I asked Phil Rosenzweig to take a look at what I've written about the halo/horns over the past couple of postings and offer his comments. Also, check out his post about the problems with Jim Collins' Good to Great/Built to Last research.

For starters, I'm not an IT professional--I worked in finance and marketing, and now am a business school professor, but of course I work around IT every day. I'll make three points.

First, my guess is that IT isn't less respected or less appreciated than a number of other functions.  Years ago, I worked in cost accounting at Hewlett-Packard and my colleagues and I were gently but routinely derided as "bean counters" by the engineers. Accounting is often seen as a useless overhead function performed by corporate drones in green eye-shades.  But that's nothing compared to Human Resources, the Rodney Dangerfield of functions.  At least accountants are thought to have some ability with numbers, which is more than most people say for HR professionals--who are often thought to be well-meaning folks with no particular skills whatsoever!  I suspect most functions fear that they are held in disregard by their peers, and on that score, IT is probably no worse than average.

Second, about halo and devil effects, people have a natural tendency to make inferences about things that are fuzzy and ambiguous on the basis of other things that are seemingly tangible and concrete.  As it relates to company performance, we often take things that are apparently rigorous and precise--financial figures of sales, profits, and share price--and make attributions about things that are less precise--such as corporate culture, vision, leadership, and so on.  The problem in so much research about company performance--including some of the best-known best-sellers of recent years--is that the things they claim drive performance are in fact attributions based on performance.  The studies are fatally flawed because the independent variables are not truly independent of the thing they are trying to explain, namely performance. 

As it relates to IT, I suspect that what is most salient and memorable are those few but very aggravating episodes of lost data, systems down, and so on.  What is overlooked are the many more times when systems work well.  Our minds grasp what is salient and exaggerate its importance, making sweeping attributions about IT incompetence or lack of dependability.  We remember the failures and attach an exaggerated importance, a little like the way that the occasional airplane crash makes the news, but the very many car accidents don't.  That's the paradox--the better IT works, the more salient are the few occasions when it lets us down.

Third, the way to keep the halo/devil effect in check is to try to maintain a sense of perspective, and not let a few salient episodes shape opinion.  One of my colleagues defines customer satisfaction as "pQ/e," where Q is quality, p is perception, and e is expectation.  Satisfaction is a function of perceived quality relative to expectations.  The key point is that most of us spend out time trying to improve the Quality of our service or product, but forget that it's also important to set expectations and to shape perceptions.

One example: recently I was going through security at the Copenhagen airport.  The sign said: "We promise that 95% of the time, your waiting time will be less than 5 minutes."  There was also a graph that showed waiting times throughout the year, which showed that 98% of the time, the waiting was less than 5 minutes--they were beating their promise.  Now, I was reading that while in a line that took quite a bit more than five minutes.  My perception?  "Gee, I must be in the unlucky two percent!  These guys are really good--my experience is exceptional!"  Rather than remember how long my wait was, I remembered how, overall, the airport security is efficient and fast.  They set my expectations, shaped my perception, and I left satisfied--despite a long wait.

The corollary for IT is to make sure that expectations are well communicated, and also that the record of performance is correctly perceived.  If IT is in fact, from an objective standpoint, very reliable, have you made sure that people understand that?  Otherwise, it's quite possible that people will blow occasional failures out of proportion--and reach unfortunate conclusions.

 

Some good points here. The only one I really take issue with is the first. To say that there are functions that are treated worse than IT misses the point, I think. I think that IT has become so deeply tied to the business that to treat it like just another function is anachronistic. IT won't shed its horns by feeling good about being less derided than HR...

You do not have flash or javascript support.

Average (0 votes)

Reply