Exception reporting, a.k.a. alerting, is a common response to the “No news is good news” attitude of stressed-out managers. Perhaps, it is a sign of pessimism to say that when top managers receive a message, it has to be something bad. Wait…or is it optimism in the sense that if they haven’t heard anything, nothing THAT bad could have happened? Maybe, too, it is just the desire to receive less – or even no – news for a change in light of the information overload that managers battle each day.
At any rate, exception reporting is designed to suppress information. Instead of regular intervals, the report consumer should only receive information in certain circumstances that are defined in advance. Thresholds, however, are dangerous in this lump sum approach. If a variance exceeding 5% is relevant, does that mean that a 4.98% variance is not? Which cutoff should we make on which level of aggregation?
The information systems expert Norbert Szyperski already warned of these problems back in 1978:
“Fixed thresholds are dangerous in combination with the fiction of management by exception. Managers should be curious, in other words, should proactively seek out new relationships among information and not just dose off attentively like a person monitoring a switchboard.” (translated from Mertens/Griese, Integrierte Informationsverarbeitung 2, Wiesbaden 2002)
My personal preference? Design standard reports that are so attractive and information dense that the readers enjoy receiving them and would never even consider trying to suppress them.