Thursday, July 12, 2007

Predicting the future

One of the first things I did for our startup was take a long hard look at the existing market for Business Intelligence (BI) software packages. To say you could probably throw a cloak over all of the marketing strategies in this space is generous, you probably only need a small towel (maybe a loin cloth).

But of all the product claims and positions I came across, one that struck me the most was the claim of 'predicting the future for your business'. Usually backed up by claims of 'predictive modeling', and 'forward looking metrics'.

Having used statistical tools for a lot of my career, I have no problem with a well specified model based on sound data that illuminates relationships and gets you thinking about how everything works together, and will possibly work in the future. However, having used these tools I know they are almost always overly simplified estimations of reality - especially business reality. And in no way do they predict bursts of creative thinking or 'out of left field' game changing shocks to your business. They just can't. Yet these things (that happen more often than you think) can render the best model obsolete.

Which is why, in good conscience, I could never claim a business tool is 'forward looking' or 'predictive' - it can only be that in a limited way at best. And at worst, can give you a dangerous false sense of security.

My advice is learn to live with uncertainty, and love it!

Of course, that tag probably isn't going to win you as many customers.

Dam I hate marketing.

Digg this
Sphere: Related Content


sk said...

Is there anything more annoying than claims that researchers, or worse, forecasters, who claim they can predict the future. I find that accurately describing the present is hard enough. Which is all to say: I'm with you in spirit and will follow your links with interest. I'm always open to new measurement techniques, though i bring a skeptical eye to them all. Curious, though, any experience (good or bad) with prediction markets? And by the by, thanks for the early post to my young blog as well.

Paul Soldera said...

Thanks sk. I like your blog. I think we're pretty much on the same page.

And yes, I'm naturally skeptical about all measurement techniques, and usually try to look at choices in terms of 'least error' rather than 'most accurate' - technically the same but psychologically vastly diff.

Never had a chance to play around with prediction markets - they seem to be very in vogue right now. Taleb's book (The Black Swan - that I talk about in an earlier post) is perhaps the best example of 'prediction skepticism' I have ever read. He touches on prediction markets, but I just can't find the reference (o why can't book contents be searchable?). I'd give chapter ten a skim - The Scandal of Prediction.