When last we met our our faithful and intrepid blogger, newly under the influence of thought leader Clay Christensen, I was anointing myself as IDIOT Director, to wholly unanimous agreement. (That was more than a week ago, I must note. Christensen may have driven me to disrupt UVU, but I’ll be damned if he is going to disrupt my Thanksgiving.) I had proposed replacing institutional research and effectiveness with a future-oriented, theory driven planning organization, the Institutional Disruption and Innovation Office for Theory. Such an organization will be far more effective for planning, assuming that data tells us nothing about theory; that theory can make useful predictions without data; that theory without data is somehow supported by more than cherry-picked anecdotes that support a predetermined conclusion; and that there is a fundamental ontological break between the past and the future. You will recall that not even my iPad found this persuasive.
Modest proposal number 2: Implement innovations with the highest potential for disruption. Using the theories developed by the IDIOT, UVU will adopt instructional practices designed to maximize access to higher education will minimizing cost and reducing quality. This will involve eliminating virtually all directed instruction, whether by in person by tenure-track or adjunct faculty or online using content developed by faculty or instructional designers. Lecture capture, for instance, does not reduce costs or quality sufficient for full disruptive potential. UVU’s instruction will be delivered instead in a purely peer-driven framework. Instruction will thus be opinion-based; consist entirely of participation in unmoderated discussion forums; and result in the award of gold stars for having an opinion.
Initially it was considered that UVU could partner with Fox News and MS-NBC to provide some content around which discussion could take place. However, these sources are not, in fact, better than non-consumption; studies suggest that viewers of these outlets are less informed than those who do not have any news exposure. Instead, courses will take place through Craigslist discussions. Students who demonstrate more intelligent opinions on the topic than the average Craigslist user, as determined by Big Data, will be deemed to have demonstrated that their knowledge is superior to non-consumption and receive the gold star. Students will be awarded degrees based in the number of stars received in discussion forums relevant to a topic, with 300 stars leading to an associate’s degree and 600 leading to a bachelor’s degree (roughly equivalent to one opinion per traditional course per week).
Christensen’s theory of disruptive innovation attempts to explain innovations that radically transform markets rather than making incremental change. Implicitly, one must note, disruptive innovations fundamentally change the firms participating in those markets: Christensen offers no examples in which an existing major player in a market disrupts the technological status quo ante; in all cases the disruptions come from outside the existing firms, either from new firms (e.g. Netflix) or from existing firms not currently involved in the market to be disrupted (as with the iPhone). The CD fundamentally changed technology in the audio market beginning in the late 1980s, but the existing firms all adopted the new technology; hence the CD is not cited as a disruptive technology. Central to Christensen’s story of the PC as a disruptive technology is that it destroyed Digital Equipment Corporation. This is very much a theory of firms rather than of technologies.
In routine, non-disruptive innovation, Christensen argues that while technologies initially emerge as less capable than their users, the rate of technological innovation is sufficiently greater than the rate of user learning that technologies rapidly exceed the capabilities of their existing users. His example here is curious, though: he notes that running Word Perfect on an Intel 286-based PC (which he inexplicably called the fastest processor in the world at the time) required one to type slowly so that the processor could keep up. but that modern processors use only 15% of their capacity.
Obviously, Christensen is not a data scientist. Or a gamer.
This leads him to conclude that technologies can innovate to the point where a competitor entering the market can disrupt the existing participants. The current, relatively small market, is already using technologies that exceed the buyers’ needs, so further incremental innovation is useless. Instead, the new market entrant enters with a product that is cheaper, more accessible—and worse.
Yup. The new technology must merely be better than non-consumption in order to disrupt the market.
