Tag Archives: Statistics

Analytics Defined

For this post, I provide a link to my article entitled ‘Analytics Defined: A Conceptual Framework’ that was published in the June, 2016 issue of ‘ORMS Today’. A PDF copy of the article has been placed on the Indiana University – Purdue University Fort Wayne website. The link follows:


The core ideas in this article were first presented in four posts in this blog:

These ideas were further developed for a panel discussion on 11/1/15 (https://cld.bz/KAj90ao#104/z [bottom of page]) and an invited presentation on 11/3/15 (https://cld.bz/KAj90ao#269/z) at the INFORMS National Meeting in Philadelphia.

Analytics And String Theory

The notion has arisen that analytics is an emerging field that represents a convergence of the quantitative decision sciences. For example, in a 2015 paper in Interfaces, the authors state: “…the emerging definition of analytics as a field of expertise that subsumes OR.”. If true, such a convergence would be a surprising and remarkable development, as it would represent a dramatic reversal of the trend in human history toward specialization. It is worthwhile therefore, to examine this idea, and consider its logical consequences.

A convergence of the decision sciences into a single field would imply that this new field would contain all the knowledge and methods currently included in the decision sciences, and would lead to one of four possibilities being true.

A new unifying theory is developed. Currently, hundreds of theoretical physicists are working on the development of string theory, which they hope will mathematically unify quantum mechanics, particle physics, and gravity. Unlike string theory, there is no history of an analytics theory going back to the 1960’s, no founders of an analytics theory, no seminal papers introducing an analytics theory, and no conferences where an analytics theory is discussed.

In fact, in a 2014 paper in the European Journal of Operational Research, the authors found only 15 articles in theory oriented journals that were listed in the International Abstracts in Operations Research database with the term analytics in the title or abstract. Moreover, the types of analysis that analytics thought leaders offer as examples of analytics, such as advanced statistical analysis, econometrics, and optimization, are actually examples of existing methods from existing disciplines. (See for example, ‘Competing on Analytics’.) So, while string theory may provide a unifying ‘theory of everything’ for physics, there is no evidence that a unifying theory of the decision sciences exists, or any reason to believe that one could be developed.

Analytics practitioners must master all the knowledge and methods of the decision sciences. Without a new simplifying mathematical theory, mastering the knowledge and methods of all the decision sciences would require five or six Ph.D.s and 80 to 100 years of experience. Since the human life span is insufficient to accomplish this, we can reject this possibility.

Analytics practitioners produce simplistic or superficial work. Lacking a new simplifying mathematical theory, or a sufficient life span, analytics generalists would be unable to perform at a level comparable to current experts in the decision sciences. I will assume that those who believe that analytics is an emerging field, those who would employ analytics practitioners, and everyone else, will view this outcome as a highly negative development, and therefore, will not accept it.

Analytics is practiced by individuals specializing in different areas. Since the other possibilities are impossible or undesirable, we must conclude that specialization is necessary.

Allow me to list these specialties for you: statistics, computer science, operations research, industrial engineering, economics…

Confusion Over Analytics

The term analytics emerged in November, 2005. In the chart shown below, the relative number of Google searches on the term analytics from 1/1/04 to the present are displayed.

Chart Showing Searches For Analytics
Searches For Analytics

In addition to the dramatic growth in the use of the term analytics, there has been a proliferation in the way it is used, with phrases such as text analytics and healthcare analytics common.

Numerous definitions of analytics have been proposed, but consensus and clarity have been elusive:

  • The lead article in the December, 2013 issue of OR/MS Today was entitled ‘The Evolution Of Analytics’. In the body of the article, the authors presented a 200 year history of statistics!
  • In recent surveys, operations research professionals have expressed widely differing views on the relationship between operations research and analytics.

However, there is one point on which there is agreement — analytics is related to many different disciplines:

  • Davenport, Cohen and Jackson, in the May 2005 research report ‘Competing on Analytics’ mention statistics, operations research, industrial engineering, econometrics, and mathematical modeling as examples of analytics.
  • Rahul Saxena, co-author of the December, 2012 book ‘Business Analytics’, on slide #5 of a slideshare presentation, lists 14 disciplines as being antecedents of analytics. The list includes Business Intelligence, Computer Science, Statistics, Operations Research, Industrial Engineering, and Finance Planning & Analysis.

A Simple Explanation

How can we explain the sudden appearance of the term analytics and its connection to so many different disciplines? Did a new meta-discipline, representing a new problem solving paradigm, suddenly emerge?

Let me offer a simpler, and more plausible explanation. In November, 2005 a new concept emerged, and went viral: the idea that it is useful to be able to group together several distinct disciplines, and refer to them collectively with a single term.

This concept makes it possible for statistics, computer science and operations research to be represented by the word analytics in the same way that biology, chemistry, and geology are represented by the word science. Organizations can now consider quantitative decision sciences collectively for purposes of planning and resource allocation.

In current practice, the term analytics is used to represent two different groupings of disciplines:

  • a base grouping (statistics and computer science) – e.g., data analytics;
  • an extended grouping (all quantitative decision sciences) – e.g., business analytics.


When analytics is understood to be a conceptual grouping of quantitative decision sciences, confusion disappears:

  • the sudden emergence of analytics and its relationship to other disciplines is explained simply and logically;
  • the varied usage of the term analytics becomes understandable;
  • it is not necessary to postulate the emergence of a new meta-discipline;
  • there is an additional way of thinking about the quantitative decision sciences.