Nick Carr was right and I was wrong. Sort of, anyway.
I confess to having roughed up Nick Carr quite a bit (in print, of course) over the past couple of years. I thought he was plain wrong in his book Does IT Matter? His argument there could be summed up as the spread of IT throughout the economy makes it impossible for any company to achieve competitive advantage through IT; therefore, companies should settle for a commodity IT perspective, spending no more than the minimum necessary to perform basic functionality.
I feel that his characterisation of competitive advantage is a straw man; he seems to imply competitive advantage is a permanent advantage, immune to attack by other market participants. Information technology, just like every other form of competitive advantage, is subject to erosion over time, whether through mimicry by competitors or being supplanted by newer, better forms of competitive advantage. Moreover, with the spread of IT throughout the economy, IT is more important to companies, not less, since it infuses whatever a company puts forth as its competitive advantage. To take an example, certain high-end hotel chains emphasise their personal service, remembering that you like early am wakeup calls, or always want to use a particular limousine service, etc. Those chains don't depend upon individual staff to remember those details; they use IT to support the service offering-behind every smiling staff member who seems to know what you want before you do stands a sophisticated database system which stores information on every guest, carefully graded by frequency of visit, typical spend patterns, etc, etc. So one could argue that IT isn't their competitive advantage; on the other hand, one could argue that IT is fundamental to their competitive advantage.
Therefore, I felt Carr was well off the mark with that book.
His more recent book, The Big Switch , however, is a very different proposition. In that book, a discussion of the emergence of cloud computing, he outlines a very real trend and trenchantly observes that it is transforming the way IT is done. One of Carr's real strengths as a thinker and writer is his ability to examine historical developments with a fresh eye and discern important lessons germane to the topic at hand. So, for example, in Does IT Matter?, he discussed the rise and role of railroads, observing that they initially gave huge competitive advantage, but once they were built out, there was little competitive advantage available from railroad availability; nevertheless, despite the reduced advantage, enormous sums continued to be poured into railroad development, destroying everyone's margins as a result. Likewise, he concluded, IT was suffering from the same trend: initial advantage eroded by wide availability, with excess investment nevertheless being wasted by shortsighted corporations.
In The Big Switch, Carr reaches to another 19th century development to help understand today's IT-electric power. In the early days of industrial electricity use, companies had to generate their own power. Each company built its own powerplant, arranging for fuel delivery, generating capacity design, and so on. Eventually, electrical utilities were formed, making it possible to gain the benefits of electricity without needing to self-generate. Because the utilities had access to more capital, could manage load factors across a larger pool of users, and could bring more expertise to bear, company reliance on self-generation dwindled, replaced by public grid power.
Carr analogises from this historical trend the inevitable end of corporate data centers. Whereas it was necessary for companies to run their own computing infrastructures in the early days of computing, today utility computing companies like Amazon and Google have sprung up. With cheap capital, ability to manage load across multitudes of customers, and specialised expertise in data center design, Carr forecasts that company-owned data centers will fade away, replaced by access to a public computing grid.
I was dazzled by Carr's historical analogy (and, in fact, was enthralled by his description of the rise of electricity). However, on my original reading of his book, I was put off by his, to my mind, over-glib description of how cloud computing works. He described how a friend of his put together a website, snapping together individual applications like blogging, photos, etc. From this, he concluded that corporations would therefore move to the cloud, getting rid of all those pesky developers and extra IT headcount. I felt (and feel) that Carr does not grasp the complexity of typical corporation computing infrastructures and production applications. Moreover, I felt (and feel) that he underestimates the rapid pace of technology change in IT, all thanks to Moore's Law (and also thanks, perhaps more directly, to Moore's company.)
The topology of application delivery has changed several times throughout the (roughly) 50 year history of computing. Originally, everything was centralised in an on-premise mainframe. Then timeshare, supported by primitive public networking, came along (in fact, nearly every time I speak on cloud computing, someone asks "Isn't this just like timesharing in the old days?"). Then, with the dropped cost of computers, came centralised minicomputers. Distributed PC-based client/server followed that. Then came remote computing via the Internet. Now comes cloud computing.