Big data is the big thing with big corps. Walmart, Target, Amazon, Shoppers, and Facebook know more than you think about what you buy and therefore who you are. They are hiring mathematicians in their marketing departments. The ads on your Google page are tailored to wherever you roam on the Internet. It’s not Big Brother anymore, it’s little mouse. But lately, there have been whisperings that these mathematical models are not as sophisticated as purported. Remember the hedge fund debacle? Perhaps the shaping of your behaviour is more complicated than the math moguls predict.
I’ve had my own experience with small data.
We have an elegant and efficient in-house inventory control system for the 30
thousand or so titles that need to be manipulated. In a normal month the
algorithm that is used works just fine. What we have sold is invisibly crunched
to tell me what we need to order. But this does not work at Christmas because, as you know, the sales velocity in the week before Christmas is too quick to allow us to reorder in time. So I have to predict, and the only way to
predict is to look at what’s selling daily and make some intuitive decisions.
Now, those decisions aren’t just intuitive; I study our sales every night and
try to keep abreast of who is going to be on the CBC for the week, what
Chapters is out of, what a certain bookseller is handselling. It’s a lot of
work but actually keeps my brain in the game.
But here’s the rub. It’s extremely hard to
predict buying patterns. In the ten days before Christmas, one day our
bestseller might have been I’m Your Man, the new Leonard Cohen title, the next
day we might sell none of it and the top buy might be Dear Life by Alice Munro.
I sometimes feel that there is an unseen principle in the world based on the
fact that if you prepare, what you thought would happen won’t, and if you don’t
prepare, it will. There must be a name for this. Perhaps this is the same thing
that is causing all of those mathematical minds to shake their heads. And that
is a good thing!
- Barb
- Barb
No comments:
Post a Comment