Monday, August 8, 2011

A problem with some analytics practices

I like Analytics, but it complements, doesn't replace, technical business knowledge.

Consider the task of loading sand onto a truck. A solution-focussed person sees a worker with a shovel, then optimizes that and creates a robot that shovels better than any human.

This is solving the wrong problem with advanced technology: a better solution is to have a conveyor belt, a vacuum-cleaner like appliance, or better yet put the sand in a funnel with a servo-controlled spout. (These are all technological solutions borne of a problem-focussed rather than solution-focussed way to look at the world.)

And this is the problem with some current Analytics approaches: they apply a lot of brain power to solve problems using an amateur approach (i.e. make a robot to shovel sand), as opposed to actually learning from the field that has studied the problem for a long time. (By some I mean a minority, though a noisy one; in my opinion – never humble ☺ – the rise of Analytics is one of the three best things to happen to management in the last ten years.)

When a principal from a Analytics startup proudly tweets the equivalent of "now we no longer have to shovel sand by hand" as if this was a great discovery, it's clear that there's a need for professionals who understand technical business material.*

(Some people think that business material is "common sense," at which point I ask them for their common sense way to value a complex derivative, to figure out the core resources of a competitor, to brief an advertising agency, to determine the appropriately-loaded cost for a decision, to reorganize a continuous flow production process given a change in preventative maintenance schedule, or to program a full media-supported staggered promotion through multiple distribution channels.)

For some reason a sizable fraction of people working in Analytics have an acute case of Dunning-Kruger's. Smart people, acting dumb.

-- -- -- --
* To avoid embarrassing the person in question, I don't link to the tweet, but it was epic in its ignorance; it suggested that now we could stop doing something that hasn't been done by anyone with minimal marketing skills since the 70s. And anyone who had read a intro marketing textbook would have known that.

Observation: Serendipitously, when I was editing this post, iTunes chose to play Pink Floyd's Another Brick in the Wall (Part 2), the chorus of which is "We don't need no education!"