The Tragedy of Compaq
The Compaq Computer Company's early years of absolutely insane growth remain the stuff of legends. Founded in 1982. First year revenue? $111 million. 0 to $111 million. One year. IPO, December 1983. And year 2, 1984? $329 million revenue, 200% growth. Year 3, $504 million, 53% growth and the Fortune 500. Later, Compaq hit $1.2 billion in revenue for 1987, the fastest ever in history.
Along the way, Compaq led an insurgency of IBM PC clone-makers against Big Blue, overwhelming the old lion and unlocking the PC standard for a new generation of PC-makers. That is when the problems began. The tragedy of Compaq is that they led the revolution. And then as it so often happens,
the revolution turned on them. In this video, we take a look at the fall of Compaq. ## Engine: Product Compaq in its early years had several growth engines. First, they offered a breakthrough product that was both compatible with - yet sufficiently differentiated from - the original IBM PC.
You can't exactly work with the Compaq Portable sitting on your lap - though I would like to see you try - but you can lug it around from one place to another in a shopping cart. The original PC cannot do that, and IBM did not bother to introduce their own portable until February 1984, over a year later. Add the cheaper price, and you can see why the Portable was a suitable second computer for working professionals.
In addition, Compaq went above and beyond in producing a computer compatible with IBM's formidable ecosystem of business software. This includes its flagship productivity spreadsheet software Lotus 1-2-3. Such compatibility not only required the use of a cleanroom - employing two teams working on re-implementing the IBM PC BIOS code - but also fidelity to the IBM PC's hardware, right down to the keyboard layout. ## Engine: Speed Number two. IBM's product development cycle was slow and deliberate. Development does a prototype. Manufacturing reviews it for mass production and gives feedback. After that, sales reviews it and makes a plan for management. Very orderly,
taking about 12 to 18 months in all. So it took some time for IBM to add new features and technologies to their PCs. Compaq's product development cycle on the other hand sets loose all of their departments at the same time. Messy and chaotic, but taking just 6 to 9 months from end to end. This let them bring new technologies to the market faster than IBM can. The classic example is the iconic Deskpro 386, the first IBM PC clone to be equipped with leading edge silicon. It hit the market a full year before IBM had something similar. They were also faster than anyone else - even IBM - to fill the shelves with product. IBM’s early struggles to keep up with demand meant frequent stock-outs.
Compaq presented a compelling alternative. The best ability is availability, as they say. ## Engine: Distribution The third major engine was the company’s relationship with its dealers. IBM had long sold through its own sales force. But with the IBM PC,
IBM started selling through retailers like Sears, BusinessLand and ComputerLand. However, conflicts arose. Those retailers did not appreciate that their partner was competing with them. Early on, Compaq differed in that they made the retailers their only distribution channel. Cofounder and founding CEO Rod Canion liked to say:
"Keep the seller sold" and "A dealer done right can be the best distribution mechanism". It helped them save money by not having a big and expensive sales force. It also engendered dealer loyalty. A salesman at CPU Computer Centers said in an interview at the time: > I view that as a real big plus for me. I can sell a corporation a single IBM, maybe two. If they decide they want 100 of them, they go to IBM and get a 30% discount and I lose a big account. > But for Compaq they have to come to me.
This was an early advantage that helped the company achieve its fabled growth. But as we will get to later, it also eventually became their fatal weakness. ## The High After successfully launching the Deskpro 386 and leading a revolt of PC clone makers against IBM, Compaq seemed to be firing on all cylinders. Everyone admired their instant success. The business media fawned over their ability to pick the right products at the right time. Magazines like BusinessWeek wrote praise-filled profiles about Compaq’s seemingly unique style of business.
In 1990, Rod Canion sat for an interview with Harvard Business Review. Harvard heaped praise on the company's consensus-oriented culture and seemingly-contradictory style of doing business. > But what is most intriguing about Compaq is a series of counterintuitive notions that combine to create the company’s management process ... > In an industry that is driven by innovation, Compaq defines innovation as staying within the boundaries of accepted industry standards ...
