Skip to content

Since Data Scientists Don’t Do Business Decisions – Businesses Need Decision Scientists

Erik Larson Jan 18, 2022 9:00:00 AM

Twenty years ago, there was a great reckoning in the market research industry. After spending decades mastering the collection and analysis of survey data, the banality of research-backed statements like “consumers don’t like unhealthy products” belied the promise of consumer understanding. Instead of actionable insights, business leaders received detailed reports filled with charts and tables providing statistically proven support for research findings that did little to help decision-makers figure out what to do.

So, the market research industry transformed itself into an insight industry over the past twenty years. To meet the promise of consumer understanding, market researchers are now focusing on applying business knowledge and synthesizing data to drive better insights and ultimately better decisions.

A similar reckoning is at hand for data scientists and the AI models they create. Data scientists are mathematical programmer geeks that can work near-miracles with massive complex data sets. But their focus on complex data is much like the old market research focus on surveying consumers – without a clear focus on applying business knowledge to support decision-making, the results are often as banal as the old market research reports. Improving data and text mining techniques and renaming them Machine Learning, then piling on more data and calling it Big Data, doesn’t change that fundamental problem.

Today’s data scientists can learn a lot from companies like Johnson & Johnson, Colgate, and Bayer. These leaders have successfully transformed their market research functions into insight generators and decision enablers by combining analytical tools with the business skills required to drive better decisions.

Data scientists could follow a similar path, but what if we took a much bigger leap?

Leaping to Decision Scientists

According to Merriam-Webster, science is “knowledge about or study of the natural world based on facts learned through experiments and observation.”

Applying that definition to data science highlights the critical disconnect. Data scientists do not exist inside companies to study data – they are there to generate knowledge to help business decision-makers make better decisions. By itself, collecting more data and analyzing it faster does not result in objective knowledge that can be relied on for better decision-making. The science of data is not the science of business.

Now imagine the evolution of a new decision scientist role. Since decision-making effectiveness is almost perfectly correlated with business performance, a focus on studying and building knowledge about business decisions and decision-making will directly advance business goals.

Rather than studying how to generate, process and analyze generalized business data, tomorrow’s decision scientists will focus on tracking, understanding and improving business decision-making.


Data Scientist

Decision Scientist

How can we create a complete enterprise data lake?

How can we create a complete enterprise decision map?

What data is available for analysis?

What insights do decision-makers need?

How can we create more accurate analytical models?

How can we create more predictive decision-making models?

How effective is our analysis process?

How effective is our decision-making process?


Decision scientists will live in an exciting new world where the current lack of scientific knowledge about business decision-making means major discoveries will happen every day. Even better, every decision is a natural experiment, setting the stage for incredibly rapid learning. Finally, businesses will benefit directly by applying new understanding to make better, faster decisions.

Decision scientists will focus on mapping the structure of business decisions and understanding the process business people use to make decisions, including the mix of data, experience and intuition needed to guide recommendations and decisions that deliver business value. As this white paper (PDF) explains, the new decision-back approach is a valuable catalyst for this vital work.

Decision Scientists Will Decode the DNA of Business

By mapping the decisions that drive a business and then tracking those decisions’ inputs and outputs, decision scientists can bring decision-centric, business-focused scientific direction to the disconnected layers of today’s data-centric world. Much like a complete map of our DNA highlights the genes and interventions important for better health, a comprehensive map of our decisions can focus efforts to drive the most business impact.

  • Decisions – Who makes decisions, and how are their choices made? How can we measure decision success and map that back to the optimal mix of decision inputs and most effective decision processes?
  • Business Issues – How can decisions be used to model a predictable connection between the inputs and outputs of a business? How are business goals connected, via decisions, from the CEO to the front-line manager?
  • Insights – What is the scientific definition of an insight? What is the best role for human synthesis in generating insights, and can that human role be modeled and even automated for faster, more efficient insight generation?
  • Analytics – What analysis and mathematical models are needed to support critical decisions? How can we align top-down analysis that starts with key decisions and business issues with bottom-up analysis that starts with key variables and data sources?

Decision scientists will shift the focus from the science of inputs (data and analysis) to the science of outputs (recommendations and decisions). Of course, data science will continue as an important activity, except now it will be directed not only by the technical challenges of complex data sets but also by the complex needs of decision-makers. This shift will significantly improve the business value of data and analytics, making tomorrow’s decision scientists an indispensable business resource.