BPS Division Forums

[Dan Levinthal]

Dell: From Poster Boy to the Doghouse

Jay, here is my take on Dell.

Since Thucydides commented on the Peloponnesian Wars, historians have documented the rise and fall of empires. Business commentators seem to have identified a contemporary story of rise and fall emanating out of Round Rock Texas. In the 90s, owning shares in Dell stock was likened by Business Week to owning a lottery ticket with a gain from 1990 to 1997 of 20,000%. In contrast, the last few years have been a story of earnings disappointment and sagging stock price. How can we make sense of this saga of “rise and fall”? Does Kevin Rollins, the post Michael Dell CEO, lack the right stuff? Or, per a Greek tragedy, is this a failing ordained by the fates --- or the basic economics of competitive dynamics?

To understand the downfall, it is useful to recap the success story. Dell entered an already established personal computer industry with a different idea about how to compete. The well understood part of the success story is that, at the time of Dell’s entry, the IBM PC had become an industry standard. An upstart entrant didn’t have to develop an operating system and supporting software and establish credibility with consumers regarding the technical features of the machine. Rather, one could buy some chips from Intel, load on the Microsoft Operating system, and hold out the banner of IBM compatible. That’s the obvious part of the story, but if that had been the whole story, Michael Dell would not have graced the covers of leading business magazines and his company would have enjoyed the not so glamorous fate of Gateway or perhaps the even less glamorous fate of the unheralded “white-box” makers of commodity PCs who have entered and exited the industry and toiled away at miniscule margins.

Michael Dell started out in his University of Texas dorm-room customizing machines, adding memory and the like, to meet the particular needs of some corporate clients. This was the starting point for what has become the remarkable Dell “build-to-order” system. Build-to-order has implications for both production costs and the value proposition to consumers. First, on the cost side, in an industry with remarkable rates of price decline in key inputs (microprocessor and memory), by not holding inventory one is twice blessed. There is the standard “blessing” of conserving on working capital. However, there is also an industry-specific blessing that stems from the fact that buying components 12 weeks later than a competitor results in dramatic cost savings. An extra treat in terms of working capital is that customers pay Dell before Dell pays its vendors --- working capital becomes an earnings generator as the firm earns money on this float.

What about the value proposition side of this production model? Corporate customers, such as American Airlines or the New York Stock Exchange, like the fact that their supplier can generate large number of machines in a short period of time with customized hardware configurations and software installed. “Liking” in the language of business strategists translates into a high willingness-to-pay. Dell was not simply an efficient production machine, which is the standard sense-making of its success. Gateway sold build to order machines as well, but their target market were consumers whose computer needs could be satisfied with the usual small, medium, and large product line of retailing (or as reinvented by Starbucks, tall, grande, and venti) and as a result didn’t get the “liking” willingness to pay bang out of the build-to-order model that Dell did.

Dell marched through the land of corporate clients and captured a large share of that domain. Other warriors, such as HP and IBM, had committed relationships with retailers and resellers that made it hard for them to respond to the Dell “model”. Life was good in Round Rock and for Dell shareholders.

What happened? I offer two answers: one corresponding to the “fates” and the other to acts of man.

First, consider the fates. It seems to come as a continual surprise to the business press and stock analysts that, while the economic universe does continue to expand, that rate of overall expansion is quite modest and on the order of 3% per year. Individual sectors and individual firms can grow at rates vastly greater than that (all Google shareholders stop now for a moment and send a sizeable donation to a worthy cause), however they can only do so for a finite amount of time. Dell can do well in PCs, but its market share is bounded by 100% and PCs, after penetrating the set of potential customers, will experience a replacement and upgrade pattern of demand. Dell’s stock did extraordinary well over the years as the success of the build to order model and its ability to garner a large fraction of the market came as a (reoccurring) surprise to analysts. One legitimate negative surprise for the investment community has been that the build-to-order model appears to be less successful in Europe and in Emerging Economies than it had been in the U.S. context – although maybe it shouldn’t have been such a surprise that success in selling direct without intermediaries in a society of folks “bowling alone” was not going to be indicative of success abroad where direct supplier customer relationships appear to be more valued. However, the fact that the size of the economy is finite at any point in time and grows at a modest rate and that these facts have some bearing on the growth prospects on any given sector and, in turn, firm should not have come as a shock. But again, this basic property remarkably never seems to fail to surprise observers – even highly paid ones.

Now we come to the acts of man. Security analysts like companies that grow earnings. Company executives, in turn, impose mandates on their business unit heads regarding growth targets. This symbiosis leads firms to desperately seek out new growth opportunities and, often in that act of desperately seeking growth, to engage in initiatives that are inconsistent with the firm’s basic competitive logic. One place Dell went looking for growth was the printer business. It has also been seduced by the wonders of Plasma TVs and of course the hottest PC-like product in recent years – the iPod. While these products open up opportunities for growth, there is a fundamental problem in the logic of competitive advantage. Where is the build-to-order mass customization advantage in these products? When did anyone, corporate or consumer, need a printer or music player, with high levels of customization? Dell suffered this lesson earlier in its history when, at one point, it offered its computers through traditional discount retailers. While Dell may have the most efficient manufacturing and logistics system for making build-to-order products, it is not likely to have lower costs than some plant in China making consumer electronics products in fixed product lines of “tall”, “grande”, and “venti”. Dell may be able to bring the build-to-order magic to the market for servers and data storage devices, but their powers would seem modest in the land of standardized products. Yes, they could simply re-brand someone else’s Plasma TVs, but does the world really need another Amazon.com, particularly given the modest (and in most quarters negative) profitability of the original one?

