Blind Faith

Baldwin Louie — Mon, 16 Apr 2001 15:00:00 -0700

In November, at an upbeat meeting with an analyst, the chief executive for Canadian telecom-equipment titan Nortel, John Roth, blithely projected that his company would grow at the same healthy clip it had the year before. "The demand from society and from corporations in particular is not going away," he reported.

That same month, Cisco Systems (CSCO)' John Chambers, a CEO who has produced 11 years of steady revenue growth for the network equipment maker, told 500 analysts that he had never been more optimistic. And Hewlett-Packard (HWP)'s Carly Fiorina came out days later with ambitious growth projections of her own.

All of them were dead wrong.

In the past four months, the tech industry has experienced the sharpest business downturn in its history, with the shattered forecasts of company after company in the sector provoking a $3 trillion loss in the total market value of the Nasdaq in just over a year.

The cratering has leveled the industry and slowed the deployment of new of chips, software and Internet gear that was to spur a leap in business productivity and global expansion.

The hubris has been replaced with a deep lack of confidence in earnings projections, with "limited visibility" the new industry buzzword. It means a firm's leaders can't say where their company is going. Why it happened is a story of how blind ambition, terminal optimism and technology hype brought an overconfident industry crashing down to Earth.

Technology companies are not really that different from firms in other industries when it comes to developing products and generating sales projections: They talk to customers and gauge market expectations.

Indeed, as recently as five months ago, technology companies were thought to be better-positioned for even a sharp downturn; after all, they were using new software systems for forecasting demand and preventing dreaded inventories buildup. But nothing could protect them from the intense pressure to improve earnings quarter after quarter.

"Nobody wanted to flinch," says Edward Barnholt, chief executive of Agilent Technologies (A). "Everybody was building for 50 percent market growth. We didn't want to miss the opportunity on the upside if business was to go for another six months."

This shoot-for-the-stars attitude was evident at Palm (PALM). A determination to dominate the market for handheld devices caused the Silicon Valley company to pursue ambitious expansion plans, building inventory for the launch of two new products.

Unfortunately, the company missed warning signs that it would fall short of third-quarter earnings projections; Palm is now saddled with inventory that rose 200 percent to $102 million. Palm chief executive Carl Yankowski labored to explain the sales slowdown with words like "jarring" and "abrupt."

Palm wasn't alone in misreading the tea leaves. In the business-to-business software market alone, venture capitalists poured $25 billion into 1,000 companies in 2000. Old-economy companies, threatened by Internet-savvy upstarts, accelerated technology purchases. That helped lift IT spending to unprecedented levels: 20 percent growth each year from 1997 to 2000, an increase over the 10 percent growth during the previous seven years.

The spending orgy led to overcapacity in everything from the Internet to semiconductor production. The seeds of the downturn were sown. And the few who urged caution were largely ignored. "No one wanted to believe our numbers because our numbers didn't impress investors and didn't help their business plans," says Hilary Mine, telecom analyst with Probe Research, which forecast an aggressive 100 percent growth. "But everyone wanted to hear 800 to 1,000 percent."

Michael Crosno, chief executive of software company Epicentric, admits that by September he noticed sales were taking longer to close: 150 days instead of 115. "It was building up like a water balloon," Crosno now recognizes, yet he was one of many executives in high tech who ignored the inevitable and did nothing. "We all got a little fat and sassy," Crosno says. "You started believing your own hype."

If Crosno and his fellow execs had been really listening to customers, they might've foreseen the decline in spending. Staples spent aggressively in 2000 building out its fledgling online sales operations to handle sales that were rising twice as fast as expected. The company wrote check after check to vendors such as Digital River (DRIV), IBM (IBM), Trigo Technologies and WebMethods. It invested heavily in software to manage inventory, track orders and link to its existing pricing database, and it added more networking capacity to handle heavy Web traffic.

But the company wasn't planning to continue that buying binge, because "a lot of the hard stuff has already been put into place," says Anne-Marie Keane, VP of b-to-b e-commerce.