> Because the company’s labor costs are so low, says the CEO, the company must hire its people very carefully ... > In a company that has achieved such remarkable financial results, cost ranks relatively low on the list of manufacturing priorities. ## Attack of the Clones The issue with these statements - the last one in particular - was that price did end up being very significant. If you recall from my video on Hard Disk Drives or HDDs, the reason why nobody made money in making and selling HDDs had to do with their modularity. There are four major components in a HDD. Each is precisely engineered, but they are
all largely independent of one another. So if someone improved a particular aspect of one component, they can quickly slot it into the final product and bring it to the market. The same with the IBM PC clone ecosystem. Fundamentally speaking, a PC is a modular box with a keyboard and monitor - powered by an Intel chip sitting on an IBM-derived hardware architecture, running Microsoft OS software. In the early days of the PC clone industry, it took careful engineering to put these components together and make them work. But by mid-1985, the ecosystem’s quality had improved to such an
extent that you can throw them together and they will work together with relatively few issues. Thus in the end, just two things mattered to a computer's perceived performance: Microsoft's OS software and Intel's silicon. This became increasingly obvious with the Deskpro 386. That iconic computer inspired a generation of market entrants from the United States and East Asia, in particular. It was easy for
them to get a 386 and MS-DOS and introduce their own high-powered PC "super-clone". Take the Fall 1986 edition of the Computer Dealers Exhibition or Comdex show. Held in Las Vegas, Comdex had once been amongst the biggest trade shows in the world. The Deskpro 386 release triggered a multitude of 386-equipped PC clones at that Comdex show. Many were just prototypes with mass availability in 1987 - implying a rush to the market.
These other super-clones from Dell, Gateway and such were just as powerful as Compaq's products, and cheaper too. The price difference might be because the computers themselves were made with cheaper materials like plastic or assembled using offshore labor. Or it could also be the business model. Companies like Dell or CompuAdd cut out dealer
markups by directly selling through telephone or mail-order - like how Sears once sold stuff. Dell in particular often ran big advertising campaigns at consumers and small businesses, comparing their offerings - and prices - to Compaq's. Taking no prisoners. Consumers consistently paid a 67% premium for a Compaq compared to a similar Gateway 2000 system. Yet Compaq was slow to respond to this, or even acknowledge it was happening.
In December 1990, Michael Swavely, President of Compaq North America and Canion's second-in-command, told reporters: "You can charge a premium price if you build the best machines." Just a few months later, Swavely retired from his position as president. ## New Innovations Was Compaq really making the best machines? They were certainly fast to integrate new things into their computers. The issue however was that they produced few of those new things themselves. In the late 1980s, they invested only about 4% of their revenues into R&D.
So they found themselves beholden to critical partners for their own progress. In one notable event, Rod Canion writes in his memoir "Open" about confronting Intel's paranoid mastermind Andy Grove in late 1989. Canion needed Intel to finish their 486 chip. But Intel was instead pulling their best people to prioritize RISC development to fight Sun Microsystems and their SPARC chip. In that story, Grove eventually decides to shift back to the 486 and the Pentium after that. But it illustrates how Compaq was dead in the water without its partners. But beyond that, Compaq also missed opportunities to open and corner new markets. In early 1984,
Rod Canion badly wanted Compaq to do a laptop - apparently because he wanted to work on one. But Compaq never produced something without first doing market research. So they did that, and a young market researcher showed that though a market existed, it was then not that big. Canion accepted this and set it aside. He later cited this incident in the 1990 Harvard Business Review interview as an example of their pursuit of informational truth.
Then in 1986, Toshiba released the Toshiba T3100. It was not quite a true laptop because it needed an external power source, but it did have that laptop "look" and sold quite well. Despite what seems like an obvious shift in market demand, Compaq took until October 1988 to introduce a laptop of their own - the SLT. A fine laptop. But it was starting to feel as if Compaq - built to take on IBM - was ill equipped to keep up with this rapidly changing business environment.
## Europe Yet despite the intensifying competition and lack of self-generated innovation, Compaq continued to grow revenues and profits. While other computer companies like Apple Computer and Sun Microsystems struggled to resist the attack of the clones ... in the three years from 1986 to 1989, Compaq's revenues multiplied five times over to $3 billion. The secret behind that growth was the company's European operations. Those were run by a former Texas Instruments executive from Germany named Eckhard Pfeiffer.
