Consider a typical management report showing % change in revenue last week vs. same week a year ago. Is this change meaningful? That is what management always assumes, and demands an explanation for it. But what if the change is just due to chance (the normal variation in weekly sales)? You don’t know this until you view the whole context of weekly sales history. Graphical presentation gives you a sense of the randomness and variation in the data that is missing from simple point-to-point comparisons. I recall the horrors of sitting through meetings where executives demanded explanations (why did sales go up? why did sales go down?) for what was simply RANDOM VARIATION. How much time do companies waste on this ridiculous exercise?

Martin pointed me to Donald Wheeler’s book several years ago, and it has been a constant companion. It provides a savage critique of traditional management reporting, along with lots of practical ways to improve decision making by presenting data in the appropriate visual context.