Reality hit the industry hard by December as companies prepared pre-announcements that would bring the world's most dramatic rise of industry and wealth to an end.

As market researchers adjust to the sobering reality of a deep downturn, lower expectations are emerging. In mid-March, analysts from Gartner Group chopped back an ambitious forecast for transactions that will flow over the Internet by 2004. Transactions should now total $6 trillion instead of the $7.3 trillion Gartner projected in January 2000.

"It was a big party," says Agilent's Barnholt. "The bigger the party, the bigger the hangover."

So why was the industry caught off guard? After all, high-tech firms are supposed to know all about technology and how to use it to predict what's around the corner. Surely they should have been able to see the warning signs within their factories, warehouses and sales ledgers. In June, chip factories saw chip-testing equipment operating at more than 97 percent of capacity, says G. Dan Hutcheson, president of VLSI Research, a market research firm. By August, that number had fallen to 91.4 percent, a clear indication of idle machines. By October, chip prices had fallen 9 percent, Hutcheson says, foreshadowing slowing demand.

In public, executives maintained their bullish outlooks. And in private, companies such as Cisco were reluctant - and late - to tell suppliers to slow shipments and manufacturers to idle assembly lines.

Their unwillingness to admit to lower forecasts caught distributors by surprise. Suddenly truckloads of components and circuit boards started coming back marked "return." In response, distributors began charging restocking fees to offset lost revenue, costing tech firms millions of dollars.

These problems were supposed to be solved by the companies' newly installed software systems. But the products didn't live up to expectations - an ironic twist given their high-tech owners. The systems were supposed to link suppliers, vendors and customers in a continuous data network, so a canceled customer order would slow component purchases throughout the supply chain. The goal was to make it possible for companies to catch changes in demand and prevent the buildup of inventory.

"I wouldn't blame the software," says Steven Katz, CEO of PowerMarket, a maker of supply-chain software. Many systems are "relatively new and not fully deployed. They did what they were supposed to do with the limited data that was being put in and the limited number of people who had access to that data."

As a result, inventories started rising, climbing 29 percent at Cisco, 15 percent at Intel (INTC) and 28 percent at Nortel. Outsourcers such as Flextronics and networking partsmakers like PMC Sierra and Broadcom (BRCM) were also overwhelmed by their stockpiles.

The current downturn is pushing companies such as Motorola (MOT) and Flextronics to install new global supply-chain repositories to enable executives to view the entire picture.

It is also bringing about a new sense of discipline. For all its technical prowess, the forecasting failure illustrates just how much the industry still depends on human instincts and gut decisions - as the old Silicon Valley hardware company rubric goes, "Have a hunch, buy a bunch."

Forecasting software is only as good as the people who run it. It relies on subjective assumptions about economic growth. When growth is constant, the software tends to work well; otherwise, the software isn't as reliable.

Every quarter, executives collect data from their sales forces to estimate business prospects. With this sales data in hand, they assign probabilities to each deal: How big is it? Can a competitor steal it away? But doing so requires them to make assumptions about factors like market demand. "The thing we don't have is a crystal ball," says Oracle (ORCL) senior VP Jeremy Burton. "We don't have visibility into the state of mind of the man who signs the contract."

No crystal balls, indeed. Suddenly the same high-tech execs who've had no problem projecting huge growth through the boom years complain that they have little insight into the immediate future. HP's Fiorina said in February that "visibility remains extremely limited." Inktomi (INKT) CEO David Peterschmidt echoed to analysts last week: "As for visibility, there isn't any."

But as much as it sounds like an easy excuse - don't bother asking 'cause I don't know - there is something to it. The technology industry, in fact, has entered a remarkable phase, a time when many assumptions about technology's promise and prowess are proving wrong, and criticisms about the industry's weaknesses and bullishness are proving right.

The forecasting problem erupted because those who were supposed to know where the industry was heading didn't, but thought they did. Until their forecasts and strategies reflect reality rather than ambition, limited visibility is the best they can offer.

The Industry Standard - Blind Faith - Comments

Blind Faith