The Networking Nightmare of '97
December 30, 1997
By Noel Preston Lindsay
Before last year, it looked as though the party would never end. But for much of 1997, the volatility in data networking stocks was enough to leave even the most experienced investor feeling bewildered, wondering: What conditions caused this volatility, and are they likely to recur in 1998?
Data networking equipment vendors have routinely faced a variety of challenges, such as aggressive pricing, product and technology transition, pauses in purchasing because of customer confusion over the "right" network architecture, irregular purchasing patterns by large customers and, of course, economic weakness, especially within specific geographic regions. In the past, the industry's strong underlying growth had enabled the major players to coast effortlessly through any one or two of these difficulties. Unfortunately, all these conditions coincided in early 1997.
A Bit of History
As the data networking industry developed, each of the leading companies consolidated their positions within a particular portion of the network. Cisco Systems Inc. of San Jose became the leading vendor of the routers used to build the infrastructure of corporate networks and the Internet. Santa Clara, Calif.-based 3Com Corp. emerged as the leading supplier of network adapter cards that are installed into PCs and servers, as well as other low-end networking systems that could also be sold through its two-tier distribution network. Bay Networks Inc. of Santa Clara and Cabletron Systems Inc. of Rochester, N.H., battled for the middle ground, selling the hubs used to connect PCs and servers to each other and to the router-based "backbone" of the network. Cabletron used its direct sales organization to focus on controlling specific accounts, while Bay leveraged its giant reseller channel to achieve broader market coverage.
But during the past two to three years, all these companies have broadened their focus; all are now targeting the entire market. As a result, direct competition has intensified dramatically, although for most of this time, strong demand mitigated the need for price competition, allowing the major players to realize relatively high gross margins. By January 1997, however, economic problems in Japan and some European countries had begun to affect the leading vendors, which had come to rely on the high growth rates in some of these regions, particularly Japan.
Also by this time, Internet service providers were completing a buying binge that had dramatically expanded their capacity during the latter part of 1996. As these service providers attempted to digest their new capacity, a series of highly public debates erupted over the best way to design wide-area network architectures. Predictably, this combination of excess capacity and confusion caused a protracted pause in purchasing that didn't end until the second half of 1997.
This pause, combined with regional economic weakness, triggered an outbreak of severe price cutting among industry leaders. By late January 1997, even industry powerhouse Cisco Systems began to caution investors that it might not meet analysts' expectations. Not surprisingly, Cisco's stock paid a heavy price for this warning.
Troubles at 3Com
Meanwhile, at 3Com's analyst meeting in January 1997, the company warned investors that its distribution channel had accumulated excess inventory, which could temporarily slow revenue growth. Days later, Intel Corp. of San Jose announced price cuts of 40 percent on Fast Ethernet adapter cards, and 3Com immediately matched Intel's price cuts. But unlike Intel, 3Com had not yet finished developing a low-cost version of this product, so the price cut significantly affected the company's overall gross margins. This one-two punch of bad news led analysts to lower earnings estimates, triggering such a severe sell-off of 3Com's stock that it even surprised the company's critics.
Throughout February, March and April, data networking stocks performed poorly, although Cabletron started to recover in late April. Earlier, management had said that the company had not felt the pricing pressure that other vendors had reported, nor had it experienced any unusual weakness in the geographic regions, which had plagued the other industry leaders. Consequently, many people thought that Cabletron would be the rising star of the sector.
Then on May 6, the clouds parted, the sun shined, and Cisco announced earnings that were in line with expectations. Both Cisco and 3Com recouped much of their declines. By that time, Bay Networks was already beginning to make a major comeback, and its stock finally reversed its 14-month slide.
Cisco's announcement gave Cabletron's stock price an additional boost, continuing an increase from $30 per share in late April to $46 per share in late May. But in a surprising pre-announcement on June 2, Cabletron said it would miss analysts' expectations, citing a sudden and severe slowdown in demand from Europe. The next morning, Cabletron's stock lost one-third of its value.
