Decision Sciences is the discipline of solving business problems by
using first principles with a mix of mathematics, business knowledge,
and technology. Familiar examples include data mining and predictive
analytics. It emphasizes the adoption of a structured, hypothesis-driven
approach to break down the problem, analyze data, and make business
Executives often make poor decisions when they rely on intuition and try
to predict the future based on gut feel. Decision Sciences provide
effective methods to examine alternatives, separate facts from biases,
and improve decision making. Many Fortune 500 companies and industry
pioneers are leveraging data-based methods to identify efficient ways to
do work, adapt to swift changes in the economic environment, and react
to changing customer needs. But despite the growing acceptance of
Decision Sciences, myths about the field persist in the marketplace.
Myth: Decision Sciences requires clean data.
Reality: It is true that building enterprise data
warehouses would help institutionalize data-driven decision making, but
it is not a prerequisite to getting started. “Dirty data" -- raw pieces
of information that are not cleansed, are stored in disparate sources
and contain missing values and outliers -- can still yield value. And
processing data can induce bias and skew results. The most important
requirements for getting your organizations to start making data driven
decisions are raw data and analysts who are capable of operating with a
Myth: Decision Sciences raise the risks of breaching data privacy and security.
Reality: Techniques to mask customer sensitive
information transform data without altering the intrinsic patterns that
exist and are necessary for analyses.
Myth: Decision Sciences requires a large budget and expensive off-the-shelf tools.
Reality: Sophisticated analytics software is not a
pre-requisite; first principles-based thinking and a hypothesis-driven
mindset are the key elements. Force-fitting generic models to the tools
available in the market will give incorrect to average results. A
qualified analyst can cut and slice the information available without
any expensive software to yield insights that guide businesses to make
the right decisions.
Myth: Decision Sciences requires highly trained personnel.
Reality: One does not need a PhD, formal training in
statistics or proficiency in programming logic to get started on
decision sciences. Understanding the information available, applying it
in the right context and asking the right set of questions are the keys
to successfully practicing Decision Sciences. To quote a famous
statistician John Tukey, “An approximate answer to the right problem is
worth a great deal more than an exact answer to an approximate problem.”
Demystifying Decision Sciences helps an enterprise open doors - each
open door leads to actionable insights and new doors. It is a
constantly evolving maze and business managers need to use the right
combination of business, math and technology to prioritize issues and
open the right doors - making it more ubiquitous and guiding all
Firms that approach Decision Sciences expecting an "aha!” moment will
likely be disappointed at the lack of dramatic insights or incedible new
market opportunities. The reality is that Decision Sciences provides
improvement and focus to existing business processes, not million-dollar
Dhiraj Rajaram is the founder and CEO of Mu Sigma, an analytics services company that helps clients institutionalize data-driven decision making.