This article originally appeared in [https://techcrunch.com/2015/12/04/decision-science/] and belongs to the creators.
Nearly everyone is familiar with the DirecTV commercials in which actor Rob Lowe and NFL quarterbacks Eli Manning and Tony Romo appear as unflattering alter-egos of themselves: “peaked in high school” Rob Lowe, “bad comedian” Eli Manning, “artsy craftsy” Tony Romo, etc.
The adoption of data science in companies like Uber, Netflix and Amazon evokes a similarly striking contrast with legacy companies such as Yellow Cab, Blockbuster and Sears/Kmart.
Data science vs decision science
Data science is often used in conjunction with many other science-related terms — algorithms, machine learning, artificial intelligence and predictive analytics. All of these terms are intended to indicate when computers are being used to detect signals or patterns in data that drive better business outcomes.
While data science is perhaps the most broadly used term, “decision science” seems like the more fitting description of how machines are assisting business leaders in solving problems that have traditionally relied on human judgment, intuition and experience.
It may not be the sexiest phrase in the world — I’ve never seen it in any marketing materials — but “decision science” aptly encapsulates how computers are helping to systematically identify risks and rewards pertinent to making a business decision.
Decision science incorporates an economic framework — a consistent, rational and objective system to “price” each possible outcome, taking into account risks and rewards. It is simply a better way to make decisions.
Behavioral Science, Democratized
We make 35,000 decisions each day, often in environments that aren’t conducive to making sound choices.
At TDL, we work with organizations in the public and private sectors—from new startups, to governments, to established players like the Gates Foundation—to debias decision-making and create better outcomes for everyone.
Why use decision science?
Such a framework allows us to separate the bias and pitfalls often introduced by emotion and ego that are otherwise impossible to overcome. Remember the hot water Uber got into last year over its “God View,” which allowed the company’s staff to track both Uber vehicles and customers?
Calculating risks and rewards for all possible outcomes, taking into account all available data, is time-consuming. Computers can do these computations far better than humans. Machine learning techniques are now allowing computers to discerns patterns in very noisy data.
"Decision science is most effective when it’s treated as, well, science."
As a result, computers are now better able to cull massive amounts of complex data and act as powerful scenario-generation engines, identifying the risks, rewards and uncertainties in a variety of important decisions. With this analysis at his or her fingertips, a decision-maker can confidently overlay judgment and experience to make better decisions.
Getting the most out of decision science
As an aside, decision science is most effective when it’s treated as, well, science. It’s a mistake for companies to invest massive amounts in building data warehouses to try to get a single “view” of data without really knowing what to do next.
Basic scientific principles call for doing some experiments and testing outcomes before you think about scaling. A scientific approach calls for testing with sample data, validating it and then integrating and expanding.
If you’ve seen the TV commercials featuring Bob Dylan talking to a computer, you’re familiar with a prime example of how computers are getting better at pattern matching and enabling better decision science. The ads are for IBM’s Watson, the “Jeopardy”-winning technology that takes in data from a multitude of sources and interprets it to expose patterns, connections and insights — such as the best course of treatment for a cancer patient.
The AI Governance Challenge
Software company SAS is another notable player in pushing the edges of decision science. Google and Microsoft offer modeling capabilities, though they tend to suffer from requiring too much work to adapt to the context/relevance in a functional area of the enterprise, such as sales.
Thanks to decision science, things that traditionally have been considered difficult or impossible to predict are proving, with the right tools, to be forecastable after all.
Using decision science to predict the unpredictable
Take HR, for example. HR departments typically have lagged behind the rest of their organizations when it comes to harnessing data, but more and more are starting to use data analytics to better find the right person for the job, as well as retain employees. Talent Science is an application vendor in this space.
“Predictive policing” has become one of the hottest emerging areas in law enforcement. According to the National Institute of Justice, “predictive policing leverages computer models, such as those used in the business industry to anticipate how market conditions or industry trends will evolve over time, for law enforcement purposes, namely anticipating likely crime events and informing actions to prevent crime.” Companies like PredPol and HunchLab aim to make real the crime-stopping technologies predicted in the 2002 movie Minority Report.
Simply put, decision science is a marriage of technology and business perspective to solve complex challenges. Executives who don’t catch on may find themselves with a DirecTV-style alter ego: “Hi, I’m yesterday’s business leader.”