The other leading networking stocks seemed unfazed by Cabletron's problems, however. Cisco resumed its typical upward march, and Bay Networks streaked skyward as it became the momentum stock du jour. That all changed in late October, when the Asian markets melted down, reversing much of the gain that had been made in networking stock prices.
October's Asian Surprise
The Asia crisis, combined with a merely "in line" earnings release from Bay Networks, resulted in a loss of the critical upward stock-price movement required to maintain the company's momentum investors. So while all networking stocks were hurt by the Asian economic crisis, Bay's stock lost nearly a third of its value in 12 trading sessions. At that time, network-switching product maker Xylan Corp. of Calabasas, Calif., derived about 25 percent of its revenue from Asia and more than 10 percent from South Korea. As one might expect, its stock was also hit hard. Also at about the same time, it became apparent that the channel inventories at U.S. Robotics were higher than investors had realized. Given that 3Com had just acquired the modem maker, its stock price slid from more than $55 per share to less than $31 per share in about a month, as analysts again cut their estimates.
In the midst of the crisis, however, there were some bright spots. On the day after the Dow's biggest one-day point drop in history, MMC Networks Inc. of Sunnyvale, Calif., completed its IPO, and its stock price rose from an opening price of $11 per share to more than $25 per share.
1998 and Beyond
Does understanding the events of 1997 give us any insight into the likelihood of a repeat performance in 1998? It is possible that the data networking industry could again see economically induced weakness, another pause in purchasing because of confusion over new technology or another round of aggressive price competition. But are all these factors likely to coincide, as they did in 1997? After examining the sequence of events, a repeat performance would seem unlikely. It also seems unlikely that so many company-specific issues would again occur with such uncomfortable regularity.
Assuming that we don't have an encore of 1997's "issues," what should we expect in terms of growth in the networking sector? Experienced investors have recently wondered if the underlying growth in demand for networking equipment has slowed. It's difficult to tell for sure, but the same engines of growth that have driven the data networking sector for the past two to three years still appear powerful:
The need to upgrade and build existing networks. Today's networking infrastructure is quickly becoming obsolete, as network managers seek to increase data rates and better support delay-sensitive traffic such as voice, video and real-time interactive applications. Of course, there will always be the need to continue building the network as the size of the organization grows.
The need to connect small branch offices and individuals working remotely to the enterprise network. As inexpensive remote-access routers and servers have become available, it has become much more feasible to connect even the smallest remote sales office to the rest of the enterprise network. Also, more people are dialing into the network each year, either from the road or from a home office. This trend is likely to accelerate as new "virtual private networking" technology makes it feasible to replace long-distance dial-up links with "virtual connections" via the Internet.
The need to deploy new networks. While most of the large companies in North America have already been networked, it appears that an increasing number of small businesses will need to deploy a data network for the first time. Outside North America, many large companies have yet to deploy their first network.
The need to continue expanding the infrastructure of the Internet. I don't want to sound like a hype-monger, but the Internet is growing at a truly explosive rate. Virtually every element of the Internet infrastructure is in a constant state of expansion or re-engineering, and this growth should continue to drive demand for networking products.
The need for consumers to gain better access to the Internet. In years past, consumers have hungered for ever-faster modems. However, even the fastest analog telephone modem is pathetically slow relative to the needs of the emerging "wired" consumer. If the Internet is to realize its full potential, consumers will need to aggressively deploy high-speed cable modems, xDSL modems and ISDN adapters for use in the home.
The volatility of 1997 should remind us that short-term dynamics can indeed have a profound impact on the investment quality of an entire sector, regardless of the underlying strength in demand. However, the fundamental engines of growth that have propelled networking stocks in the previous few years appear to be strong. Indeed, the weakness in 1997 may set the stage for the strongest growth yet in 1998.
Noel Preston Lindsay is the director of equity research at Deutsche Morgan Grenfell's San Francisco office.
from UPSIDE |
