A colleague of mine made an interesting observation today relative to a discussion about using data to draw conclusions.   “If you torture the data long enough, it will confess to anything”

There must be a reality

This is an interesting twist of words.  There certainly is some truth to the implication that a given set of data can support a variety of truths.

The reality, however, is that data does not speak, only the human interpreters do that. The point is really about salesmanship.  Interpretation of data has a subjective component.

My own personal experience with this comes from years of developing business cases.

I used to brag (cynical me) that if I was advised the organizational threshold for rate of return (aka the hurdle rate) I could document a business case that would result in approval to move ahead, every time.

The sad part of this boast is that I have been crafting business cases since 1980 ( for various organizations) and in over 35 years I have never had one business case proposal rejected. (I did not count them but we are talking about lots including one for several hundred million)

Back to the Point

Based on my own experience of torturing data, truth is, getting the data to say anything was not about the data but all about my ability to interpret, organize even shape the story the data told.

The data, well it was silent putty in my hands.

I do strongly recommend questioning and validating proposals, however my advice would be to focus on the story as oppossed to the data itself.