January 29, 2019

Failures in Data-Driven Decision Making

Bryan Hughes

Bryan Hughes
President /FirstService Residential Massachusetts

Share This Post

"... we need to be aware of the why behind the data, not just the data itself."

 

Nearly all business colleges require at least one course on the process of data-driven decision making. The concept is that if people can see the details of past data, plus current trends, that they can best make future decisions that range from staffing headcount to inventory purchases. This is a worthy endeavor and almost a science in itself. It is not, however, a religion. Many leaders review the data as gospel truth, but do not balance that data with the subjective realities of the business. This isn’t for all businesses of course, but primarily for those utilizing people as the main service mechanism.

It has been a trend for organizations to adopt a balanced scorecard and with it comes the Dashboard. Within a single page, managers have at their fingertips real-time data in areas such as marketing conversion, revenue trending, productivity by headcount, or any number of KPIs that are considered valuable to the organization. The expectation is that this will allow people to quickly make adjustments and take actions based on what they see in real-time. This should provide the organization with faster response times, and it does, but should we always act on this data?

If we review the data, we need to be aware of the why behind the data, not just the data itself. If we start with the basics though, the data needs to be accurate. This may seem to be a given, but it isn’t. I was associated with an organization recently where KPIs were discussed at great length. My challenge to the organization was why they were tracking these specific numbers since the process of gathering them was flawed. The organization was making judgments of team effectiveness by evaluating results immediately post-event, yet admitted that the real numbers wouldn’t surface until months six or even twelve. It was justified that they “adjust” for that reality and act accordingly. I again asked what good the numbers really were? They were not accurate, timely, and any action based on them could be flawed. Why even take the time to track and report?

I was also recently associated with an organization that had a great and colorful dashboard that tracked real-time data, but also extrapolated that data to make end-of-month and year predictions. Again, great idea, but the formulas were wrong. Because of the type of business, end of month forecasts were expected on a weekly basis but because of variability in this industry, the dashboard predictions would swing wildly from day to day, causing panic at higher levels. Further, the dashboard lost credibility with the team because they didn’t believe the data could be trusted.

To quote from Joel Shapiro, who wrote for the Harvard Business Review, “Too often we think of analytics as representing some sort of unbiased and dispassionate truth.” The truth is the key takeaway from this. I have been with individuals and organizations that would show me the truth based on what they had on a dashboard. I would ask a question about how a segment of the operation was going and they would refer to the dashboard to answer the question. I would ask then what were the contributing factors behind the score shown on the dashboard and better than 50% couldn’t tell me even how the number was generated! They had no clue of what was going on their organization from a people, morale, and energy perspective. They hid behind their dashboard rather than taking the time to be in the field to understand what each of those numbers really represented.

Dashboards are wonderful as a single-screen measure of real-time data. The trick, however, is to first ensure that the data is real and relevant, but then to understand the contributing factors behind each of those numbers. If your organization is so big that the customer and the C-suite are distant, then ask, listen, and counsel with mid-level and frontline managers to better understand the why behind the numbers. This will not only help for better data-driven decision making, but will also better avoid difficult conversions with boards and analysts who expect forecasts to be reasonably accurate.

Comments? You can contact me directly via my AdvisoryCloud profile.

Share This Post