Cisco Shopped till It Nearly Dropped

It was an all-too-typical deal for Cisco Systems Inc. (CSCO ) Monterey Networks Inc., an optical-routing startup in which Cisco held a minority stake, was a quarry with no revenue, no products, and no customers--just millions in losses it had racked up since its founding in 1997. Despite those deficits, Cisco plunked down a half-billion dollars in stock to buy the rest of the company in 1999.

But within days of closing the deal, all three of Monterey's founders, including its engineering guru and chief systems architect, walked out the door, taking with them millions of dollars in gains from the sale. "I came to the realization I wasn't going to have any meaningful impact on the product by staying," says H. Michael Zadikian, a Monterey founder. Eighteen months later, Cisco shut down the business altogether, sacking the rest of the management team and taking a $108 million write-off.

That dismal tale hardly jibes with Cisco's widespread reputation as an acquisitions whiz. Not since the conglomerate era has a company relied so heavily on its ability to identify, acquire, and integrate other companies for growth. CEO John T. Chambers believed that if Cisco lacked the internal resources to develop new products in six months, it had to buy its way into the market or miss the window of opportunity. Some put a new name on it: acquisitions and development, a way for the company to shortcut the usual research cycle. Its belief in the strategy has led Cisco to gobble up more than 70 companies in the past eight years. Analysts and academics heaped praise on Cisco's acquisitions prowess in articles, books, and business-school case studies.

In the early days, some of this praise was deserved, as Cisco morphed from a router company to a networking powerhouse. Its first acquisition, Crescendo Communications Inc., guided Cisco into the switching business, which generated $10 billion in sales last year. All told, acquisitions have laid the foundation for about 50% of Cisco's business.

But in early 1999, with exuberant investors enticing a growing number of unproven companies to go public, Cisco suddenly had to acquire companies at a much earlier stage. Cisco had long claimed an unprecedented success rate of 80% with its acquisitions. Chambers now says it fell to something like 50% during the Internet craze--still above the industry average. "We bet on products 12 to 18 months out," concedes Chambers. "We took dramatically higher risks."

Chambers often maintained that his acquisition strategy was aimed at acquiring brainpower more than products. But an analysis of the 18 acquisitions Cisco made in 1999 shows that Monterey was no fluke. Many of the most valuable employees, the highly driven founders and chief executives of these acquired companies, have since bolted, taking with them a good deal of the expertise and experience for which Cisco paid top dollar.

The two founders of StratumOne Communications Inc., a maker of optical semiconductors purchased for $435 million, left Cisco. The chief exec of GeoTel Communications Corp., a call-routing outfit acquired for $2 billion, walked out after nine months. So did the CEOs or founders of Sentient Networks, MaxComm Technologies, WebLine Communications, Tasmania Network Systems, Aironet Wireless Communications, V-Bits, and Worldwide Data Systems--all high-priced acquisitions in 1999. Some simply felt Cisco had become too big and too slow. "People who crave risk don't do so well at Cisco," says Narad Networks CEO Dev Gupta, who sold Dagaz and MaxComm Technologies Inc. to Cisco in 1997 and 1999, respectively. "Cisco focuses much more on immediate customer needs, less on high-wire technology development that customers may want two to three years out."

Chambers maintains that Cisco's turnover rates are the best in high technology. "In our industry, 40% to 80% of the top management team and top engineers are gone within two years," he says. "Our voluntary attrition rate is about 12% over two years."

Difficulty holding on to top talent was not the only flaw in the Cisco acquisition machine. Cisco often paid outrageous sums for these unprofitable startups--a total of $15 billion in 1999 alone. Even some of the deals that Cisco considers successful look pretty dreadful using simple math. Its 1999 acquisition of Cerent Corp., a maker of optical-networking gear, is a good example. Cisco paid $6.9 billion for the company, or $24 million for each of Cerent's 285 employees, even though the company had never earned a penny of profit and had an accumulated deficit of $60 million. Even if earnings bounce back to 2000 levels of roughly $335 million, it would take Cisco about 20 years to recoup the purchase price.

Of course, deals such as Cerent found their rationale in Wall Street math. If investors were willing to pay 100 times earnings for Cisco's stock in 1999, then a Cerent profit of, say, $300 million could effectively increase the market cap of Cisco by some $30 billion. Call it bubble economics. Besides, many of these deals were done for highly inflated Cisco stock instead of cash. Even so, that wampum could have been used to buy other assets that could have delivered greater returns.

Only in the months since the bubble burst has it become evident just how muddled Cisco's mergers-and-acquisitions strategy became. In its haste to do deals, Cisco often purchased companies it didn't need or couldn't use. In some cases, the buying spree led to overlapping, duplicative technologies, political infighting, and just plain wasted resources, as Monterey shows. "M&A works to some extent, but at Cisco, it got out of hand," says Iqbal Husain, a former engineering executive at Cisco.

After losing many of the leaders of these businesses, product delays and other mishaps were not uncommon. When Cisco closed down Monterey, for example, the company still hadn't put a product out for testing, which alone would take as long as a full year. "By the time the product was there to test, the market wasn't," says Joseph Bass, former CEO of Monterey.

Chambers says he has moved to correct the flaws. Its acquisition binge has slowed--from 41 companies from 1999 through 2000 to just two purchases in 2001. While Chambers expects to do 8 to 12 acquisitions this year, he insists that market conditions will let Cisco wait at least until a target company has a proven product, customers, and management team before cutting a deal. "We're making the decisions to acquire a company based on a later point in time, which dramatically lowers the risk," Chambers says. Anything more ambitious, Cisco now knows, may be foolhardy.

By John A. Byrne and Ben Elgin in San Jose, Calif.

    Before it's here, it's on the Bloomberg Terminal.