by Andrew White | March 28, 2019 | Comments Off on The Role of Data and Reason in being Data-Driven
It’s all the rage- we need to be ‘data driven’. Unfortunately important ideas and advice are often dumbed-down or simplified to singular messages for ease of consumption and this can lead to yet more problems. To be data driven is a great example. What is such a thing?
Conceptually to be data driven sounds like a situation whereby a decision ought to have its roots founded in data. In this case, data really means all manner of information from raw data, analyses of that data (e.g. an analytic or metric), or some other more sophisticated representative of a model or algorithm. The difficult part here is with ‘root’ and ‘founded in’.
Does data driven mean that whatever the data says, goes? No. The idea, I believe, is that we should seek data, even new data in some cases, to help augment the decision making process. In some cases the (new) data will be overwhelming in reinforcing one choice in a decision over another (e.g. efficiency). In other cases the (new) data will proved inconclusive. In yet other situations the (new) data may change the decision itself and the process in which it sits, completely negating the choices of the original decision (e.g. transformation). Both results are still data-driven: it all comes down to context and interpretation.
There was an interesting article in today’s US print edition of the Financial Times. It was an Opinion piece and it was written by Anjana Ahuja, a science commentator. It was called, Beware making a fetish of an arbitrary number. The article highlights a dialog between scientists and pundits who use ‘statistical significance’ as a proof of a fact, argument or casual relationship between factors. In tests every day, if something is found to be statistically significant then it must surely be. The problem is this is not necessarily so.
You can easily conceive of a test that is not valid, that can actually be developed to explain or prove an argument – See lies, damned lies and statistics. The point is that the test for statistical significance was never meant to be a substitute for fact or reason. Statistical significance is a flag, a signal, yet another piece of data. It is not a pseudonym for fact. We might even argue that facts don’t actually exist, but that’s another story for Gary and me to work on later…
So our data driven efforts should never make the mistake that data speaks for us. Data, in all its myriad colors, insights and shapes, is just that- data describing something. We still should be aware of who asked the question being tested, why was the question asked, and more importantly, who is the beneficiary (or the object) that is now justified (or vilified) by the data? As I have so often heard, ‘follow the money’ and this will highlight who is setting the direction in which data is being driven.
Comments or opinions expressed on this blog are those of the individual contributors only, and do not necessarily represent the views of Gartner, Inc. or its management. Readers may copy and redistribute blog postings on other blogs, or otherwise for private, non-commercial or journalistic purposes, with attribution to Gartner. This content may not be used for any other purposes in any other formats or media. The content on this blog is provided on an "as-is" basis. Gartner shall not be liable for any damages whatsoever arising out of the content or use of this blog.