Mad Dog 21/21: The Innovators’ Droll Lemma
March 23, 2015 Hesh Wiener
Robert Samuelson got it right, and memorably so. Last June, the Washington Post columnist characterized a spat in print between two Harvard professors as “an intellectual food fight.” The fight is over the value of business professor Clayton Christensen‘s theory of disruptive innovation, and Christensen’s adversary is history and literature professor Jill Lepore. Christensen’s theory, right, wrong, or some of each, may help explain IBM’s perplexing situation, and arguments about it are undoubtedly underway at the top of IBM. And if they aren’t they ought to be.
In Christensen’s 1997 book, The Innovator’s Dilemma, he asserts that there are two significant types of change in the behavior of a business. One is “sustaining innovation,” his term for the process of improving a company’s existing products and services. The other is “disruptive innovation,” Christensen’s term for the development of a new product that threatens an established one by entering the market at a lower performance level but also at a lower and therefore more attractive price.
Paradoxically, Christensen argues, the companies that are knocked down and sometimes out by disruptive innovation are the industry leaders who have worked hard on sustaining innovations. The leaders who will be deposed by upstarts have listened to their customers and worked hard to fulfill these customers’ expectations for more and better of a particular product or class of products. These threatened and sometimes doomed companies are doing the right thing by their customers, but they can catastrophically fail when a disruptive rival shows the customer a less costly (even if less capable) alternative. Christensen’s classic example is the hard disk drive business. Small and relatively inexpensive drives originally made for use in PCs eventually supplanted larger, faster but more costly devices. IBM was hit hard by this. It lost sales of high-end disks to RAID products based on disruptive small drives. It ultimately reinvented its storage business to incorporate products that were not what IBM’s enterprise customers formerly said they wanted. IBM had to learn to make and market disk arrays to compete with the stunningly successful RAID offerings of EMC, among others. Today, IBM tries to be a leader not a follower in storage technology, but it never fully recovered by the battle in which its sustaining technologies were vanquished by EMC’s disruptive technologies. Moreover, the core electromechanical technologies in the trade, spinning disk drives, keep improving as the industry, which no longer enjoys IBM’s membership, consolidates and tries to survive an assault from an even more disruptive adversary, solid state storage.
Disk drives may be a dramatically disrupted and disrupting product in Christensen’s universe, but the companies producing them are not the most visible examples of industries altered by innovation, because they are buried inside information systems. To see how a really large and visible industrial company is affected by innovation, it is perhaps better to examine a corporation with ubiquity and a big public profile; IBM and Hewlett-Packard are examples. But another company, Digital Equipment Corporation, may provide an even more informative case for calm discussion, because it is long gone, as is its co-founder, Ken Olsen, who died more than four years ago. Olsen ran DEC for 35 years and was forced out in 1992. By that time, his company, which at its peak was second only to IBM in the computer business, was in woeful shape. Everything about DEC was unique, beginning with its home (for most of its existence) in a former woolen mill that made army blankets during the American Civil War. The mill is in the center of Maynard, Massachusetts, and boasts a delightful history. Clay Christensen says DEC was both a disruptive innovator and the victim of other disruptive innovators.
DEC produced radically inventive computers that were quite different from the data processing systems made by IBM at the time. Its first blockbuster hit was the PDP-8, a small, relatively inexpensive machine with a tiny instruction set that would later be seen as the first major example of a RISC processor. Computers in the PDP-8 family eventually made their way into IBM glass house installations, serving as front-end processors for IBM System/360 and System/370 mainframes. DEC PDP-11 machines were market successors of the PDP-8, but not technical offspring of the earlier DEC minicomputer. The PDP-11 attracted a lot of attention from the computer industry’s intelligentsia, particularly at Bell Labs, and became the first host for Unix, the C language, and their constellation of software. The PDP-11 evolved into the VAX family of minicomputers, systems so attractive that formerly devoted IBM customers often installed them to do critical jobs that were far more costly or awkward to implement on an IBM mainframe.
The VAX, more than any other DEC machine, is generally held to be the disruptive product that forced IBM to create the 4300 series of small mainframes and price them vastly lower than any prior IBM machine. The dramatically better price/performance of the 4300 tripped up IBM’s mainframe sales force, because customers shown the 4300 expected similar price/performance from IBM’s much larger mainframes. IBM did adjust prices and financing options to an extent, but eventually found a way to market large mainframes at much high price points than smaller compatible machines. IBM was shaken up, but survived in part because its direct rivals in the large systems market, Amdahl and Hitachi, couldn’t build really big machines much more cheaply than IBM, and the software and hardware technology to harness large numbers of more efficient 4300s simply didn’t exist.
