Joan Lillich

How change management can improve decision-making

Behavioral science in business applications is a fascinating topic.

In his book “The Undoing Project,” Michael Lewis describes behavioral economics and how even the best experts’ decision-making is not as accurate as predictive modeling or analytics. He cites an Oregon Research Institute study where radiologists were consulted on the best methods to make an accurate clinical diagnosis of stomach cancer. The doctors described their decision-making process and researchers developed an algorithm to closely mimic that process.

It turns out the algorithm produced more accurate outcomes than the doctors. The experts who knew what to consider were not as good as the machine in using the decision factors to make a diagnosis. The doctors’ diagnoses disagreed with one another; and, when presented duplicate images of the same tumor, every doctor submitted different diagnoses for the duplicated (identical) image.

The simple message here is that our humanity makes us prone to stereotypes, fatigue, likes and dislikes, prejudices, and good or bad days. As a result, predictive modeling is a very good tool to assist humans and businesses to make better decisions.

The goal for many organizations is to become “insights-led.” In business and IT consulting, we often develop predictive models that use accurate data to help clients make effective decisions. But what happens when the insights presented aren’t trusted? This can occur even when decision-makers were the ones who selected the most relevant data to use, determined the proper weighting, assigned the staff to enter the data into the system of record, and selected which data to report.

Why is it that sometimes executives don’t want to trust data? Often it is because they value their hard-won experience and pride themselves on their “gut feel,” and, moreover, they want to be “comfortable” with the calls they are making. We need to support leaders to move out of their comfort zone, and to have the courage to make insight-based decisions supported by quality data, even though it may go against their natural instincts.

Given that data-driven insights play such a key role in advancing digital transformation, which is a top priority expressed by our clients, how can we increase our stakeholder’s comfort level with data-driven insights? The practice of organizational change management can help on two levels:

  1. Strategically, where leaders and their business objectives need to be supported with the data-driven insights
  2. Tactically, in areas where subject matter experts and managers can build a groundswell of support for learning from the data and incorporating continuous improvement


The simplest strategic approach for leadership is to develop a success model that determines a handful of metrics they expect to drive as a result of the data-driven insights. The activity to drill down to only a few metrics is among the best ways for the business units and IT leaders to come to agreement on the most important metrics that will help solve business problems. Once there is agreement on those key metrics, the entire organization knows the key priorities and can focus on baseline and target performance objectives to guide their own, day to day, decision-making.

A key component of tactical change management is to focus on the middle of the organization—an area I refer to as the “squeeze zone.” In any transformative change, managers are squeezed to meet current business metrics while concurrently dedicating the most experienced staff to help design the insights-led initiatives. They sacrifice time from all staff to learn and ramp up on the new predictive modeling procedures and, equally important, to dedicate their own time to move the culture from transactional to analytical.

Any one of these manager tasks seems like a full time job. CGI’s experience has been that transformation success is tied to managers’ engagement throughout the process. We enable them to provide input to the designs as well as advanced communication and demonstrations, early training (before their staff), and support during implementation.

I invite you to read more about going beyond common sense to harness the true value of business intelligence in an article by my colleague Steve Lennon, in Australia, and also to learn about CGI’s change management capabilities.

Blog moderation guidelines and term of use