hegemony over the world’s financial systems at about the turn of the century, most periods of financial folly have been fully intermediated by the banking system. That helps to make the boom less flamboyant, partly because the personal risks and rewards for a typical commercial banker are less tempting than for an investment. But it also makes it much more dangerous, because of the role the banks play in accepting deposits and transmitting payments. Banks are too important to lose, so governments have to bail them out. Intermediated bubbles drag in everyone.

So the IPO surge is a relatively harmless way of soaking up a temporary mania. We have been rewarding those investment bankers so lavishly for performing the socially useful role of diverting irrational exuberance away from the banking system and into the public markets, where it can be absorbed with greater short-term foolishness but less long-term damage.

The web bubble has given us some fascinating new businesses, and some extremely interesting experiments, as well as a flood of me-too ventures that now face extinction (remember all those online pet stores?). The apparent initial success of the dotcoms pushed established business to react in months rather than in years. So the bubble produced a much more rapid transformation of existing businesses, and the much more lavish creation of a global electronic infrastructure.

That means that the web era has only just begun, which is the fifth lesson I draw from these past remarkable years. Stable technology, more mature business models, interpenetration of the “real” and the virtual world, the arrival of always-on broadband connections, and the creation of the rich, permanently connected handheld device - all mean that there is at least as much interesting innovation ahead as that which lies behind. There just won’t be a capital-markets frenzy to speed the process.

Copyright The Financial Times Limited 2014. You may share using our article tools.
Please don't cut articles from FT.com and redistribute by email or post to the web.

SHARE THIS QUOTE