Wednesday, June 07, 2020

Getting Your Numbers Right

It’s good to see that Harvard Business Briefing is proffering some good advice on how not to be deceived by statistics and the numerical indicators so loved by today’s conventional leaders. It seems to me, however, that they’ve partly missed the point. It isn’t just that statistics are so widely misunderstood and misused. The real problem is the worship of quantification, and the accompanying belief that what can be quantified is, in some way, superior to what cannot.

Mathematics has become an essential tool in the more glamorous sciences like physics and astrophysics. Maybe because of that, it’s seen as uniquely “scientific,” at least in the popular mind. Or perhaps the culprit is the computer, which can handle numerical information in any amount, but is totally unable to process qualitative information. Either way, the result is the deification of numbers. Only the other day, I heard a corporate executive tell a radio reporter their company was well managed because: “we’ve been able to quantify our management approach.”

Mathematics is a language, albeit a uniquely rigorous and logical one. It’s a way of expressing relationships between things in ways that are far more precise than words can offer. Still, it remains a language: a tool to think with. Whatever can be quantified can be processed with mathematical and statistical tools. Used properly, this can reveal profound insights. Used childishly, which is how many companies use it, it’s like a three-year old declaiming Shakespeare: profound thinking reduced to gibberish.

Reality consists of some things that are quantifiable and many others that are not. Beauty, truth, goodness, honesty and joy are qualitative factors. They cannot be reduced to numbers. Nor can trust, service, satisfaction and loyalty. You can impose a numerical look to, say, customer satisfaction by means of a survey, but the outcome is never more than broadly approximate. Who knows whether all the people who rate your service 3 out of 5 mean the same thing? Or whether a rating of 2 or 4 is significantly different? Or whether the average rating from the survey means anything useful at all?

The lure of quantifiable approaches is based more on spurious simplicity, speed and the ability to avoid the tedious parts of number crunching by using computers. Numbers and quantitative indicators should never be seen as more than a starting point for inquiry that adds qualitative data to the mix. Making decisions for the purpose of changing a numerical indicator alone is plain dumb. You might as well ignore the road and the other traffic and drive with the sole intention of keeping rigidly to the speed limit.

Paradoxically, in the days when the numbers had to be calculated on paper, these approaches might have been more useful. At least doing the calculation personally forced people to look more closely at the data. Nowadays, the answer appears in a few nanoseconds and no one looks at what went into it or why.

Wise managers don’t reject quantifiable data, buy they don’t trust it either. They use it to spark questions and suggest areas for inquiry. It has no greater validity than purely qualitative data. Often it has less, if the data the calculations are based on is itself questionable. Many corporate numbers are Fool’s Gold. There are no numerically-based short-cuts to good leadership, as there are no other short-cuts to good leadership either. Good leadership takes time, careful thought and a deeply questioning state of mind; which probably accounts for its rarity in the typical “grab ’n go” organization.

Add to Technorati Favorites Stumble Upon Toolbar


Lee Iwan said...

Excellent article.

A must-read for anyone involved with business and performance statistics.

I have always felt that the current urgency to quantify and put numbers on every part of the business process, and then to concentrate and focus on managing these numbers just isn't right.

8:59 AM  
Carmine Coyote said...

Thanks, Lee.

9:29 AM  

Post a Comment

<< Home

Creative Commons License
This work is licensed under a  Creative Commons Attribution-NonCommercial-NoDerivs 2.5 License.