Pfeiffer first joined Compaq back in 1983 to build out the fledgling company's European operations. With little guidance from the American HQ - their attitude was basically "if it happens it happens" - he transplanted the dealer-only structure across the UK and Europe. Compaq ignited huge growth in what was a largely untapped market dominated by IBM, Apple and Olivetti. The British subsidiary turned a profit in its first month. He then opened a swath of new subsidiaries all across Europe and led the construction of a new Scotland factory to keep costs low.
By 1990, over half of Compaq's revenues came from abroad - mostly Europe. Pfeiffer was the company's second highest paid executive after CEO Rod Canion himself. ## 1991 Unfortunately, things start going downhill in the middle of 1990.
In August that year, the First Gulf War breaks out, causing oil shocks that roll through the rest of the economy. There is a mini-recession. Various Information Systems departments around the country had to grapple with lower company budgets. This price-consciousness accelerates the clones' growth and forces Compaq to cut hardware prices. But more worrying than the macro stuff was a growing divide between how the company's management and its other stakeholders perceived the landscape surrounding Compaq. Chairman Ben Rosen, one of the original investors, grew concerned about Compaq's ability to keep up with the competition. In an oral history for the Computer History Museum he said: > Well, those two decades, the ‘80s and the ‘90s, are massive changes of the industry, and in both cases, our management didn’t move as fast as the industry.
> We started moving faster than the industry by creating it. But then there are a lot of other smart people ... and good companies in the world, and they went after us, and we didn’t move fast enough. In May 1991, Compaq pre-announced that their second quarter 1991 profits would fall 81% from the prior year. The stock clanked 27%, losing $1.2 billion of market value. But company management disagreed with Rosen and the market's assessments.
They argued "short-term perturbations", insisting that sales will come back when the economy picked up again, and that the 27% crash in the stock price was an overreaction. In the actual 2nd quarter 1991 earnings conference, management said that the next quarter might be even worse. The company braced for big changes ahead. At the end of summer 1991, Rosen demanded that Compaq cut headcount and bring out a low-end PC as soon as possible. Canion agreed to the former, but said that it would take a year to do the latter.
In October 1991, the company reported a quarterly loss - its first in years - and that they would be laying off 12% of its work force, about 1,400 people. A few days later, the board of directors held a major meeting. After a brutal 15-hour back-and-forth, Rosen dropped his ultimatum: A low-end Compaq PC in three months. Rod Canion replied that it was impossible to make a PC in that amount of time without sacrificing the company's renowned quality control and engineering excellence. Rosen then said, "Yes you can". Unbeknownst to Canion,
Rosen had secretly sent two Compaq guys to the Comdex show. They bought several off-the-shelf parts from vendors and then assembled two demo PCs in their motel room in three days. Rosen then had those demo PCs brought out, saying that he did this with just two people in three days. But Canion again refused to bend. That sealed it. Rosen offered him a seat on the board to save face. Canion refused and Rosen fired him. Eckhard Pfeiffer stepped up to the CEO position. In an interview with the New York Times shortly afterwards, Rod admitted that the management team probably recognized things later than they should have, but added that they were now moving quickly. So why did Rosen replace him? Canion said:
> I’m still not completely clear on what went on in his mind. Changes were being made, but it had more to do with looking to the future. He was looking for a single leader who was going to make the tough decisions and felt that called for someone more like Eckhard than like me.
> In some ways, it seems like a very severe move for not a very big disagreement Much later, Rod reflects in "Open", his corporate memoirs: > In truth I was somewhat burned out by the intense, nonstop pace of those ten years, especially during 1991. That year we were caught off guard by an economic recession, by six of our top ten dealers merging into three, and by an unexpected strengthening of the dollar ... Laying off hundreds of dedicated people weighed very heavily on me. ## New Changes Though Compaq was already turning towards the low-end when Pfeiffer took over, the German was seen by outsiders as having the intensity, drive, and the execution skill to take on the clones. He was less focused on the computers' technical aspects, and more on hitting the company's numbers. The strategy he unveiled to all stakeholders was simple: Make millions of low-cost PCs, and turn it into a volume game. A few months after becoming CEO, Pfeiffer launches a price war against his fellow clone-makers like Dell and Gateway. First,
he cuts prices across the board on every product Compaq was then selling to the public. Then in June 1992, he unveils 16 new products - including the low-cost ProLinea desktop and Contura laptop lines. Such plans began under Canion, but Pfeiffer fast-tracked their development by pitting Compaq's internal component divisions against external suppliers. Now, the cheapest 386 Compaq was 50% cheaper in 1992 as compared to 1991. The ProLinea desktops quickly became Compaq's best-selling lines. In addition, Pfeiffer expands distribution - signing deals with large computer superstores focused on the home consumer.