Meanwhile, DEC was becoming the darling of academic institutions with its flagship 36-bit computers, systems that were much easier and more economical to share via remote terminals than IBM’s 24-bit and later 32-bit systems. These big machines, the largest to come from DEC, were the processors in the PDP-10 line. They machine had a precursor called the PDP-6 and a successors called the PDP-20, but it was the PDP-10 that inspired creative programmers to develop affordable and friendly timesharing systems. They became the most popular systems on Arpanet, the precursor of the Internet, and the hub of a considerable amount of application development. Microsoft’s first products were developed using tools built for the PDP-10 that included simulators for Intel chips. The development was made possible by extraordinary text editing software, programs with names like EMACS and TECO, that made the code development process fast, interactive and, for people with free access to the many university facilities based on the PDP-10, essentially free. (The PDP-10 editors produced offspring in the personal computer world, EMACS yielding a word processor called FinalWord and TECO a programmer’s editor called Vedit. While these products have largely faded away, they have their DNA in contemporary development systems used to create apps for mobile devices and myriad other software products.)
Quite a few observers say DEC was killed by the rise of the very microprocessors its computers helped develop and apply. IBM, many say, got its revenge as the PC and the tiny servers based on the same Intel technology encroached on the minicomputer business and forced DEC’s margins to unsustainably low levels. DEC tried to compete in the PC business, but its products never made it to the mainstream. The brilliance evident in the wildly popular PDP computers never showed up in a line DEC called its Rainbow machines.
It might have been corporate hubris that kept DEC from getting into the PC clone business which, at that time, was rising very quickly. IBM’s managers similarly overplayed their hand, bringing control of the PC group that had succeeded by virtue of its oddball (by IBM standards) practices in line with the rest of Big Blue. Once PCs became more IBMish, the clones, led by Compaq, had a field day. In the end, Compaq grew so large it felt it could buy its way into the server business, and in 1998, it bought DEC for more than $9 billion. By that time DEC had been stumbling, trimming down and selling off various operations in an effort to become more svelte and fit, much the way IBM has been divesting itself of manufacturing. Four years later, in 2002, Compaq, the Icarus of computing, was taken over by Hewlett-Packard.
IBM is pretty well aware of Clayton Christensen’s theories about innovation and its impact on the trajectory of a business. It is also undoubtedly aware that Christensen’s most visible critic, history professor Jill Lepore, thinks Christensen’s theories about innovation are, as Henry Ford said about history, mostly bunk. She made her views impossible to ignore by publishing a harsh essay in the New Yorker. Christensen’s initial retort, in the form of an interview published by BusinessWeek, indicates that the battle is a lot more than an academic dispute shouted too loud to be confined to Harvard Yard and its B School annex.
IBM fought back against the disruptive innovation of the minicomputer in the 1970, saw it couldn’t beat the PC revolution and decided to join it in the 1980s, moved into the services segment that had the potential to control the large systems market during the 1990s, sold off its disk drive business in the 2000s and seems poised to complete the disposal of its manufacturing operations and reorganize its operations during this decade. The company now talks about its pivot toward mobile technologies and its decision to embrace Apple as a new best friend.
While IBM has repeatedly shown itself to be far better at adapting to change than DEC, its past adjustments to disruptive innovation have come in the form of whole new products or, in the case of the PC, whole new divisions. But now IBM is divesting itself of products. And while it still makes Power and mainframe servers, its MIPS growth may already greater in the cloud than in its installed base. At the very least, it seems to be heading that way. Is the cloud the next disruptive technology that will be successfully exploited by IBM? That remains an open question. For IBM the cloud could turn out more or less the way the Rainbow PC turned out for DEC, close but no cigar.
And Clayton Christensen isn’t likely to go public with a prediction any time soon, particularly now that a colleague and historian has said his theories of innovation are less than useless when it comes to forecasting. Still, I would be surprised if IBM was moving ahead without its top executives spending some serious time talking to Clay Christensen. He’s the kind of person most IBMers would look up to, if for no other reason than he’s six foot eight.