Does IT matter?

- by Paul Murphy, -

The May 2003 publication, in the Harvard Business Review, of an article by Nicholas Carr titled "IT Doesn't Matter" raised so much controversy that opportunists, like Keehn and Norris writing in Syllabus Magazine, can now use a denunciation of his views as a kind of publicity virus to get their own rather special pleadings wider media attention.

Carr's fundamental argument is that every widely used business technology conferred significant strategic advantage on early adopters but lost that potency as it matured because widespread use levelled the playing field. If, as he says, IT is now mature, then it must also have lost its ability to confer strategic advantage and therefore companies which continue to invest in strategic IT initiatives are just wasting their money.

Strategic advantage means being able to do something contributory to winning the competition which your competitor can't. Thus Carr's basic argument is a gimme: my bronze sword or voter registration list, for example, confers a strategic advantage only as long as you don't have one, so the issue isn't whether this happens, but whether it has already happened to IT.

He uses commoditization as a proxy for maturity, and says the IT market has become commoditized; ergo mature. Most of the major players on Wall Street agree with him; but almost everybody with a stake in the industry disagrees. If you ignore conspiracy theorists and people being paid to puff the Itanium as a commodity processor, the basic pro argument comes down to the idea that the market has spoken with only Microsoft, Dell, and IBM likely to survive the coming shakeout. Thus most of the more highly publicised attacks listed on Dr. Carr's website can be disregarded because people like Bill Gates, Steve Balmer, Craig Barret, and Carly Fiorina are just faking outrage to bolster sales among the gullible.

Focus on Carr's argument and there seems little doubt that he's right with respect to the world of IT as he sees it. It's tough to argue, for example, that upgrading from Windows 2000 Server to Windows 2003/XP Server or from one generation of x86 processor to the next, conveys any strategic advantage to those doing it. On the contrary you could see the march from Office 5.0 on Windows 3.11 to Office/XP "Professional" on Windows/XP as exactly the kind of expensive but valueless product churn you'd expect in a commoditized market where everybody has to keep up with every body else, but the new features gained at each step aren't useful enough to let early adopters get ahead long enough for the difference to matter.

Notice that the issue here is relative business advantage, not technology. Few would argue that Windows/XP isn't technically better than Windows 3.11 or that a P4 running at 3.2GHz with 512MB of RAM doesn't out perform a 386 running at 16MHz with 8MB of RAM. It's quite reasonable, however, to argue that these changes have had very little, if any, net positive impact on user productivity or the business benefits companies actually achieve with the technology. Thus, had Dr. Carr titled his paper something like: "Wintel doesn't matter" instead of "IT doesn't matter" he would have had to change very little in the body of the thing and been essentially irrefutable on the facts.

Most of the people commenting on his work suffer the same failing; silently refusing to recognise the implications their assumptions about Wintel dominance have on both his conclusions and their support or rebuttals. For this reason the repeated the editorial responses in magazines like Information Week remind me more of the of the drunk searching for his car keys under the street light - because that's where the light is - than of Porter lecturing on comparative economic advantage in business.

At Fortune, for example, both Kirkpatrick and Alsop are so deeply committed to their faith in Microsoft and Dell that their perception of IT opportunities appears to be both defined by, and limited to, the PC industry press release. Unfortunately their ability to see Dell as a manufacturer is a lot like their willingness to call Microsoft an engine of American innovation; it contradicts their basic argument on commoditization while making marketeers blush with pride and the rest of us flush with embarrassment for them.

One of the stranger things about these articles, columns, and editorials, whether pro or con, is that the writers involved generally seem to assume that IBM's size gives it an automatic role in IT but to have no idea what that might be. As a result they discharge their obligations to objectivity and IT beyond the PC through mercilessly upbeat references to IBM's leading role in whatever technology they're waving off while ignored the rest of the industry entirely.

What Carr missed, and neither his supporters nor detractors seem to have noticed, is that they're all trapped inside the Windows box. Step outside, and there are lots of opportunities for companies and industries looking for ways to gain strategic advantage from IT.

