Sign up FAST! Login

The Journey from Raw Data Sets to High-Value Information for Decision Makers


http://anametrix.com/blog/2014/03/13/the-journey-from-raw-data-sets-to-high-value-information-for-decision-makers/

rare steak

Who in the organization deals with data in what format? 

  • Software engineers: bits and bytes
  • DBAs: context (and the bits and bytes)
  • Data analysts:  information
  • Decision makers: knowledge

Notice that as we cross each interface, additional details, metadata, or analysis are required for that next group to use the data as depicted. Yet we expect the system to function as if all the necessary pieces of information are available at each step. It’s difficult for software engineers, for example, to write code to gather data, along with all the additional details and metadata needed to support tasks further along the process, without the perspectives of the latter three groups.  It could be summed up almost as this:  collect data (numbers and text) without knowing why or what will be done with them. The other groups are in somewhat similar binds, often lacking input into the previous step and understanding of the next step. Ideally there would be a feedback path (not necessarily a loop) from the decision makers back through the data analysts and DBAs, and on to the software engineers, all of whom can support the groups before and after them by sharing their perspectives, as well as requirements and even usage cases. 

Analytical projects can fail for many different reasons. Even good data and good analysis can fail to produce useful insights for decision makers because of how requests and projects get translated between data analysts and the marketing team or other business people. For this final connection, I’ve seen at least two prominent issues cause difficulties. The first of those, and most common, is one that many non-analysts will recognize: a presentation of an analysis focusing on modeling results and details, and not on the meaning or interpretation of those results in the context of business goals. While certain conclusions might seem obvious to a data analyst, the presentation is in the wrong “language,” so to speak, for the audience. The second issue occurs when decision makers have strong convictions about certain preconceived notions. In An Essay on Criticism, Alexander Pope explained this well when he wrote,  “A little learning is a dangerous thing.” In such cases, a decision maker might determine the direction of the analysis, e.g., “I just want a linear regression done on the data” or “Give me sales forecasts for the next six months.” Unfortunately, linear regressions are often the wrong approach and sometimes forecasts are requested when the actual goal is to understand drivers of performance. 

It is important for the business team to delineate the decision-making process and determine where in that process more information is desired, and then describe what type of information would be useful. The data analysts should study the process and determine what data and which analyses, if any, are appropriate for providing such information. In addition, the data analysts should offer suggestions for other places in the process and other types of information that data and analysis could support with evidence-based results. The business team should be very clear about what they are trying to gain from the analysis and how they intend to use the information. In turn, the data analysts should present results of their analyses with a focus on providing the business team with what they requested, but without omitting vital elements, often of a statistical nature, that could constrain or moderate the usage of the results. That way the data get analyzed and the results interpreted and translated into knowledge that supports decisions consistent with the evidence presented. This interface is the most complex of the three since it is the least amenable to specification sheets or requirements documents, and necessitates ongoing, iterative interactions between business team and data analysts. Read more: http://anametrix.com/blog/2014/03/13/the-journey-from-raw-data-sets-to-high-value-information-for-decision-makers/#ixzz2vuwcXXCK

Stashed in:

To save this post, select a stash from drop-down menu or type in a new one:

Raw data is like raw meat. We need spices, cooking utensils, and cooking appliances.

You May Also Like: