Compaq Leads the Revolt

The press conference went very smoothly. Although the Compaq team had done almost all the work to this point, we were very careful to make it clear that this was a coalition of equals, with every company heavily involved. In reality, we had contributed about 90 percent of the effort, HP about 9 percent, and all the others together, about 1 percent.

It was disappointing that neither Bill Gates from Microsoft nor Andy Grove from Intel attended the event. Even though the chairmen of both companies had given speeches at Compaq’s 386 launch in September 1986, apparently both felt this would have been a bigger problem for their relationships with IBM. Both had made it clear to me they wanted the coalition to succeed, and sending other top executives showed strong support.

On September 14, the day after the press conference, the coalition ran a jointly sponsored full-page ad in the Wall Street Journal. In giant bold type running down the middle of the ad was the punch line: “OVER THE PAST 8 YEARS, BUSINESS HAS INVESTED MORE THAN $100,000,000,000 IN INDUSTRY-STANDARD PERSONAL COMPUTING. TODAY IT JUST PAID OFF.” We had been beating up on IBM about abandoning all the industry-standard hardware and software investment ever since the PS/2 announcement. But this ad was like shouting directly in its face.

Remarkably, IBM decided to make an announcement on the same day—and at the same time—as the EISA press conference. IBM’s announcement was held at its headquarters in midtown Manhattan. One part of its presentation focused on the PS/2 Model 30 286, a low-end computer that used the industry-standard bus instead of its own Micro Channel. As a low-cost, low-end product, the computer made perfect sense. But announcing it on the same day as the EISA announcement didn’t. It gave the press and analysts a tempting comparison that was unfavorable to IBM.

IBM had almost always followed a policy of not commenting on competitive announcements, but that day, its executives spent much of the Q&A period defending Micro Channel. Although they knew very few facts about EISA, that didn’t stop them from attacking it. The IBM spin machine was running at full speed.

Since there was no World Wide Web at the time, the first media outlet to carry the story was television, in particular, business news programs such as Moneyline on CNN. In one of the first segments to appear, Lou Dobbs opened with: “Big Blue is doing some embarrassing backpedaling in marketing. The backpedaling brought on as many of its competitors mounted an historic challenge to IBM’s role as creator of standards for the PC industry.” Dobbs then introduced Steve Young to report the story.

The video cuts to a production line of soft drink bottles moving swiftly along. Young says, “What, you ask, does New Coke have to do with IBM?” He explains that when IBM introduced a new personal computer “flavor,” the PS/2, it rendered useless $12 billion worth of software that businesses had purchased. That caused IBM’s market share to drop significantly, he noted, and as a result other computer companies were selling two-thirds of the PCs bought from dealers. “Faced with overwhelming resistance to the flavor of ‘New Compute,’” Young says, “IBM today quietly, very quietly reintroduced ‘Classic Compute.’ It looks like the new computers, but has a souped-up version of the old AT brain. Is it an effort, as many analysts contend, to regain market share by bringing back ‘Classic Compute’?”

Later Young suggests that IBM may continue to lose market share because its competitors agreed to use the newly announced EISA bus instead of IBM’s. It cuts to a shot of me speaking at the EISA press conference saying that because EISA is compatible with the existing industry standard, it protects the investment of millions of PC users worldwide. Young wraps up the report with, “Some analysts think it was a day that left Big Blue looking black and blue.”

I couldn’t believe it when I saw this news report. I thought the story couldn’t have been more positive toward us if I had written it myself. I had used that “New Coke” analogy during the April 1987 interviews just after the PS/2 was announced. It had worked well before and apparently was an irresistible analogy for the media again.

Other news reports were similar, but not quite that positive. They included Neil Cavuto on the Nightly Business Report, Doug Ramsey on the Financial News Network’s Business This Morning, and Bob Jamieson on NBC’s Before Hours, among others. The common theme: The industry IBM had led for seven years was in revolt, and it didn’t look good for IBM. They all pointed out that the company had been losing market share since the introduction of its PS/2 and reported our EISA announcement could accelerate the trend. Most of the reports quoted IBM executives. While they each had smooth deliveries, they all came across as defensive.

Soon there were major articles in all the business newspapers and magazines. For the first time in Compaq’s history, it seemed to me the analysts’ quotes were generally negative toward IBM and favorable toward Compaq’s, and EISA’s, chances for success.