Technology advantages tend to be short lived with companies which succeed in converting the combination of some new technology and modified business processes into competitive success then converting that success into longer term advantages like those of scale or market dominance while the rest of the industry plays catch-up. That's how successful technologies reshape their industries and eventually level the playing field for new competitors.

The most obvious and widely cited positive example of this is the role of the Mac in the printing and related industries. Here use of the Mac as an interface to Postscript printing perfectly complemented existing ways of doing business and thus raised productivity among adopters while those who stayed with older software, mainly from DEC or IBM, stagnated and those who opted for PC based systems experienced cost and productivity disadvantages. Thus use of PostScript and the Mac eventually became pervasive and remodelled an entire industry around itself.

Not all such transformations are positive. Change in the accounting industry, for example, seems likely to be seen by history as an industry wide IT failure because the widespread adoption of the Wintel PC led the big firms to make catastrophic changes in their operations and internal organization. Prior to the PC, partners were senior practitioners whose collective conscience determined the firm's primary product: it's overall credibility. After the PC, partners had become faceless project managers focused on packaging, selling, and managing standardized services delivered by laptop toting juniors billing by the quarter hour.

What forced this metamorphosis along was the fact that senior management could deliver substantially increased returns by selling more lower cost junior time combined with their belief that they could nevertheless maintain quality by embedding standardized methodologies in the PC software used by those juniors. Thus when KPMG became the last major firm to change from the Mac to the PC, individual productivity fell considerably but overall recoveries increased sharply simply because a much larger portion of the firm's total billings was coming from cheap junior time.

It now appears, however, that this strategy has failed as the industry's credibility has became based on branding rather than people, and the firms have lost focus on their role as ethical arbitrators and decision advisors. Thus the current spate of SEC investigations suggests that it may be past time to bring back the professional accounting practitioner, and therefore the former focus on personal responsibility and productivity. Indeed it seems likely now that bigger firms which don't soon make a strategic commitment to more productive forms of IT for senior people -the Mac or the SunRay- will be destroyed by litigation and client attrition.

Linux has yet to really transform any single sector and so offers huge strategic oportunities for knowledgeable early adopters. It's obvious that adopting Linux provides cost and reliability gains over using Windows and that there should, therefore, be strategic advantages to be gained through adoption. So far, however, relatively few companies have clearly succeeded in doing this. My belief is that this is mostly because businesses tend to adopt Linux but perpetuate the MCSE way of thinking about computing and thus gain short term cost reductions but limited strategic benefit. Both Linux and BSD have been widely adopted, for example, in the remote hosting industry, but the fundamental operational model there continues to be that of the Microsoft Server with both customers and suppliers insisting on dedicated gear where shared would be far more efficient.

The areas where the use of Linux or BSD on x86 has delivered major strategic gains are mostly limited, so far, to academic and small company systems development environments where doing more with less is often critically important. In that environment Linux is delivering strategic advantage now with virtually every new computing initiative or new piece of software appearing first on some form of x86 Unix, but a thirty person company that doubles in size every year is still going to take eight to ten years to make the big leagues. As a result, we probably have several years to go before we see industry wide transformations brought on by x86 Unix.

That's true too of the Sunray and other forms of smart display computing. There is a silent revolution going on here too, but it's mainly focused in the defence and other high security environments with relatively few civilian adopters so far. Sun, however, manages its own business using Sunrays and the Unix business Architecture, nicely demonstrating the potential strategic advantage by needing fewer than 30 people to support a high reliability worldwide deployment of over 30,000 desktops.

Carr, and the people responding to him, have missed all of this. To them, if it isn't Windows, it doesn't exist, unless maybe it's some semi-magical IBM thingee they can wave their hands over. In his own context he's absolutely right: limit IT to Wintel, and IT doesn't matter; step outside his assumptions, however, and he's dead wrong.

Paul Murphy wrote and published The Unix Guide to Defenestration. Murphy is a 20-year veteran of the IT consulting industry.