Counting Candy

David Williams, PhD

The candy count exercise simulates one of the foundational concepts in the science of improvement: understanding variation. Here at IHI, we use it in the IHI Improvement Advisor Professional Development Program. Candy count is a quick and easy way to learn the concepts of special versus common cause variation with a small group in just 15 minutes. After doing this exercise, you'll be able to define common cause variation and special cause variation, and discuss why they're important to process improvement.

In improvement science, we learn to recognize two types of variation: common cause variation or variation that is part of the daily normal work, it's built into the system random, and affects everyone; and special cause variation, variation that's attributable to a cause, and non-random. If we want to improve a process, we need to know whether the results we're seeing are inherent to the process that we're trying to improve or due to some identifiable cause. For example, if you're measuring your commute time to work, you probably see some common cause variation that's part of the normal process. It might be longer or shorter depending on the traffic lights, pedestrians, or your own pace. That's common cause variation. Every now and then though, you'll come across something unusual, a car crash, that causes a different result. That's special cause variation.

The setup is simple. Your facilitator will hand out a small pack of M&Ms to each participant. Count your M&Ms in secret, and then share the number with the facilitator when you're called on. Your facilitator will plot out the number of M&Ms in each bag on a run chart. The X-axis shows the number of participants across the bottom, and the Y-axis shows the count of the M&Ms. Remember, no eating yet. It's just counting.