The longest article was in the Wall Street Journal. On September 14 in an article titled “Nine Firms That Make Personal Computers Gang Up Against IBM,” Paul Carroll and Michael Miller wrote, “Nine big personal computer makers are doing what computer companies have dreamed of doing for decades: ganging up on International Business Machines Corp. They are joining forces in an audacious attempt to wrest away from IBM the power of setting the standard for how personal computers are designed, and they seem to have a chance of succeeding.”

Several analysts and consultants were quoted in the article. Aaron Goldberg, vice president of IDC, noted that this was the first time the compatible industry was deciding its own future. Tim Bajarin, a consultant at Creative Strategies Research International, said that if there was ever a chance to force IBM to lose its dominant industry position, this was it. Jonathan Yannis, a Gartner Group analyst, noted that there was a “better-than-even chance” for the coalition to consign the Micro Channel to the subset of buyers who go with IBM simply because it’s IBM. And Charles Wolf, an analyst at First Boston Corp. who had been pessimistic about Compaq after IBM introduced the PS/2, did an about-face and predicted Compaq would be the big winner if the effort was successful.

The article concluded that Compaq had the bulk of the risk: “Because in recent years it has been fast building a reputation for the best technology in the industry, it has the most to lose. And few companies that are No. 3 in an industry—Compaq’s position behind IBM and Apple—have achieved the sort of influence that Compaq is seeking.”

Its comment on the coalition’s impact on IBM was chilling: “For IBM, the gang’s announcement yesterday is at best a dust storm of confusion, and, at worst, a dagger to the heart of its PC strategy.”

Similar, but shorter, articles appeared in the New York Times, BusinessWeek, the Financial Times, and USA Today. In a September 14 USA Today article titled “IBM rivals challenge its standard,” Katherine Hafner closed with, “A group of manufacturers that prospered by aping IBM’s standards may succeed in calling the tune on personal computer design. And Big Blue could end up cloning the cloners.”

Peter Lewis, in a September 11 New York Times article titled “A Big Battle Over Computer Buses,” wrote, “Just to confuse things further, IBM will also make an announcement Tuesday, introducing, among other products, the latest model of the Personal System/2 family—a Model 30 that uses the old AT chip, the 80286, and—surprise!—the old 16-bit AT bus architecture. That makes it sound as though IBM is making a U-turn into oncoming traffic.”

Coincidentally, the Salomon Brothers Microcomputer Conference was held in New York the day after the EISA announcement. With both IBM’s Lowe and me speaking, there was no way to avoid contact. Actually, I was glad to have the opportunity to set the record straight with regard to the spin and misstatements coming from IBM. If the press thought I had suddenly grown claws when I became more aggressive after the PS/2 announcement in April 1987, they were now going to see me crank it up another notch or two.

I was sitting in the audience when Lowe gave the luncheon address. He hadn’t seen EISA’s specification, he said, so he couldn’t make an authoritative comparison between the two buses. But he added, “We think Micro Channel is more robust.” He failed to say specifically on what he based that assertion. Later, he said, “We’ve protected our customers’ investments in that Micro Channel ensures you can take advantage of new technologies for years to come.”

When a reporter approached me after lunch, I was ready. IBM had been promising that machines with Micro Channel would be able to take advantage of “new technologies” and “new applications” for almost a year and a half since the introduction of the PS/2, and still had nothing to show. Also, when our engineers were analyzing Micro Channel and developing EISA, they discovered a severe weakness in the Micro Channel–based PS/2 machines IBM had already shipped. Those computers lacked a memory cache for the processor, which meant that most didn’t have enough performance to deliver on IBM’s promise.

I told the reporter, “I think his answers were really misrepresentations. . . . Without cache memory for the processor, those 1.5 million Micro Channel machines will be unable to take advantage of applications that require higher bus bandwidth.” The cracks in IBM’s story were starting to get bigger.

A week later, Dataquest Research International held a PC Conference in Napa Valley, California. The scheduled topic was “Visions of the ’90s,” but the major subject of interest centered on EISA.

An article in Computer Reseller News on September 26 titled “Dataquest Conference Focuses On EISA” pointed out that almost all the conference attendees supported EISA. Ed Anderson, chief operating officer of ComputerLand Corp., said “The new EISA standard corrects (an issue) that has left the customer stranded in terms of investments” and that EISA was what Micro Channel Architecture should have been. MicroAge Computer Stores Inc. chairman Alan Hald called EISA “a natural extension of Compaq’s concept of continuity.”

Over the next two weeks, there was a lot of coverage in PC-industry publications, including Information Week, MIS Week, Computer Reseller News, Computer World, Computer & Software News, PC Week, and InfoWorld. Overall, the coverage was evenly mixed, with some articles strongly supporting EISA while others strongly supported Micro Channel and cast doubt on EISA’s survival.

