BI is About Questions, not Answers

One good question can give rise to several layers of answers, can inspire decades-long searches for solutions, can generate wholeness fields of inquiry, and can prompt changes in entrenched thinking. Answers, on the other hand, often end the process.

Stuart Firestein, Ignorance: How it Drives Science

Recently I had the chance to visit with a member of the Utah House of Representatives, Val Peterson. Val related his experience last legislative session about the process and eventual decision related to the Sales Tax on Food. Those involved in the discussion had spent weeks reviewing charts and graphs and consuming volumes on information looking for the data that would guide them to the best decision. Taxes and government are always sticky. There are always winners and losers. Ultimately they focused on what actions could remove the volatility in the Sales Taxes. It was determined that any efforts to change the sales tax on food would NOT impact the overall volatility as it accounts for less than 10% of the Total Sales Tax Revenues in the State.

I asked Val if they could have come to that conclusion without going through the journey, without spending the two weeks consuming all the data. He didn't think they could have. From my assessment they needed the data to frame the discussion to clarify they were asking. Was the experience lost? Was the legislators energy wasted? NO! They now have more context for the next discussion about Sales Taxes.

In today's world of Analytics and "Intelligence" we have to allow consumers of our goods to have their journey. Those asking questions need to be able to develop a level of expertise in a safe, non-judgmental zone. Everyone has ego, especially executives. If the Analyst gives the answer then they end the process and diminish their own ability to contribute to the organization. An Analysts value increases the more varied the questions are that get asked, whether the answer exists or not. An unanswered question becomes a driver for change.

There are professional analysts that spend more time asking questions than providing solutions. Growing up I had a neighbor who was a very successful management consultant. He used to tell me the two things he took with him everywhere were his laptop and golf clubs. The laptop to write up the recommendations for his clients and the golf clubs to ask the questions. You see people would talk on the golf course about things and open in a way they wouldn't in the office - his questions could go deeper and reveal more outside the setting of the office. His success as a consultant increased when he learned how to get to ask more questions.

There is something statistically comfortable about asking more question, the more you ask, the higher chance you have of asking a meaningful question, right? Why then in the BI space do we spend so much effort to limit the questions asked and explain the "Single Version of the Truth?"

I come from a religious background, plus I have done graduate academic study. From my perspective there are two ways to define truth:

  1. Divinely decreed
  2. Peer-reviewed, validated, reproduced results

In the absence of a Deity explaining the "Truth," it appears we need to spend more time socializing the data and the processes around it and less time explaining why "my number is right and their number is wrong."