Two of my favourite academics, Andrew McAfee and Erik Brynjolfsson, recently collaborated to publish a fascinating article on Carr’s Disease. For those who haven’t come across it, Carr’s Disease is a relatively rare condition, cycling the patient through intermittent bouts of selective blindness and 20:20 vision.
Andy and Erik set out to test some intriguing formal theories they had developed, focused on the market share concentration and sales turbulence associated with industries that had a high IT intensity and associated investment. If the data supported the theories, they would have formal evidence of the difference in the competitive dynamics between low-, medium- and high-IT-intensity firms, thereby proving the psychosomatic nature of Carr’s Disease.
I quote from the article:
The link between IT and competition surprises many researchers and executives for two reasons. First, most companies buy technology to gain control over their environments, not lose it. Enterprise systems help companies create consistency and reduce randomness, so it’s ironic that a high level of such investment would be associated with a more frenzied competitive environment.
Second, some observers have argued that information technology is so pervasive that it no longer offers companies any big advantage. If many businesses in the same industry bought the same type of large-scale commercial-enterprise software, there is reason to believe they would subsequently become more similar, and the competitive field would level. Instead, something close to the opposite has taken place.
Fascinating. Read the article for yourself, and do follow the dialogue at Andy’s blog. Here’s an excerpt:
We reasoned that since IT spending increased substantially beginning in the mid 1990s, and since not all industries spend equally on technology, one good way to assess IT’s competitive impact would be to see if there was a difference in differences: if industries that spent a lot on IT (‘high IT’ industries) experienced different competitive dynamics after the mid 1990s than they did before, and if there was a difference in this respect between high and low IT industries. If both these differences existed, it would be a strong indication that IT mattered — Â that it was a driving force behind the observed changes in competition.
But how to measure competition? Â One common metric is concentration: the extent to which market share is held by a few big firms, rather than many small ones. We also looked at turbulence, or the extent to which firms in an industry jump around in rank order from year to year (If the #5 firm in sales one year is #10 the following year, this is a pretty turbulent industry).
The results of our analyses were clear. High IT industries experienced significantly greater turbulence and concentration growth after the mid 1990s than they did before, and these differences were not as pronounced in low IT industries. The Business Insight article contains graphs that show these differences, and I’ll post and discuss more results here laterÂ . Â For now I just want to point readers to the article (a paper written for an academic audience that discusses the research design and results in more formal terms is available here) and solicit their reactions.
These dialogues are important, so could I encourage you to participate?
My personal take on it is quite simple. First, the concentration growth. Any industry sector facing increased commoditisation and consequent margin pressure should display the concentration growth characteristics found in the study, as participants strive to hold on to the vanilla while seeking to grow the higher-margin complex; there is always a shake-out and a consolidation. I think this happens in low-, medium- and high-IT-intensity sectors; the difference is that the process is accelerated by IT and therefore more pronounced in the high-IT-intensity sectors.
The sales turbulence is far more interesting. My guess is that it shows how poor we really are in sustaining the value generated by IT investments. This happens for a variety of reasons: suboptimal prioritisation processes; an unwillingness to make the requisite people-process-technology lockstep changes that will crystallise the value; a lack of understanding of the sheer speed of change in the high-intensity competitive environment; inconsistent approaches to IT investments within different divisions of the same firm, creating artificial and unnecessary variances in performance; a cyclical tug-of-war between product and service innovation champions as executives revert to type; the list is endless.
One way of looking at it is that we need process innovation in order to sustain the value generated by IT investment; I think that the greater the IT-intensity of a sector, the more radical the process innovation has to be. For most of us, this is uncharted territory: we haven’t quite grasped the “Don’t Automate, Obliterate” mantras that Hammer was espousing in the late 1980s; instead, we still have this tendency to “pave the cowpaths”, as he termed it. I think this is caused, at least in part, by the battle between professions that Andrew Abbott writes about, as we move gently towards a Wilsonian consilience.
In essence, I believe that the market share concentration growth is a phenomenon that manifests itself in all industry sectors facing commoditisation, and that this phenomenon is accelerated in high-IT intensity industries, an “environmental” effect. On the other hand, I believe that the sales turbulence increase is a self-inflicted wound, showing our relative immaturity in sustaining the value we create. This is not surprising given our ambivalent attitude to such investments, itself a reaction to irrational markets and, sadly, also a response to Carr.
Perhaps I should have left it to Sy and Ella to explain what I think (or maybe Fun Boy Three and Bananarama for younger readers). Do let Andy know what you think, or comment here if you feel lazy, I’ll make sure he gets it.