A main argument against EISA was one IBM had started. By the time EISA machines actually arrived, IBM said, Micro Channel would already be entrenched. This theory overlooked the fact that almost none of the installed Micro Channel machines were suitable to run applications requiring an advanced bus, due to the lack of a memory cache for the processor.

Another argument came from an IBM engineer, who said that his team had looked at doing a bus like EISA, but rejected it because of inferior performance. In the past, IBM could get away with making a statement like that without being challenged. But we had built a strong reputation as a technology leader and could make credible counterarguments.

Interestingly, both PC Week and InfoWorld published editorials that cut to the bottom line. An editorial in the September 26 issue of PC Week titled “EISA: More than Just Another Bus” stated, “The new 32-bit bus design, EISA for short, deserves a thunderous round of applause . . . EISA is nothing less than a path not merely for survival, but prosperity. It’s the sanest, strongest answer to IBM’s pullout of the PC mainstream last year in favor of doing something proprietary (Micro Channel Architecture) and threatening anyone who tried to clone it.” The copy of the editorial that I have didn’t include the author’s name and I wasn’t able to find it.

William F. Zachmann, Vice President of Market Research for International Data Corp., wrote an editorial for the September 29 issue of InfoWorld titled “What’s So Confusing? EISA Extends Capabilities of Industry-Standard AT Bus.” In it, he addressed “the supposed confusion” introduced into the marketplace by EISA.

I nearly fell out of my chair when I got to the fourth paragraph. It said, “I’m not sure whether the folks writing this stuff simply are confused easily themselves, working for IBM’s public relations department, or just downright stupid. . . . However much the folks at IBM would like everyone to believe this sort of nonsense, the idea that EISA confuses matters is ridiculous. EISA is not a ‘third’ standard at all. It is simply a straightforward extension of the de facto industry-standard AT bus . . .”

Later, he goes on to say, “The supposed advantages of the MCA have been more a creation of the imagination of IBM’s public relations and advertising departments than of IBM’s engineering departments. The MCA has failed to demonstrate any real functional or performance advantages . . . From the start, the MCA was designed more with IBM’s interests than customers interests in mind. It is small wonder then that many users have rallied behind the slogan: ‘MCA—Just Say No!’ EISA merely strengthens the case for those who prefer to continue enjoying the benefits of intense competition among vendors that an open architecture makes possible.”

Although the battle was far from over, the momentum had definitely shifted in the direction of the open industry standard.

Zachmann got it exactly right: All of IBM’s spin and “FUD” (fear, uncertainty, and doubt) couldn’t confuse something as simple and straightforward as an extension to the existing industry standard. If there was any confusion created among customers, it was in those who were about to choose Micro Channel. Most were buying it because of the IBM brand. They may have paused to consider how much better it would be if they could continue buying machines compatible with their installed base. And anyone leaning toward industry-standard machines now had one more reason to stay with the standard.

Even with the first EISA-based machines a year away, we had effectively thrown a roadblock in front of IBM’s march toward proprietary control of the industry. Although the battle was far from over, the momentum had definitely shifted in the direction of the open industry standard.

Looking back at all this, I’m still amazed we had the guts to attempt it, and by how well it actually worked. There was definitely a lot of risk involved, but from our perspective, long term it was the least risky path of all. Even though we were experiencing record sales growth during the second half of 1987, I was convinced that IBM would eventually win if customers weren’t shown a better choice. And fortunately, our success hadn’t deluded me into believing we could win a one-on-one battle against IBM. I didn’t want an Alamo-like last stand.

But no one could have anticipated a solution like the one we came up with. Then, to minimize the overall risk, we identified each individual risk area and devised a plan to offset as much of the risk as possible. After that, it was simply a matter of excellent execution of each part of the plan. All the tens of millions of dollars and countless hours of work we invested in pulling this off were clearly worth it.

COMPAQ’S SALES CONTINUED TO BOOM throughout 1988, and we ended the year beating analysts’ expectations by delivering over $2.1 billion in sales. We continued to introduce new products throughout the year, including three new desktop models and our first entry into the laptop market. All made significant contributions to our revenues. We also stayed ahead of IBM as the performance leader in the 386 segment by introducing the Deskpro 386/25 on June 20, 1988. Although IBM announced a competing 25-megahertz, 386 product at about the same time, we were shipping our product in volume immediately while IBM didn’t begin shipping its products for about three months.

Plus, the Compaq Flex Architecture enabled us to significantly outperform IBM’s 25-megahertz 386 product, even though its machines included the Micro Channel.

