Back at You, IBM

THROUGH THE FALL OF 1983, Compaq’s management team had been weighing options to diversify our product line. A major factor in our success was deciding not to offer a product directly competitive with IBM’s desktop PCs. The fit of our portables beside IBM’s desktops was perfect for both dealers and customers, because the issue of their loyalty to IBM had effectively been avoided. In many ways, we believed that the safer path was to continue to be a “Portable Company” and offer a wider range of portable products.

To offer a desktop competitor to IBM would be a major change, and one with unpredictable consequences. There was risk that a desktop might be viewed as a late-to-market, “me too” product and not taken seriously. On the other hand, we had also established a reputation for ruggedness, compatibility, quality, and reliability, all traits that were important in desktop PCs.

Still, it was difficult for us to make the decision to leave the safe haven we had enjoyed.

One thing was very clear: Any desktop product we introduced would have to be significantly different from IBM’s. It could start with all the key differentiating characteristics from our portables, but it would have to go well beyond those and deliver additional real, useful, and innovative features without sacrificing any software compatibility. And we would have to deliver all that at a competitive price.

The fourth quarter of 1983 had been extremely busy. We launched the Compaq Plus, settled the lawsuit with TI, and completed our IPO by early December. Valuable time was slipping away, so we made the critical decision in mid-December to start a new desktop project.

Once that decision was made, getting to market as soon as possible became a top priority. As the scheduling process began, we immediately faced an unexpected problem. We wanted to position our desktop against the XT, which meant we needed to announce it before IBM came out with a product based on Intel’s new 80286 chip. Plus, we felt strongly we should not announce a new product during the summer months, when much of the press and analyst community would be on vacation. If we missed June, we would need to wait until September, adding the serious risk that IBM might announce its new product before ours. To make a late-June announcement meant we had only six months from start to finish, which seemed unreasonable for even a highly skilled, highly motivated team. After all, the Portable had been completed in record time—just eleven months.

I knew one way to speed things up was to isolate the development team as much as possible. We immediately formed a separate division led by three of our top executives: Kevin Ellington, division vice president; Mike Swavely, marketing vice president; and Gary Stimac, engineering vice president. The project was code named “Bullet” because we had to move “faster than a speeding bullet” to complete it on time. Office space at a remote location was rented for work to begin. Everything had to be done in parallel, so the marketing and manufacturing teams began their efforts simultaneously.

As 1984 began and rumors of an impending portable from IBM froze orders for Compaq Portables, what might have been shaky confidence in our plan at first soon turned into a certainty that we had made the right decision. IBM—not Compaq—had been the first to break the unspoken truce and invade our territory. Whether our desktop would be successful was still uncertain, but fear of upsetting the delicate balance was clearly no longer an issue. We were now fully committed to entering the desktop market, and in a way that leveraged the reputation we had worked so hard to achieve.

If we had not been so committed to achieving a high degree of differentiation from the IBM PC, it would have been tempting to design the Bullet product as simply a repackaging of the Compaq Portable into a desktop. That alone would have been difficult to accomplish in just six months. But we knew the likelihood of success was low if that was all we did.

Our team scoured through every aspect of the desktop PC looking for ways to make our product better without sacrificing compatibility. That was the real rub. In the last year, most of the established computer companies had introduced their own PCs that had real advantages over the IBM PC. In every case, they had sacrificed compatibility with the largest base of PC software in the world and, as a result, had failed to achieve any real success. Compaq alone had found the holy-grail technology of being able to add innovative features without sacrificing complete compatibility. That was our ace in the hole.

While the March winds and April showers came and went, the Bullet team hardly had time to notice. The thoroughness of our efforts to find areas of real improvement paid off with an impressive set of features. The Bullet would have the same dual display capability we invented for the original Portable. It would also have the same triple-shock-mounted hard-disk drives, providing ruggedness not found in any other desktop PC, along with the performance advantage of the 8086 microprocessor.

We also solved other important issues not addressed by IBM. As an example, the 10-megabyte hard drive needed a more efficient method of backing up an end user’s important data than that provided by floppy disks. To back up all ten megabytes required thirty floppies and took so much time to accomplish it wasn’t done very often. We addressed this issue by offering a backup option with a 10-megabyte tape cartridge, so the entire disk could be backed up on one tape in a reasonable amount of time. This kind of commonsense solution to a very real problem was exactly what we were looking for.

