Creating a Data-Driven Culture

Scott Busse


April 1, 2012

Empirical evidence shows that a data-driven culture can lead to better, faster decision making, greater customer insights, fewer mistakes and ultimately a competitive advantage.

For example, through his “Moneyball” approach, Billy Beane instilled a data-driven culture as general manager of Major League Baseball’s Oakland Athletics. He disregarded old-time scouts’ “intuition” and used multivariate analysis to understand what made great hitters and pitchers. Doing this gave Beane an edge in an inefficient market that overvalued a hitter’s batting average and undervalued his on-base percentage. With this approach, Beane was able to draft players that the market undervalued and build a team that produced wins at a lower cost than any other team.

However, overcoming organizational inertia to create a data-driven culture isn’t always easy. Institutionalizing the following steps and ensuring everybody asks, “Does the data support this decision?” will help make the transition easier.

Know Your Statistics
From the hiring process on, build an organization that understands data. For example, an insurer sent customers a loyalty survey that asked, “Do you use our customer service website?” and  the results showed that those who did were 10% more loyal. However, customers aren’t more loyal because they use the self-service website. Another factor was probably driving the correlation. For example, loyal insurance buyers are more likely to use online servicing and be less price-sensitive in the first place.

In short, a thorough understanding of statistics, correlation and causality is critical to designing data-collection processes, measuring results, interpreting results and recommending action.

Employ an Expert 
At least one person on the team needs to be an expert to prevent mistakes, mentor colleagues and champion analytics. Not everyone will be analytics and statistics masters, but at least one team member should be able to validate proposed test scenarios and collection techniques, interpret results and lead the creation of a data-driven culture. One-third to one-half of this person’s time should go to unstructured data analysis and developing new insights; this person should be evaluated on how effectively the organization’s culture changes.

Ensure You Can Measure
Make measurement part of the process, not an afterthought. Creating a data-driven culture starts with being able to accurately measure and present reliable and actionable information. Recently, an insurer sought to determine if customers who know their premium is changing are less likely to switch carriers once the change happens. To measure benefits, the insurer notified some customers but not others and then measured the retention difference between groups. However, the company did not know that agents had already notified customers in the control group, thereby invalidating the results.

Remove Emotion (For the Most Part)
Good decisions rely on data first, experience second and intuition third. However, many leaders still base decisions primarily on intuition and experience. The presence of readily available data makes it easier for decisions to be fact-based and unemotional. This can help leaders use less time to make decisions and more time thinking strategically, developing people and truly leading their organizations.

Test, Learn and Fail Fast
Don’t be afraid of mistakes, but instill processes that let them happen quickly and economically, and produce lessons. Because of the importance of supporting risk taking, the ability to rapidly establish test-and-learn scenarios is critical. A financial services company recently launched a $4 million program to offer a new product, and assumed a 10% sales conversion rate. The actual conversion rate was only around 2%, and the program was a resounding failure. Rather than sinking $4 million into the program, a pilot program with a couple of agents and mock website would have been more effective.

Go for Insights (Not Data) and Analysis (Not Reports)
Usable knowledge comes from true analysis; true analysis comes from empowering analysts. Most business leaders say they get plenty of reports but not enough information. Analysts should have the right tools, training and support to spend their time doing true analysis. If your analysts are spending more than 20% of their time creating reports, then the processes that enable a data-driven culture are not in place.

It’s All About the Process
Fact-based decision making starts with building a process to measure facts and creating system that incorporates them. Metrics and measurements don’t always get the respect they deserve. They’re often an afterthought and frequently get cut when there are overruns. Metrics deserve their own budget, and an independent business group should be accountable for ensuring the right measurements are in place.
Scott Busse is director of the insurance practice at PricewaterhouseCoopers.