I.e., they need to suck just a little less than Olive Garden. #lrt
— Jeffrey Alan Johnson (@the_other_jeff) November 21, 2014
Hence Christensen’s argument that online education is disrupting higher education. He uses the example of a colleague who taught a course at a Scandinavian university from Boston using an innovative communications system and an iPad attached to a Segway-like robot. Christensen didn’t say that this was better than in-person education, but did conclude that it was good enough to get the job done. Online technologies like this, he argues, will open higher education to many people who would otherwise not have access to it. Quality will suffer, but not unacceptably.
Let that sink in: Christensen is arguing for lower quality education. Which leads to, well, this:
Clay Christiansen says the solution to disruptive innovation for Harvard is to buy University of Phoenix.
— Jeffrey Alan Johnson (@the_other_jeff) November 20, 2014
No joke: the University of Phoenix is the future of higher education. He doubled down on that later in the talk, after arguing that online education allows higher ed to keep customers “at arms’ length,” avoiding unnecessary interdependencies like the idea that content needs to add up to a coherent degree. Certification is the main objective. Higher education’s future thus lies in online badges.
As far as I can tell, Clay Christensen is dead serious about turning universities into the educational equivalent of McDonalds. To do so would be devastating to so many of the other functions of higher education. If he is right about how disruptive innovation then we may well have to accept that this is the best option left after ruling out the impossible. Online technology is destiny from this perspective.
But Christensen’s theory is deeply problematic. He offers as examples of this theory the disruption of the minicomputer by the Apple personal computer, of the U.S. auto industry by the entry of Toyota into the U.S. market, and the disruption of the smartphone market by first Apple and then Samsung. And with the relationship between theory and data that he articulates (as I dicussed in part I) we should not be surprised that the problems of theory become evident here: none of his cases hold up.
Consider the case of DEC first. Christensen’s story is that DEC failed because it stuck with its existing business customer base, which saw no need for the personal computer. The minicomputer was the only machine that had the processing power to carry out their business needs. But as personal computers, whether Apple or DOS/Intel based, became widely available DEC was stuck with an obsolete technology; the PC had beaten the minicomputer.
This is, of course, bollocks to anyone who actually works with information technology. The minicomputer was not disrupted. We simply have a new name for it: a server. When my graduate program replaced its VAX minicomputer (in the very late 1990s, I note), it didn’t get a low-end PC running Windows 95; we got a server running Windows NT that carried out exactly the same functions. When I run SQL queries on the UVU database, I don’t use my PC’s processing power; I use the PC as essentially a dumb terminal to access our database server, where the query is executed. Ultimately, DEC wasn’t killed by Apple or IBM’s PC. It died because the VAX 9000 sucked (a point on which I will accept no counterargument, having used the thing to prepare for comps), Sun and Oracle were making far better UNIX-based systems, and Microsoft ended development and support of Windows NT for DEC’s late-1990s Alpha computers. Digital may have missed out on the PC market, but could have been an effective player in the server and database markets had it been able to compete with Sun and Oracle—and not had extraordinarily high overhead costs, a factor not considered in Christensen’s model.
The case of Toyota is equally at odds with reality. Christensen describes the Corolla as the epitome of his low-cost, high-access, low-quality innovation. But the Corolla’s introduction in 1968, and its success in the 1970s, is anything but. The Corolla was certainly cheaper. But it was less accessible thanks to fewer dealers than the American Big Three (especially if one takes into account the need for parts and service; those who remember the era undoubtedly remember the birth of “import” parts shops and garages). Nor was being merely better than non-consumption a factor, as the American auto market was more-or-less saturated by that point. There simply wasn’t a large enough pool of potential buyers who were being kept out of the market for Toyota to have succeeded following Christensen’s strategy. And most importantly, the Corolla was better than anything Ford, Chrysler, or GM were making, let alone the now-dead American Motors. When the oil crises hit in the 1970s, Toyota’s smaller, more fuel-efficient, and more reliable cars were better suited to changes in consumer demand.