So were our solutions to storage expansion and plug-in board expansion. By using half-height storage devices, customers would be able to mix and match floppy disk drives, Winchester disk drives, and tape backup drives in any combination that made sense to them. By using the new 256K RAM (Random Access Memory) chips, we were able to get all 640K bytes of addressable RAM on the system board instead of taking up an expansion slot as the IBM PC and XT did. The result was four available expansion board slots in the maximum standard configuration, versus only one in the XT when comparably configured.

The Compaq Deskpro was designed to meet user needs and perform significantly faster than the IBM PC and XT.

Indeed, Compaq’s dealers carried the Deskpro, even though shelf space was about to get much tighter. Only six weeks after our launch, IBM announced the AT, the first PC to use Intel’s new 80286 microprocessor. The AT surpassed the Deskpro in performance, although not by much, but its price was much higher.

Six weeks after our launch of the Deskpro, IBM announced the AT, the first PC to use Intel’s new 80286 microprocessor.

Still, IBM regained the performance leadership position with the AT, making our claim on that title short-lived. That was a negative for us, but the AT’s high price point meant it would have limited real impact on Deskpro sales.

In reality, the AT announcement was positive for us because it moved the industry standard forward to the next processor level and paved the way for our future products.

The decision to rush the Deskpro to market ahead of the AT turned out to be right on target. Had we not entered the desktop market, our reputation would have remained limited to portables and we wouldn’t have been able to develop a broader market leadership position. Even though the Deskpro was announced only six weeks before the AT, that gave us enough time to get our product successfully positioned against the XT and gain wide, favorable media coverage.

The Compaq Deskpro gained widely favorable press coverage.

Had the Deskpro launch missed June and occurred after the AT announcement, it would have been compared unfavorably with the AT in every area except price. It would have been perceived as a lower-priced alternative, a position we definitely did not want. In the end, our timing worked well and almost no one realized how close we had come to missing the window of opportunity.

Shortly after the AT reached the market, another positive development for Compaq emerged. The AT had major incompatibilities with the existing software base, much more so than the XT. In spite of our having repeatedly broadcast the problems created by a lack of backward compatibility, IBM still paid no attention to achieving it. Perhaps it was a matter of not knowing how to accomplish backward compatibility, since the company had never done that before. A more likely explanation was simply that IBM still held the belief that everyone would move to where it was, because it didn’t need to pay attention to such details. That had always been the case. Likely, that would have continued, if we had not developed the technology to deliver backward compatibility across all our products and successfully trumpeted the advantages of doing so.

While IBM’s XT had a relatively small degree of incompatibility, the AT’s much higher degree became a problem noticed by almost all AT customers. IBM might have assumed that we would not be able to solve the incompatibilities of the 80286 and any issues would blow over. But it underestimated our capability and drive.

As 1984 CAME TO A CLOSE, we breathed a collective sigh of relief. It had been an incredibly challenging year, starting in the beginning of January with the sharp drop in sales of our portable ahead of the impending mid-February announcement of the IBM Portable. At about the same time, the Apple Macintosh entered the market. Then there was the extreme pressure to get the Deskpro out by the end of June, followed by the mid-August announcement of the IBM AT and midyear price cuts on its PC and XT.

The price cuts could have been a big problem for us, and most observers predicted they would lead to our downfall. These were the same naysayers who had predicted our demise due to the IBM Portable and later the AT. This was our first real pricing challenge, so we didn’t have any previous experience to guide us.

We believed that our upscale positioning would protect us somewhat and decided to make only token price cuts at first to test the strength of our positioning. The clear differentiation of the Compaq Portable, Plus, and Deskpro had the effect we hoped for and our sales continued to climb.

In spite of all these challenges, we achieved another record financial performance in 1984. Our sales tripled to $329 million and our profit margins improved significantly. We had established ourselves as a major player in the desktop market and strengthened our position as the leading provider of portable computers. The new factory building we had started in February came online in December, significantly increasing our manufacturing capacity.

We completed our new factory in December 1984, just 11 months after breaking ground.

We also got off to a great start in the European market with the establishment of subsidiaries and product launches in the UK, France, and Germany. Eckhard Pfeiffer had established Compaq’s European headquarters in Munich and recruited Joe McNally to head the UK and Bernard Maniglier to head France. The European market was lagging the American one, so our timing turned out to be excellent.

But even before 1984 was over, we had to deal with the challenge the IBM AT represented. The AT’s significant lack of backward compatibility presented us with an opportunity to demonstrate our compatibility technology, but we were going to have to create a lot more differentiation than that if we wanted to continue on the path to market leadership.