The InformationWeek 500 was published today for 2009. The study describes IT priorities, spending and directions for a lot of companies and then in some sense ranks the IT shops. Other organizations like Gartner Group do the same thing as well as other magazines like CIO Magazine.
I’ve got a fairly strong statistics background and I have to say, it really bothers me that these studies will always show the average IT spend as a percent of revenue by industry, state, group or whatever. But they never show any 2nd order statistics. It would really be helpful if they go that next step and tell us the spread about the mean. If the Manufacturing group in the study has a average IT spend as percent of revenue of 1.8%, I’d really like to know the spread around the average. Is the spread tight or wide? That would really give me a deeper understanding of the data. Specifically, if I’m away from that ‘average’ is that unusal or not? If the variance is large, I’ll draw a different conclusion from a tight variance.
So to those of you doing these rankings, studies, evaluations, please start putting 2nd order statistics in your results. Thanks from a CIO.