Compaq also grows their 1992 R&D spend to $173 million. Though as a percentage of total revenue, this was about the same historically, the absolute number was higher than virtually every other clone maker. Compaq's ads liked to crow that "at most other computer companies, R&D stands for 'Replicate and Duplicate'." And to help service large corporate clients with increasingly complicated systems needs, Compaq strikes a deal with system integrators like EDS. Though for the
most part, the company continues to rely heavily on PC distributors and retailers. ## Turnaround Pfeiffer oversees a remarkably fast turnaround for Compaq. Thanks to the broad price cuts and new products, Compaq sells 150% more computers in 1992 than 1991. That year, revenues shoot up to $4.1 billion, up 25% from the prior year. Net profits grow to $213 million, up 63%. The turnaround continues in 1993. As Pfeiffer predicted,
the massive volume helped find new efficiencies and cut costs. Their factories in Houston, Singapore, and Erskine, Scotland have to ramp up to 7 days a week to keep up with demand. A new factory is established in the city of Shenzhen in the People's Republic of China. In August 1993, Compaq introduces the Presario, a line of computers for the home and home office. It was an all-in-one machine with a monitor, sound card, CD-ROM drive, modem and software tailored for the computer novice. Just plug it in and turn it on. The Presario becomes a massive hit,
Compaq's new best-selling line. I think my parents had one in my home growing up. In a year when Apple had to take a $320 million restructuring charge, Dell lost $75 million and IBM broke even, Compaq went from $4 billion revenue in 1992 to $7 billion in 1993. Profits hit $867 million. In early 1994, Fortune wrote: > "In the two years since Eckhart Pfeiffer became CEO of Compaq Computer, he has engineered such a stunningly complete turnaround that it's surprising that the company still has the same name" Even the competition had to agree that Compaq pulled a rabbit out of its hat. Dell's Vice President of Northern Europe Bruce Sinclair said: > "What Compaq has done is going to go down in history as one of the great turnarounds of all time ... we can both increase our market share at the expense of others" Bruce was not wrong. The 1990s would be a massive run for the PC, and there was plenty of room for both Dell and Compaq to grow.
## The Boom Years From 1990 to 1999, the world PC industry as a whole grew an average of 18% each year. From 24.2 million PCs shipped in 1990 to a staggering 113.5 million in 1999. Much of this growth came from consumer sales. Between 1990 and 1997, the percentage of US households owning a computer increased from 15% to 35%. The amount of time people spent on computers tripled. Pfeiffer turns Compaq in a mass-market volume PC maker - selling PCs at all price points and functionalities from $25,000 servers to sub-$1000 laptops. In 1994, Compaq does $10.9 billion in sales. The year after that, 1995,
they grow to $14.9 billion, leapfrogging IBM and Apple to become the world's largest PC maker. Compaq and its CEO Pfeiffer seemed to be at the top of the world. ## A Vision and a Problem This victory in 1995 can perhaps be seen as a turning point. It marks the end of easy growth in the home and home office PC market. Afterwards, competition once again begins to intensify - particularly from Dell. Up until 1996, Dell's growth came at the expense of other PC makers. But
now their notebooks and servers were starting to eat away at Compaq too. Compaq had increasingly few options to respond. One thing that remained consistent throughout all of its history was a dependency on external distributors and retailers. This network helped Compaq sell PCs in over 100 countries. But Dell's model of selling directly to the consumer was starting to gain ground. The Internet now made it easier than
ever to cut out distributor margins and pass the savings to customers. Yet Compaq cannot easily adopt that direct-selling model because it would piss off their existing distributor partners. Such partners might start pushing Compaq's rivals to their customers instead.