Christensen also maintains Apple disrupted Blackberry in the smartphone market, but is now just hanging on in the face of disruptive innovation from Samsung’s Android phones. It’s an odd claim, considering Samsung lost nearly a third of its market share between 2013 and 2014. Apple has never been above 20% of the market, and Android’s growth has come mainly at the expense of Symbian—recall that Symbian formerly powered phones by Sony and Samsung, which switched to Android, and Nokia, which switched to Windows Phone. In fact the iPhone didn’t disrupt the smartphone market as much as it did the market for (what we call without the slightest hint of irony) feature phones. You know, those phones that were far cheaper than the iPhone and that could only text using a numeric keypad.
Apple’s mastery of disrupting markets with premium products points to a general failure of Christensen’s independent variables to predict the dependent variable. Disruptive innovations involving costly, elitist, and far better quality products are quite common: witness the rise of the craft beer industry, where restricted access is a selling point, price is no obstacle, and quality has spawned a level of pomposity previously attained only among wine snobs. Netflix is cheaper and more accessible than Blockbuster ever was, but it is also far better, with vast selection and none of the logistical efforts needed to rent from—and return to—a Blockbuster store. And if bottled water has done anything useful, it is to prove that a product that is more expensive, less accessible, and (thanks to the problems of transporting, storing, and disposing of waste) lower quality can be disruptive if marketed right.
We can also point to many cases that should have been disruptive but weren’t. If Christensen is right, Toyota should have itself been disrupted by Yugo in the 1980s, a much cheaper car that truly was just barely better than non-consumption; it isn’t entirely snark to say that virtually any product of the Soviet economy should have disrupted Western markets in the Cold War era. Linux may have come to have a respectable share of some IT markets such as web servers, but even being free, downloadable, and notoriously difficult to keep running as a desktop environment its market share is below 2%. In the real world that Christensen’s theory purports to predict, the dependent variable (disruption) simply does not vary with the independent variables (cost, access, quality), something a rigorous method would spot.
Christensen’s argument about distance education fails similarly. The technique used by his colleague is certainly not the state of the art in distance ed. His colleague didn’t educate more students than he could have in person, and he did so much more expensively than being physically present in the class where robots and new software would have been unnecessary. This case should not be disruptive. More routine kinds of online education, such as LMS-driven online content delivery supplemented by discussion and multiple-choice assessment, fits his mold better. But it isn’t disrupting the producers of high-end higher ed; Harvard remains one of the least accessible institutions in the country while the net cost of UC-Berkeley is higher than Yale. Regional state universities are the ones most being challenged by online education, but that is driven as much by political factors such as politically appointed trustees and elected officials that demand that their institutions be seen on the cutting edge. It is unlikely that any institution will be forced to change simply because of competition from the University of Phoenix, and telling that for-profit institutions, which generally rushed to embrace online learning, are the ones going bankrupt.
This is a serious problem for higher education. Christensen’s argument is essentially a form of technological determinism: a new technology will determine a single course of action for all actors; only those who go along with this course will survive. Twenty years ago political economists were making the same argument about globalization: that it could not be resisted, that it would force the same changes on everyone, that the only possibility was to embrace its necessity. But those who studied political and economic institutions knew better. A single stimulus can result in very different responses from different actors. Countries might have different institutions that turn conditions into policies. They may have a different mix of economic sectors that are more or less affected by globalization. They may stand to gain or lose, or be in different positions to force their views on other states (as, for example, the International Monetary Fund and World Bank did). Most had to respond to change, but they responded differently.
Higher education is no different, a point Sherman Dorn makes in a blog post on Southern New Hampshire University. “I don’t believe the shape of institutional change is predetermined even if the existence of change is predictable,” he writes. Looking at SNHU’s online programs, he notes that specific conditions and characteristics of the institution have played important roles in the institution’s direction. This is a more sound approach: look to the circumstances of the institution to see how the inevitable change will affect a specific institution. It is unlikely that the answer will be the same for everyone.
In the conclusion of this series I’ll take up Christensen’s ideas about what education is really about.