You won’t find true signals of changes in performance by looking at month-to-month comparisons, or trendlines, or moving averages. The signals you really need to know about, the only signals that you ought to respond to, are revealed through one particular graph only.

There is a statistically proven sweet spot for how many KPI data points you need before you can start using them and accurately interpreting signals from them.

They're so hot right now, sparklines. They thankfully bring some sanity to performance dashboards (goodbye goofy gauges). But they do have a limitation no-one seems to be talking about. (more…)

A typical way to set targets is to calculate a target for the year, and then apportion it to each month in that year. The idea is to keep everyone focused on achieving the target, so increase the likelihood the annual target will be met. But it doesn't work. ...

KPI thresholds are not improvement tools; they make people react in a rush to quick fixes and gaming that end up making performance worse. (more…)

I fundamentally disagree that we should be striving to hit our KPI targets each month (or whatever frequency we measure at). Natural variability guarantees it's purely a fluke if we hit a target in any given month. We need targets to make more sense.

Lag measures cannot be improved directly; only through changing one or more of their lead measures. And there are six basic steps to make this happen. (more&h...

Analytics doesn't magically produce insights from data like rabbits out of hats. To be worthwhile, analytics must be directed by deliberately chosen and strategically important business questions.