In his new book, “The Well-Timed Strategy: Managing the Business Cycle for Competitive Advantage” (Wharton School Publishing, www.whartonsp.com), Peter Navarro offers abbreviated case studies of companies from around the globe.
Navarro, a professor in the Merage School of Business at the University of California, Irvine, doesn’t have to go far for two of his subjects,Orange County’s most prominent chipmakers, Broadcom Corp. and Conexant Systems Inc. Navarro praises Broadcom and pans Conexant for how they forecasted and reacted to the technology slump at the beginning of this decade.
This article is taken from the book.
Over-Optimism
The rags-to riches back to ragged Conexant flew like a meteor through the Internet bubble skies,only to fall on the sword of its own hubris.
Spun off by the defense company Rockwell in January of 1999, the company had sales that rocketed up to over $2 billion by the year 2000,a truly astonishing burst.
During this time, Conexant’s stock price increased more than sixfold. However, by the end of 2002, the company had devolved into an embarrassing “penny stock” forced to divest many of its assets.
Conexant’s executive team did not totally ignore the increasingly urgent signals of recession emanating from the economy. Rather, in a supreme irony, Conexant’s supremely over-confident top execs simply ignored readily available company data indicating dramatically falling chip demand.
Indeed, within Conexant’s own supply chain management “shop,” analysts saw quite clearly that inventories at the company’s distributor locations and in Taiwan were going up, that wafer supplies were becoming plentiful in Asia as capacity factors were going down and that customers were no longer complaining if an order was shipped a few days late. These internal indicators all showed that the business and chip cycles were showing signs of turning down.
When combined with the external indicators of a falling stock market, an oil price shock and a flattening yield curve, Conexant’s own internal data should have been a strong clue for its top management to take a hard, fresh look at its business situation. Instead, it was damn the recessionary torpedoes and full speed ahead.
The vehicle for this hubris was a rose-colored demand forecast that flew recklessly on the wings of hope and past performance while flying in the face of the reality of their very own real data. On the basis of these wildly over-optimistic forecasts, production plans were dutifully made, raw materials were dutifully purchased, and, worst of all, Conexant entered into “take or pay” long-term contracts that locked them into these overly optimistic production levels so they could meet forecast volumes.
The outcome was predictable. As the bottom fell out of the semiconductor market in late 2000, revenue performance dropped dramatically. In the (bitter) penny stock end, Conexant would wind up recording almost a billion dollars in inventory write-downs and other special charges over the next two years.
As the lead off to the 2001 annual report,a year in which Conexant lost almost $1.5 billion and shareholders lost over a billion dollars in equity,CEO Dwight Decker would have this to say:
“Conexant strengthened product portfolios and market positions throughout fiscal 2001 while carefully managing cash and expenditures during the deepest, most abrupt business reversal in the history of the semiconductor industry.”
Forgive me, but there was absolutely nothing that was careful about how Conexant managed its way through a recession that was very well-signaled and anything but “abrupt.”
Dual Strategy
Broadcom and its “cross town rival” Conexant are located within just a few miles from one another.
In honor of its own name, Broadcom’s specialty is in the area of broadband communications, and the company’s ambition is to “own” the high speed networks of the future by supplying chips to all its critical parts.
Toward this end, the company makes highly sophisticated chips called “integrated circuits” that can be used in a variety of broadband applications,from digital set-top boxes and cable modems for the home TV market to DSL lines, wireless applications, servers, and other networking “gear.”
In the late 1990s, Broadcom’s executive team adopted a two-pronged strategy that would quickly differentiate it from its neighbor and competitor Conexant. One prong of the strategy involved a conscious effort to diversify its product line across all elements of the broadband supply and product chains.
This Broadcom did through a very ambitious acquisition program that,unlike, Conexant, which encumbered itself with debt,it cleverly financed with the currency of its own high-priced stock.
The second prong of Broadcom’s “fabless” strategy involved outsourcing all of its chip fabrication needs. In this fabless model, rather than insisting on manufacturing its own chips, Broadcom chose to focus solely on the higher value-added design end of the business and offshored its production to chip “foundries” or “fabs” primarily in Southeast Asia.
This two-pronged narrow diversification and outsourcing strategy proved very effective when the economy weakened. During the technology crash, it was not Broadcom left holding the huge inventory stockpiles, but rather customers like Hewlett-Packard, Motorola and Cisco.
Over time, it has been Broadcom that has continued to increase its revenues and stock price while the “incredibly shrinking Conexant” has fallen by the penny stock wayside.
Fab Flub
How Conexant devolved in three years from a fab-driven company with almost 9,000 employees into a “fabless” former shell of itself with 1,400 employees and penny stock status is a cautionary tale for any corporation that chooses to ignore the risks of the business cycle.
In fact, Conexant’s downfall began with the initial decision of the executive team not to outsource much of its chip manufacturing.
The semiconductor industry is infamous for its volatile swings; and through the often wildly swinging industry cycles, invariably it is the large fabrication plants or “fabs” that bear the brunt of periodic chip gluts and plunging prices.
Moreover, chip fabrication is a process characterized by large economies of scale and rapid technological change so that the only way to be competitive in the market is to always be the biggest and most modern fab on the block.
So it was that within a year of its 1999 spinoff from Rockwell, the relatively small Conexant found itself playing well over its head with the really big boys,huge Taiwanese fabs like Taiwan Semiconductor and United Microelectronics that feed companies like Philips and Motorola with their chips.
In this fast-moving chip arena, Conexant belatedly realized that its level of capital expenditures were insufficient to stay on the leading edge of chip production the errors of its non-outsourcing ways.
At this painful point, the company made the decision to realign its strategy and go, like Broadcom, to a fabless outsourcing model. However, in executing this new strategy, Conexant stumbled further.
The dilemma the company faced was its belief that it had developed a very important specialty process technology involving silicon germanium or so-called “SiGe” chips. Its fear was that if it sold off its fab, it might not be able to retain and further develop its SiGe process technologies.
As a result of temporizing over this dilemma, Conexant lost any opportunity to sell off its fab at a decent price near the top of the semiconductor market. Instead, it wound up almost giving away 55% of this fab to the Carlyle Group for pennies on the dollar.
