I ran across an interesting rebuttal to Harvard Business Review’s now-infamous article “IT Doesn’t Matter“. It focuses on the article’s claim that a resource’s strategic value is derived from its scarcity and that since IT is a commodity, it isn’t scare and thus has no strategic value. The rebuttal rightly points out that a resource’s strategic value is actually derived from the way it is managed. Resources like capital and talent have wide ranging impact, based on how well or poorly they are managed. (The rebuttal suggests HBR should print an article titled “Talent Doesn’t Matter”). However, I think that HBR’s article is further flawed in its basic premise that IT has become a commodity.
The article draws an analogy to the effect electricity had on manufacturing. Before electricity, manufacturing plants were constructed around a central source of kinetic energy such as a steam engine or waterwheel. A complex system of pulleys and gears distributed that kinetic energy to the various workstations within the plant. This system of pulleys and gears was cumbersome, inflexible and costly, but it was the only choice. When electricity was first introduced, electric generators replaced the waterwheel and were used to power the complex system of pulleys and gears. Eventually, someone figured out how much more efficient it is to distribute electricity directly to the workstations. This electricity powered smaller, cheaper and more flexible machines directly at the workstation. For a while, companies that were smarter about the use of electricity gained an advantage on those that weren’t. But eventually, everyone caught on and that advantage was lost.
This is amazing analogy to the current state of distributed system
technology. However, HBR draws the wrong conclusion. They conclude that
IT has become a commodity, and that while it provided a short term
benefit to early adopters, that that advantage is largely gone today. On
this issue of commoditization, HBR is only half right. Technology is a
commodity, but information is not (or at least, not yet). Taking the
waterwheel/electricity analogy forward, early computer systems were
constructed around the mainframe as the central source of information
processing. A complex system of programs and terminals distributed that
information to the various workstations within the company. This system
of applications and terminals was cumbersome, inflexible and costly, but
it was the only choice. Now, even though we have powerful information
processing systems right on our desktops, we continue to distribute
information via a complex system of programs and terminals browsers.
Building web applications is the computerized equivalent to replacing
the waterwheel.
I imagine that for the factory owners back in the early days of electricity, the value proposition of replacing the water wheel with an electric generator was very compelling. No need to replace the entire infrastructure, simply swap out the current source of kinetic energy with one that was cheaper and more reliable. After all, building the system of pulleys and gears in the first place had been expensive, and the cost of replacing it would be tantamount to rebuilding the factory from scratch. Budgets were tight back then too, after all. And I’m sure that the old system of pulleys and gears powered by electricity worked just fine, right up until the company got its ass kicked in the marketplace by a company that had embraced electricity even if it meant scraping their previous infrastructure of pulleys and gears.
Eventually, I believe forward-looking enterprises will figure out how much more efficient it is to distribute information directly to the desktop the way forward-looking manufactures distributed electricity. This information will drive smaller, cheaper and more flexible applications directly at the desktop. For a while, companies that are smarter about their use of information will gain an advantage on those that just replacing waterwheels. I assume eventually, everyone will catch one and that advantage will be lost. But there’s a long long time between now and then.