In March 1996, Compaq warned investors that it might not hit its first quarter earnings numbers. The stock takes a dive, forcing Compaq to radically cut prices again. It works and the stock recovers, but sharp-eyed viewers note that margins have now fallen to a razor-thin 20%. Pfeiffer comes to believe that Compaq's only way forward was to grow beyond mere PCs to being a global computer company - selling services like IBM did. With this push, he resolves to make Compaq into a $50 billion company by the year 2000.
## A Series of Unfortunate Acquisitions It goes horribly wrong. Compaq embarks on a series of unfortunate acquisitions. Substantial money was spent to build a networking business to challenge Cisco. In late 1995 they bought Networth, a company making networking gear. And then in May 1997, buying another networking company called Microcom.
Then in June 1997 they paid $4 billion to acquire Tandem Computer Incorporated. Tandem's speciality was in very fault-tolerant computers for companies that needed to provide rock-solid 24/7 service. So think stuff like ATM networks, process control and the like. Their computers had many redundant parts in them, which made them tricky to program. Less than a year later, Pfeiffer shocks the tech industry with his next acquisition. In
late January 1998, Compaq pays a record $9.6 billion for the iconic Digital Equipment Corporation or DEC. It was the biggest computer company acquisition to date. DEC has been around since the late 1950s. They are perhaps most famous for inventing the minicomputer category with the PDP-11 and the VAX 11/780. Their technology and R&D prowess were legendary - taking IBM head on.
Buying DEC and its $13 billion in revenues turned Compaq into a $38 billion giant. Investors and analysts mostly approved of the deal. They all knew that DEC was available for a reason. Their core high-end Internet server and workstation businesses were struggling to adapt
to disruptive competition from cheap and powerful Unix computers like those from Sun Microsystems. Yet analysts pointed at the upsides. DEC had $2 billion of cash on hand and $3 billion in tax loss assets, which can help offset the purchase price. DEC's worldwide services businesses and customer relationships still made significant money. And the company retained considerable technical assets.
DEC had their own flavor of Unix as well as an in-house 64-bit RISC microprocessor called Alpha. It was fairly fast, but the company never did anything with it. They phased it out for x86 architecture. And interestingly enough, DEC owned AltaVista, one of the major search engines in the early Dotcom era. Other than a vague notion to sell Compaq PCs through the site, they did little with it and ended up selling that to KPMG. Pfeiffer struggled to bridge the two cultures. Outsiders found it difficult to explain why.
There was little push to capitalize on DEC's considerable technology assets, which still kind of sat around waiting for someone to use them. And all the while, Compaq's core PC business continued to deteriorate. It was not until 1998 that Compaq finally began adopting Dell's direct-sales model. Dealers complained but there was no other way to close Dell's 10-15% cost advantage. But it was too little, too late.
## End of Pfeiffer In April 1999, Ben Rosen and the board fired the once-high flying Pfeiffer. Like with Canion's firing, Rosen's reasoning was that Pfeiffer had lost his touch. Too driven on hitting a revenue number, his expensive acquisition spree, and worse yet, inability to integrate said acquisitions ended his tenure. Rosen later said he wished he did it a year earlier. Pfeiffer for his
part complained to the press that the board approved of his strategy and felt that more openness on their part was necessary. After a three month search, the board chose another Compaq insider to run the company - Michael Capellas, their Chief Information Officer. Analysts were glad that someone who knew IT was back in the head spot, but overall it was not an inspiring choice. The first time CEO got to turning things around, but after a few months the results were mixed. Expenses fell and gross margins improved, but finances remained lackluster. The direct selling and build-to-order operations were gaining traction but the core PC business kept melting away.
In September 1999, Dell overtook Compaq as the leading PC seller in the United States. ## Hewlett-Packard It is the year 2000 and the iconic electronics giant Hewlett-Packard was looking to remake itself. Famously founded in 1939 in a garage in Palo Alto, HP was then one of the largest companies in the world, 13th on the Fortune 500, with $48 billion in revenue. They began in electronics like their iconic calculators and the titanic cash cow printer business. But in the early 1990s, they got themselves into the PC business. It grew well, but when that business commoditized in the late 1990s, it hit HP the same way it hit Compaq.