Dell had a great run. It was and still is a remarkable company. Business success is not an annuity and the fact that rates of return may drift down closer to some broader economy norm does not make those profits and return to capital earned in the interim less special. A company with Dell’s extraordinary capabilities and strong market position need not, and should not, merely grow into middle age complacency with the maturing of the PC market. There remain opportunities for continued growth, but at the same time discipline is required in restricting one’s focus to those growth opportunities that, in a meaningful way (where in the prosaic world of business strategy, “meaningful” is taken to be something the drives a cost position or consumer willingness-to-pay in a non-trivial way), leverage the firm’s existing competitive position. The fates may constrain the duration and magnitude of super-normal profits, but acts of man can change both their timing and degree.

(Moderator's Note: to receive an email when this conversation continues, click on "Watch this topic for replies" at the left below.)

[Jay Barney]

Dan, how can I respond to any commentary on Dell that begins with a reference to Thucydides and the Peloponnesian Wars? Especially when I’m the guy who tries to explain the subtleties of resource-based theory by discussing the success of hypothetical hot dog stands.

But I will try anyway.

As I read your commentary on Dell, it boils down to two essential points—first, it is the nature of economic life that high performers ultimately become average performers and second, in the face of looming average performance, Dell’s managers may have engaged in diversification strategies that did not take advantage of their traditional strengths—customizing PCs for corporate customers. While presented as two separate points, it strikes me that the second is largely a result of the first. That is, if Dell was not facing the prospect of average performance, its managers probably would not have been tempted into selling iPods and plasma televisions—technologies that do not require much customization and that are not big sellers among Dell’s traditional corporate customers.

I certainly agree with your central thesis—it is true that for most firms most of the time exceptional performance almost inevitably deteriorates into average performance. But for me, this is not the really interesting question about Dell. To me, the interesting question is: “Why did this process take so long for Dell?”

Look first at the industry Dell has been operating in. PCs have been a commodity for over 20 years. The “replacement and upgrade” pattern you suggest is likely to emerge in PCs has probably already existed for most of these 20 years. In such a competitive market, it is quite remarkable that Dell was able to discover any comparative advantage at all, let alone one that enabled it to both increase the willingness of its customers to pay and reduce its costs.

And then, once Dell discovered, and over time perfected, this comparative advantage, strategy scholars and the business press studied the heck out of it. Cases have been taught, articles have been published, and books have been written all about Dell’s direct model. We all now know the story—direct customer contacts, low inventory, high customization, and so forth. You know it, I know it, HP knows it, Gateway knows it, IBM knew it. And yet, Dell’s advantage lasted for many years. How can this be?

There are, of course, at least two answers. First, maybe it was something about what Dell was doing. Second, maybe it was something about what HP and Gateway and IBM were not doing.

The first answer—“it was something about what Dell was doing” has received the most attention in the strategy literature. Both positioning and resource-based language can, and have been, applied in describing the first mover advantages enjoyed by Dell and why those first mover advantages were costly for others to imitate.

But, I have to tell you, I have never quite bought this explanation of Dell’s success. To me, the direct model never seemed all that complicated. OK—we take out parts inventory (that’s a good thing), we take out finished goods inventory (that’s a good thing), we outsource as much manufacturing as we can (that’s a good thing), we eliminate retailer margins, develop a pretty friendly website that provides the same, or better, service as customers would experience in a retail setting, and then we develop call centers that provide whatever additional service a customer might need (all good things). Which one of these attributes of the direct model is really that costly to imitate? Of course, it might take some time to get it as good as Dell, but this is not rocket science here.

To me, the bigger unanswered question about Dell is really about Dell’s competition. Why did it take HP, Gateway, IBM and other PC firms so long to react to Dell? Why didn’t these firms take those elements of the direct model that were applicable in their situation and imitate them? And if the direct model wasn’t going to work for these firms, why didn’t they develop their own comparative advantages? Why did it take so long to bring Dell “back to the pack?”

In your comments about Dell, you mention that other firms—like IBM and HP—had committed relationships with retailers and resellers that prevented them from responding to Dell’s model. This is probably correct. But my guess is that the costs of not responding to Dell—its large market share, its profitability, its competitors low profitability, the cost of the merger between HP and Compaq, and so forth—were probably much greater than the cost of abandoning commitments that constrained IBM and HP choices. In the decision between responding to Dell—and improving their performance—and not responding to Dell—in the face of continuing performance deterioration—why did these firms choose to maintain commitments that led to sustained competitive disadvantage?

Indeed, it strikes me that this question—Why do firms not change their failing strategies?—remains largely unanswered in the field of strategic management. Of course, agency theorists have developed their own explanations of these phenomena. But, in this context, I am not convinced by agency arguments. The institutional pressures that faced firms like HP, IBM, and Compaq were such that “managerial shirking” and “appropriating managerial perquisites” are just not convincing explanations for their inability to respond to Dell.

My own guess is that as firms go down a particular strategic path, the language, culture, and beliefs of that path make it very difficult to see or evaluate any alternative paths. If managers discover themselves “knee deep in the big muddy,” and cannot articulate a way to get unstuck, they continue to sink until they either disappear under the waves or, in the throes of death, finally throw off their blinders and discover an alternative. This psychological language, taken from work on the “escalation of commitment,” comes closer, I think, to describing why other PC firms took so long to respond to Dell than the economic language of “the cost of abandoning the prior commitments was too high.”

So, Dan, what do you think?