This is my last serious entry before I head off for my week in Prague, so I thought I’d touch on something of which I have direct experience and which I believe can be extremely unfair to those who fall under its gaze.

That topic is Management Information (MI). When applied to the contact centre, we’re talking fine-grained, fairly complex statistical analysis of Key Performance Indicators and an often obsessive worship of numbers.

The problem, I believe, is that quite often the kinds of people setting up these measurement systems are not always the sorts of people who understand statistics or workflow. For instance, once upon a time I worked with a manager who was in charge of identifying skills utilisation. She firmly believed that one could add fractions by adding the numerators together, the denominators together, and dividing the two. In other words:

Her Logic: 1/2 + 1/4 == (1+1)/(2+4) == 2/6 == 1/3
Correctly: 1/2 + 1/4 == 2/4 + 1/4 == (2+1)/4 == 3/4

You can see where this is leading. This otherwise seemingly intelligent individual was in charge of defining MI for the competency analysis wing of a large accountancy. Peoples jobs, their futures, would be subject to the relentless implications of a single person’s lack of understanding.

Of course, being the smart-arse that I was (young geek punk, heh heh) I spotted the flaw immediately. Being a contract developer at the time, my (justified) criticism was constructive and relayed in a friendly and helpful fashion. You know what? She didn’t listen. Despite protestations, said package was completed, and launched to an innocent world.

To date, I reckon this flawed logic has probably jiggered the performance and competency assessments of tens of thousands of people. All because of a rather blinkered and mathematically challenged manager who knew best.

You might think that this is a fairly exceptional case, but believe me, it isn’t. Most medium-to-large call centres use some kind of performance measurements as a basis to set targets, measure effectiveness and ultimately in many cases set bonuses for the ordinary men and women.

Yet, looking underneath the bonnet of many MI systems, statistical irrelevance, broken logic and data vagrancy* is rife. There are two many people working in such systems who have sort of ‘fallen into it’ in some way. Marketing types who don’t know what a standard deviation is, far less how to code one into a spreadsheet; fast-track types whose rapid ascendance belies no great competence at anything, you get the picture.

The basic problem is that so much of performance measurement has not been rigorously thought out. Not enough is done to ensure that the numbers are meaningful. Often, a small mathematical error may have enormous repercussions.

In short, it’s scary!

Nevertheless, companies treat the numbers as gospel. They don’t really want to go to the effort of really talking to their staff, seeing the real performance. Numbers are easy; they fit into corporate spreadsheets which make shiny pie-charts. They look complex enough to make middle management feel clever.

It’s a bad thing…

So, what can we do? I wish I had the answer. The problem is people, at the end of the day. To paraphrase Yourdon, “there are no silver bullets". In other words, there is no easy answer - it is a non-trivial task yet if we are to avoid making serious decisions on the back of spurious data, we must get the right people doing serious quality-analysis of the constituent parts which make up MI calculations.

Your comments would be more than welcome; please add them via the link below.

John

* Vagrant Data - old data which hangs about, isn’t really welcome and smells a bit funny. There’s a lot of Vagrant Data about, sadly.