After running the company for 7 years, CEO Lew Platt decided to step down in 1999. Platt was known for preserving the company's unique culture - the HP Way. But one can also argue that Platt focused too much on the "other stuff" to the exclusion of actual performance. As a former executive said, "Your business could be going to hell, but it was like - well, you did very well on worker safety." Platt had also been concerned that the business had grown too far too fast on the back of the increasingly unstable commodity PC business. It needed to develop a new,
profitable growth business. Maybe something in the exploding Internet industry? Or services. The talk of the technology town then was how Lou Gerstner turned around IBM by moving it into IT services. Whatever it should be, HP had to change. To help lead this change, Platt and the HP board chose as CEO a fast-rising, hard-charging sales executive from Lucent Technologies named Carly Fiorina. ## The Merger Fiorina's mandate was to be an agent of change - transforming HP and bringing it into new businesses like information system services.
In September 2000, she made a $18 billion acquisition bid for PriceWaterhouseCoopers' consulting arm. Adding PWC's 31,000 information system consultants made a lot of sense. But investors hated the deal and it fell apart. For Compaq, 2000 had been a bad year. The busting Dotcom bubble not only devastated DEC's high end internet server businesses, but also unleashed a flood of excess PCs into the market. The board asked Capellas to explore a possible merger. Fiorina and Capellas first met in mid-2000 at a e-marketplace consortium meeting for selling computers online. Both being first-time CEOs trying
to turn around rigid commodity PC makers, they bonded immediately. A few months later, Fiorina calls Capellas about Compaq licensing HP's variant of Unix. Capellas quickly asks if HP wants to buy Compaq out instead. Thus on September 3rd 2001, Labor Day, HP agreed to acquire Compaq in a stock swap worth about $25 billion. Carly Fiorina would continue running the business, which boasted combined revenues of $87 billion - not far off from IBM's $90 billion. Infamously, though, the acquisition was a dumpster fire. On pure merit, it wasn't a bad idea. The two companies had product lines
that complemented one another. Only 15% of HP’s PCs were shipped direct to customers, so they can benefit from Compaq's growing direct sales programs. There were also about $4 billion of savings to be had, mostly by slashing 15,000 jobs. But investors and competitors were not impressed.
The two companies lost $13 billion in market capitalization after announcing the deal. One institutional investor said it was like "taking two stones and tying them together to see if they float". Sun Microsystem's President Edward Zander said, "When two sick companies combine, I'm not sure what you get." Michael Dell said, "Mergers of this size are very hard to do.
The opportunity it presents to us ... that's pretty compelling" And that was all before the acquisition deteriorated into a personal battle between Fiorina and Walter Hewlett, son of HP cofounder Bill Hewlett. Walter insisted that HP was paying too much and the company stick to its knitting. I shall save the battle between the two for some other video. In the end, after a frantic
cross-country campaign reminiscent of a national election, Fiorina won shareholder approval for the merger by the thinnest of margins and quickly got to work integrating the two companies. For all the chaos, the deal ended up doing well - for the shareholders, at least - despite claims to the contrary. Fiorina was not there for that - the Board dropped her as the CEO back in 2005. But her successor, the late Mark Hurd, cut $3.5 billion of costs, laid off 15,000 jobs, clarified management roles, and grew the combined company to $110+ billion in revenues over the next five years. Soon after the acquisition, HP moved to phase out the Compaq brand. But it still remained
on a few low-end computers until 2013. ## Conclusion We have largely forgotten the role that Compaq played in leading the fight against IBM. The story just skips from IBM to Microsoft and Intel. This is in part due to the massive commoditization of the PC industry in the late 1990s. It was vastly destructive,
not just to the American PC-maker companies but American manufacturing in general. In some ways, that move feels inevitable, but I do wonder if there was anything that Intel and Microsoft could have done to prevent it. They lorded over a fiefdom and squeezed the ecosystem for value and profits. Maybe they shouldn't have.
2024-11-13 11:31