# Visualizing Averages

You’re familiar with averages – whether it be from that nightmarish period in high school when you fretted over your entrance average for college or university, or from trying to figure out the historical average temperature in Cuba for the week you booked your vacation. We use averages for many day-to-day purposes, but, chances are, some of your Key Performance Indicators (KPIs) contain averages too.

Do any of these sound familiar?

• • Average Order Size (Average Market Basket)
• • Average Margin
• • Average Call Abandonment
• • Average Handle Time (Average Resolution Time)
• • Average Bounce Rate
• • Average Page views
• • Average Wait Time

Odds are, you’re missing a lot of very important details when it comes to visualizing these values. Let’s look at a very simple hypothetical example: I’m selling an earthquake warning iPhone application, and my software warns users if an Earthquake is about to occur in their location, all for \$1.99 on iTunes. Pretty useful, right?

Now, let’s take a look at my earthquake app’s average customer satisfaction.

In the app store, my product has an average of 3.66! Take a minute to really think about this number and what it means to you.

Would you download an application with a score like this? It doesn’t have a perfect score of 5, but it does score above a half rating, and it’s almost at a solid rating of 4. It sounds pretty good doesn’t it?

Let’s take a look at how this average was derived by looking at the associated comments.

• • User 1: Beautiful user interface, great responsive design. 5/5
• • User 2: Can’t beat the price compared to the competition! 5/5
• • User 3: Did not actually warn of Earthquakes!!! 1/5

Now that you can see that the application doesn’t actually warn users of earthquakes, 3.66 suddenly doesn’t look as appealing as it did before. Maybe if the average customer satisfaction ranking was presented differently, you could’ve understood the application’s real value at first glance.

Let’s take a look at a real life example of a call center where we are measuring average handle times. For those of you who are unfamiliar with this metric, an average handle time is the average amount of time a rep is on the phone with you while trying to solve your problem.

The following visualization shows the average handle time, but also breaks it out so that you can see all the details at a glance. For every call or web chat that takes place within the day that we are examining, a square or triangle is added to the chart to represent calls and web chats respectively. This type of visualization immediately allows you to see outliers, instead of bundling potentially alarming data up into a single average.

See how Dilya had a 56 minute web chat that took place today? Instead of harping on her for reducing her average handle time, Dilya’s manager can now drill into the detail of this chat and help her do her job more efficiently. Maybe customers are asking certain questions that reps do not have the tools to answer easily. By drilling into the detail on the outliers, her manager can now ask specific questions and really try to understand how to make the employees’ lives easier!

Another method of visualizing an average is to place all of the data into predefined ‘buckets’. The actual visualization that I am describing here is called a histogram, and is especially useful when there is so much data that it makes more sense to break the data down into levels. Take a look at this customer service visualization. Isn’t a distribution much more telling than showing an average?

Although averages are very easy to understand, they often do not tell the full story. And when you’re trying to improve your business and create efficiency, knowing the full story makes a huge difference.