Using Histograms

Histograms are a graphical tool we use to visually depict the frequency of something over given time periods. For example, in a customer support contact center, we might be tracking the number of calls that get dropped on a day to day basis. The scale on the left side would represent the volume of calls that are dropped. Across the bottom axis we have ten day period, and each column represents how many calls were dropped each day. With that information, we can do some basic analysis. Let us say, for example, that these represent the first ten days of the month. And let’s just pretend that the number 4 represents a Monday during this ten day period. Mondays tend to be a very high call volume day in a lot of organizations. And there’s often too much volume to keep up with it. Another explanation is that maybe we have a higher absenteeism on this Monday, leaving us understaffed for the volume of calls. Therefore, callers are hanging up because the wait times are too long. The histogram allows us to identify which days may be a problem. And this gives us an opportunity to delve into the reasons behind the problem. Constructing a histogram is fairly straightforward. Typically it’s done using a spreadsheet or manually on a piece of paper, where we plot the frequency and then the various days of the week. So for example, we could be using a check sheet or a tally sheet. We ask our call center manager to run a report to determine how many calls were dropped on a daily basis. And then we tally that up to get a view of the entire process over a range of days. Or we may have an automated method. For example, call tracking systems can capture and calculate that data for us as well.

Let’s say that we are evaluating the weight of a part or some other item. This histogram depicts the number of ounces in a cup of coffee that’s being poured. We’re trying to determine if the data has a central tendency. If we can begin to draw a line or a curve to connect the columns, is there any particular shape that forms across the data? Is there a definite mean or average with data points clustered around that average? Looking for shape is always important to give us a sense for whether the histogram is behaving in a normal fashion or in an abnormal fashion. Histograms with a central tendency tend to give us a view that we can look at like a bell shaped curve. The mode would be the most frequently recurring value across the data set. And the mean would be the average value of all the data points. The bottom axis of our histogram represents the range of the data. In an example of intelligence quotient, or IQ scores, in a population. If the population is normally distributed like a bell curve, we would expect a lot of folks to be in the middle of the average intelligence. A smaller number on the lower end of the scale with a lower IQ, and a smaller number of the upper end or genius range. We may also begin to identify some outliers to the right or to the left.

In summary, a histogram is a very important quality tool. Understanding how to create and interpret a histogram will be very helpful to you and your team when it comes to evaluating your processes.