At Compaq, A Desktop Crystal BallGary Mcwilliams
On Mar. 3, Michael Parides sat in his office nervously scanning the spreadsheet running on his personal computer. Hewlett-Packard Co. had just slashed prices 22%--stepping up a campaign to grab more of the corporate PC market. Months earlier, Parides had persuaded his bosses at Compaq Computer Corp. to back a radical plan: Use complex simulation software to manage product introductions for the world's biggest PC maker. Parides' program was designed to model trends in customer demand, pricing, and even dealer inventories. As the screen filled, he relaxed. The program had anticipated HP's pricing to within a few dollars. "It all proved out," Parides quickly relayed to his boss.
For Parides, director of business operations at Compaq's Desktop Div., it was one more hurdle met and cleared. After eight grueling months battling skeptical colleagues and fatigue, his team and its software are playing a critical role in managing the nanosecond pace of new-product introductions--arguably the company's biggest challenge. A miscue can mean disaster. A model that's late, lacks the right features, or has the wrong price quickly turns into piles of unusable components or slow-moving finished goods. Just ask IBM, which misjudged the market last year and saw a 6% plunge in PC sales--and a $1 billion loss. Analysts say Compaq passed up $50 million in sales by underrating demand for home PCs last fall.
BIG TEST. Now, the company is determined not to repeat such mistakes. "What's going to count the most, as far as remaining competitive and profitable, is how you manage your business," says Kevin L. Bohren, marketing vice-president at the Desktop Div. "It used to be every new product had six months of uniqueness. Now, it's a long weekend."
Parides' system is now facing its biggest test. Starting Mar. 8 and continuing in April, Compaq will launch a series of business and consumer PCs to replace models that now account for 75% of its $11 billion in revenues. These machines--51 new business models for the North American market alone--are expected to help swell Compaq shipments to 6 million machines this year, up from 4.8 million in 1994.
Compaq has long used computer modeling--to weigh the impact of design decisions on component purchases, for example. But Parides, a buttoned-down former McKinsey & Co. consultant, has created a kind of SimCity for the entire PC business. Essentially a giant spreadsheet, it includes far more detail than other simulations, Compaq says. As a result, it can simulate conditions such as component price changes, fluctuating demand for a given feature or price, and the impact of rival models. By modeling supplier and competitor behavior, the system lets managers consider the risk of certain actions before taking them.
How good is the model? Insiders say it played a decisive role in convincing top management on the most critical decision of 1994. Starting last summer, competitors and pundits were roundly criticizing the Houston-based PC maker for its slow switch to machines using Intel Corp.'s new Pentium chip. While rivals such as Dell Computer Corp. and Gateway 2000 Inc. were doing a booming business in Pentiums, Compaq was amassing a huge inventory of PCs based on the older 486 design. Compaq stuck to its plan, because the model said most corporate buyers wouldn't want Pentiums until early 1995. The gamble proved successful: Fourth-quarter earnings rose 61%.
The new desktop PCs that Compaq rolled out on Mar. 8 are the first models designed from the ground up using "what if" scenarios created by Parides. The machines--ranging from $2,300 to $4,500--include competitive new features such as software that tells network managers what's inside the PC.
One thing Parides didn't forecast was how difficult it would be to create the software. Early on, he teamed up with Michael J. Sheehan, a lanky analyst plucked from Compaq's operations planning group. The first hurdle was getting cooperation from department heads who had never worked together. The second was getting the right data. Compaq marketers had never estimated demand elasticity for PCs, for example. When Parides' group met with other departments last June, their requests were so unusual, they got a skeptical reception. "Early on, there were a lot of question marks--was this the right approach? The right individuals?" recalls Bohren.
As the weeks wore on, the team struggled to get the needed data. A corporate reengineering program run by McKinsey was beginning to produce some of the information--but it was only starting. Morale sank in July when the team realized it had to scrap its first model and build a more detailed one. It was Parides' partner Sheehan who kept spirits high. After taking his kids to see Walt Disney Co.'s The Lion King, he took to belting out the film's Hakuna Matata--"no worries"--song whenever the team hit a snag. "No worries" became the group's motto.
TOUGH CALL. Then Parides dug in. He canceled a late-July vacation to run new scenarios and tweak the model to show the financial risks of moving up the announcement of Pentium PCs. The findings: Shifting too quickly to Pentium-based models would wreak havoc on pricing on Compaq's many 486-based units. It would also mean paying higher prices for new components than if it waited (Intel has since slashed Pentium prices).
For Compaq, which had often triumphed by bringing new technology to market as soon as possible, it was a tough call. "We bet against everything the company believed," says Parides. He took the case to Desktop Div. Senior Vice-President John T. Rose. "A bad decision and it's hundreds of millions" of dollars lost, Parides warned.
In mid-August, Rose got the group a hearing before the corporate financial group. By then, Compaq sales managers were lobbying for an accelerated switch to Pentiums. "We got a lot of `you don't understand, we've seen this before,"' recalls Parides. He pushed ahead, however, presenting to financial executives, then to the company's top strategists.
On Oct. 28, Parides and Sheehan found themselves before CEO Eckhard Pfeiffer. Their pitch: The company could boost profits by $50 million in the March, 1995, quarter by not speeding up the release of the new Pentium machines. Pfeiffer, who listened intently, gave the pair all the assurance they needed. "It's about time we're doing these things," he said.
Within 12 months, all Compaq computers will be designed with the help of Parides' new simulation tool. If it works--and confirmation may come when results for this quarter are released--Hakuna Matata could be Compaq's theme song.
COMPAQ'S NEW SCIENTIFIC METHOD
CUSTOMERS Conducts focus groups to gauge interest in and price sensitivity to new features
DEALERS Polls resellers on inventory levels and available cash to gauge the timing of new-model introductions
LOGISTICS Schedules key purchases and production processes only after development hurdles are met
SUPPLIERS Tracks suppliers' production and, with them, develops alternative production plans in the event forecasts are wrong