March 2019 - Metrics Update

Today we make the first in a series of updates to improve Textline’s metrics offerings.  This release adjusts how we calculate some of our Service Level Agreement (SLA) metrics by switching them from a median calculation to a mean (Avg).

 

What does this mean?

Previously the below Insight metrics were calculated to look at the median number.  Today, we start calculating them to look for the mean number.

  • Time to Respond
  • Time to First Response
  • Time to Resolve

In-app we simply refer to these metrics as (Avg) now.  This is because while a median is technically a type of average, a mean calculation is the most common type of average and when referring to an average, unless otherwise specified, you are referring to the mean way of calculating averages.

 

What is the difference?

Here are some visuals to help understand the change.

 

Given a series of numbers:

3, 1.1, 0.1, 1, 0.3, 15, 1.5

A median calculation will then order the above numbers:

0.1, 0.3, 1, 1.1, 1.5, 3, 15

And then identify the “middle” of the pack:

0.1, 0.3, 1, 1.1, 1.5, 3, 15

A mean calculation, on the other hand, sums the numbers and then divides it by the count:

CodeCogsEqn.png

 

Why?

Median calculations are a type of average calculation to see the middle and offset some large outliers.  After customer feedback and research into various standards, we decided that these calculations should look at the mean instead.

There are a lot of different opinions out there about which average calculation is better.  Median is great in that it removes the impact of outliers.  In our above example, the value 15 greatly affects the mean calculation.  Remove 15 and the mean becomes 1.17 which is much closer to the median calculation.

That being said, the mean has become a service industry standard.  There are also arguments that all values, including outliers, should be counted.  It shows that you missed something and that should not be discounted.

On the Textline side, there is also a huge effect on performance when making the median calculation.  In terms of computation, ordering the list and then searching for the middle is much harder than summing and dividing by the count.  This last aspect is also important to us as we start to add in more metrics for Textline customers.



Switching to a mean calculation will change your numbers, but hopefully this change makes sense.  Let us know if you have any questions or concerns.  We’re always here to help!