For example, most data based analysis suffers from this problem – “given large amounts of data, whatever pattern you want to see, it will show up”. There is I believe a solid theory behind this, dont know all the details.

Then there is this problem (which you have already stated in the blog) – that users/marketers are using the tools without understanding the assumptions behind them. For example, one can use Monte Carlo to see the effects of a project plan, but “to take that as the sole basis for a decision” is madness.

Then there is data itself. You probably have no idea behind the scenes, but I have, of the amount of cleaning, re-arrangement, and sometimes plain shortcuts that are used to re-align large amounts o disparate data for an analytics study. Most often it is the incompleteness, wrong sampling, wrong seeding – and all sorts of hidden things that lay bare the lie of analytics.

It is nice to use technology, when it is mature enough, but even then, only as an assister. So excessive use of technology for things it is not yet ready for is itself an unbalanced response.

P.S : -) I just noticed you also used the word “bean counter”, I was hesitant to use it, but now feel more confident.

I personally think that most of those calculations are lagging indicators. That is to say that they show the result after changes have been made. It is up to the creative people to find the vision and strategy and it is up to these calculations to see if you are still on track with what you envisioned. It should be a balance between the two. We want you to bring in the most money and create the most overall value for the company with using the least amount of resources possible.