An Average Tyranny

Peter de Jager is a provocative Speaker, Writer and Consultant. His primary focus in on how we manage change, technology and the future.

In addition to speaking at conferences worldwide, he's also written monthly columns for Municipal World and Computing Canada.

His goal is always to question what we think is so, and in so doing perhaps open up new opportunities.

You can contact him at
pdejager@technobility.com

Or sign the Guest Book and he'll get back to you.

It all starts innocently enough, we wait anxiously for our marks to return from last week’s exams, and when they do, the urge to compare ourselves to our class mates is overpowering. It isn’t enough to find out how we did, we need to know how we did compared to others. More specifically how we compared to the class average. If we’re above the mean, then we glow with smugness, if we’re below – well it was a stupid test, and we’ll never use math when we grow up.

Then, as we enter organizational life, we all too often descend into realms of arithmetical silliness. We insist on asking, “How do staffing levels compare to the average of other offices?”, “How do our computer expenses vs. gross sales relate to the industry average?”, “How do our postage costs relate to the industry average?”

Yet, there are some good average comparisons; “How do our absenteeism rates compare with similar sized municipalities?”, “How do our turnover rates compare to other municipalities?”, “How do our voter turnout percentages compare to the national average of other municipalities?”

The challenge for everyone, including those of us who must put aside our mathematical objections to the ‘average’ as a useful statistical measure, is to determine when comparing ourselves to an industry average makes any sense.

The mathematical objections to using just the ‘average’ as a valid measure of performance are significant. Assume you make \$100,000 a year, and it turns out that you’re making 33% less than the industry average. Are you pleased? Furious? Upset? Should you be beating down the boss’s door and demanding a raise? It all depends. Are you a part of this peculiar population? {\$50k, \$50k, \$50k, \$50k, \$50k, \$50k \$50k, \$50k, \$1,000,000, \$100k ← you} Or are you someone in this more likely population? {\$150k, \$160k, \$130k, \$180k, \$155k, \$155k \$155k, \$155k, \$155k, \$100k ← you}

Mathematically speaking, a far more useful measure is the ‘Percentile’ – this measurement tells us what percentage of the population falls below us. In the first example above, the percentile associated with our \$100k salary is the 80th percentile, which indicates we’re doing better than 80% of the sample. Don’t let the media find out.

In the second, more normal sample? Our \$100k salary is in the 0th percentile. We’re at the very bottom of the heap. Time to schedule that meeting with the boss.

While the mathematical objections to using just the average to rate our performance are real, they pale in comparison to the apples and oranges issue. Are we comparing like to like? If not, then the comparison is useless. A word to the wise… It’s nearly always useless.

For years I worked in the IT side of several industries, from retailers to banks, from computer research companies to application development organizations, and the most prevailing measure was “IT costs as a percentage of sales”. Now those asking for these comparisons are fleetingly aware of the need for ‘like to like’ comparisons. Bankers do not compare themselves to clothing retailers. They recognize these types of organizations do different things and therefore, their IT expenditures won’t bear any relationship to each other. This much at least, is obvious, even to the most mathematically challenged manager.

They will however, insist it is appropriate to compare their expenditures to their competition in their own industry. That’s an easy assumption to make, and usually a dangerous one. Once we’re past the core accounting systems, there’s very little in common from one organization to another.

There’s some strong human psychology at work here. Comparing ourselves to others, and desiring what they’ve achieved is almost hardwired. Any human tendency warranting not one, but two places in the Ten Commandments, the two ‘thou shalt not covet’s, isn’t something we can easily cast aside. Even if I know that comparing my office costs to anyone else is meaningless, I’ll admit that my desire to know isn’t diminished.

The problem of course, isn’t the mere act of comparing ourselves to others, but it’s what we do with that comparison that makes it useful or dangerous.

Consider the two ‘good’ comparisons we listed earlier. How do our absenteeism rates compare with similar sized municipalities? How do our turnover rates compare to other municipalities?

Why are these ‘good’ comparisons? Because if our rates are higher than the industry average, then it doesn’t matter why… we must reduce them. Once we realize something is wrong we’re immediately prompted towards the next bit of research… what are they doing right that we are doing wrong? The key observation here is that certain performance levels should be in line with industry averages. Absenteeism and turnover rates are two such numbers.

Even so, is this year’s turnover an anomaly? Is it exceptionally high because the operations department was playing the Super 7 Lottery and 25 people resigned, doing a very silly dance as they traipsed out the door? Is it exceptionally low, because your Municipality has been hit by industry specific layoffs and the region is suffering high unemployment?

In many other industry comparisons, things aren’t that clear cut. Averages are not intended to be best practice performance targets. Averages are, by definition, indicators of mediocre performance.

So what if my IT costs are 75% higher than average, that is not necessarily a bad thing. Not if I’m delivering 300% more of what our constituents want delivered, not if the expenditures are reducing costs in other areas by 90%, not if my head count is 50% of other municipalities…

Sadly, while it’s obvious that setting industry averages as performance targets doesn’t make sense without taking all the other variables into account, including temporary aberrations, it’s all too common a practice. Why? Because it’s an easy measure. It is much more difficult to look at why our costs might legitimately differ from the norm. So difficult in fact, that comparing ourselves to the industry standard has become the de facto mediocre industry management standard.

© 2008, Peter de Jager – Peter is a keynote speaker, and he doesn’t believe that average performances are desirable goals – neither do his audiences. You can contact him at pdejager@technobility.com  You can read more of his early morning scribblings at: http://blog.technobility.com

1. Name
2. Email ID (Optional - will be used only by this site)
3. Your thoughts on the question.
4. Mailing list - 1-2 e-mails per month - updating you on new articles and presentations
Do -NOT- add me to the list

Resources:

Change Management

IT Management

Feedback from Clients