In the same June 1988 announcement, we also introduced a lower-cost 386 model, our Deskpro 386s, the first computer to use the Intel 80386sx chip. We positioned this product to address a broader segment, the upper end of the 286 market, by pricing it lower. Sales took off quickly.

Just a week after the EISA announcement, on September 19, 1988, we held another press conference where we introduced the Deskpro 386/20e, a smaller, lower-priced version of the original Deskpro 386/20 launched a year earlier. This introduction expanded our 386 product line to include five different performance ranges at five different price points. The 386 market had been ours ever since we were the first to introduce a 386-based personal computer. This broad product line enabled us to maintain our leadership position in that important market segment.

Compaq entered the laptop segment in October 1988 with the SLT/286.

Then a month later, on October 17, we entered the laptop segment with the SLT/286, a 14-pound, battery-operated computer. Compaq had been criticized by analysts for not going into laptops earlier, projecting we would not be a factor in that market because of our late entry. We had answered these criticisms by pointing out the technology at that time did not allow a laptop to be “full-function,” a term we often used to describe our portable products. The SLT/286 was the first truly full-function laptop, delivering a high-performance 286 processor, high-resolution LCD display, and a hard drive. It even had a removable keyboard.

The product received great reviews and achieved the highest initial order rate of any product in Compaq’s history. It took nearly six months to begin to catch up with demand. By the time we did, the SLT/286 had become the best-selling laptop through the dealer channel. And we had stayed true to our fundamental product philosophy by waiting for the right time to enter a new market. Counter to the analysts’ predictions, we quickly became the leader in that market too.

IN JANUARY 1989 I became aware of a problem that was brewing with one of our largest authorized dealers, Businessland. It had always pushed us for bigger discounts than its competitors were getting, and it had been a struggle all along to keep a good relationship. We had been very careful to treat all our dealers fairly and consistently. But now I was told that Businessland was offering its sales personnel a $100 bonus over its normal commission for each IBM PS/2 they sold. We began getting reports from some loyal customers that when they went into a Businessland store and asked about a Compaq computer, the salesperson had disparaged Compaq and tried to sell them a PS/2. This was very troubling to me. I told Ross Cooley, our vice president of sales, to be very clear to Businessland’s management that we could not tolerate it continuing that practice. I asked that I be updated regularly on where things stood.

In early February we had a meeting to specifically focus on what to do about Businessland. I found out that it had continued its bonus program, and the “bait-and-switch” reports were getting worse. I also found out that Dave Norman, Businessland’s CEO, had recently told analysts that he planned to focus most of their technical support behind the IBM PS/2. We concluded that Businessland was putting extreme pressure on us to cave into its demands for special treatment, something we couldn’t do without creating major problems throughout the rest of our authorized dealer network.

The issue was what to do about it. It was clear to me that we would be better off without Businessland under these circumstances. It would be painful to terminate our relationship with them, but we would be better off in the long run if they refused to stop what they were doing. We decided to identify and work through all the issues associated with the termination process and prepare a letter of termination before sitting down with its representatives for a showdown.

McLoughlin drove Cooley, Swavely, and me back to the airport for our flight back to Houston. On the way we filled them in on the meeting. Cooley asked if there was a chance to reconcile, and I replied that there probably wasn’t. We had decided before we came that if we actually gave them the letter, we couldn’t go back. It would be very difficult to trust that they wouldn’t be constantly looking for some way to get even.

When we got off the plane in Houston, there were several phone messages waiting for us. Several executives from Businessland had called several different Compaq executives, saying there was a misunderstanding and that we needed to work it out. But there was no turning back.

Compaq represented about 15 percent of Businessland’s sales, while it accounted for about 7 percent of our sales. Both companies suffered only slightly short term. Our other dealers were so excited that we had stuck to our commitments to them that they became very loyal to Compaq. They quickly made up the difference and more.

It wasn’t quite the same story for Businessland. In pushing IBM over Compaq it was effectively betting on PS/2 and Micro Channel. We continued to take market share away from IBM and, after our new products launched in late 1989, Businessland began to lose market share as well. Then the recession of 1991 caught it in a weakened state and it almost had to file for bankruptcy. Businessland was bought by JWP, Inc. in August 1991 for a small percentage of its previous value.

With the Businessland matter behind me, I remained focused on completing the task of defeating Micro Channel throughout 1989. I knew we weren’t there yet. We still needed to deliver EISA-based products that clearly proved their compatibility and performance superiority over Micro Channel. Everything seemed to be progressing well until mid-March, when an unexpected problem popped up.