22.2.08

Useless statistics, and the Pareto principle

I'm sure there's an easier way to get at the number I was after, but this past weekend, I needed some hard data to support my assertion that there is a very large market for a book about working in a global workplace, written for technologists. I was writing a book proposal, and wanted to back up my assertion with a little bit more than "really really lots, I'm certain of it!"

To get at a meaningful number, I worked backwards from the number of people employed on the vendor side of the global IT services business.

I took the Global Services Top-100 list, and did some very tedious pick and ax work to figure out how many employees are represented in this top 20% or so of the global IT service business. I went to each of 100 web sites, and poked around until I found a company "fact sheet" or some similar reference that reported how many employees the company has.

The good news is that 60% of the companies in my survey openly discussed their size, in what would seem to be the most relevant measurement of service delivery capacity -- the number of people who work in the company.

That means 40% of the companies attempted to obscure the size of their company. They either didn't talk about it, or blatantly obfuscated the facts about themselves. That's understandable. If you're the little guy, competing against TCS or Wipro, you may not want people to know that you only have 40 engineers. I'd argue that hiding the information is useless, but what ever... It's their company...

But some of the companies did something that's just bizarre, that I think is a great follow up to the post about Moneyball, and the way pro-baseball measures the wrong stuff.

Lots of those 40 companies that didn't like to talk about how many employees they had did like to talk about how big their company is, they just did it in very literal terms.

That is, over and over on these company's web sites, I saw references to how big the company was. In square feet. Of surface area.

Now, I really can't imagine a less useful fact to know about a company.

Square feet per employee might tell you something about how well the company treats its employees. But square feet of floor space just says, well, nothing. It's just useless information.

And I saw it over and over again.

I guess the cautionary tale is to be careful when sales people and marketeers give you impressive sounding stats like "we have over 200,000 Osminas of capacity in our Moscow facility". Think about what you want to know, ask for that data, and be very suspicious if you don't get a quick, direct answer... In units that make sense to you.

By the way, the results of my research were interesting. 60 of the top 100 companies in the Global Services top-100 list have, all in, 1,247,010 employees.

The top-12 companies (in terms of headcount) on that list of 60 employ 937,700 people.

The Pareto principle would predict that that the top 20% of companies in my list of 60 would employ 997,608 employees. That's a remarkably close proof of the principle.

No comments: