Plan requirement
| Subscription | Suite Professional or higher, Explore Professional or higher |
| Access | Agent |
Average hides what median shows. On support times that difference is not academic; it changes the conclusion.
The options
| Count | How many. For volumes. |
| Sum | Added together. For totals of quantities. |
| Average | The mean. Pulled around by extremes. |
| Median | The middle value. Unmoved by extremes. |
| Min and max | The edges. Useful for finding the worst case. |
Use median for times
Support times have a long tail: most tickets answered quickly, a handful open for weeks. Those few drag the average far above what a typical customer experiences.
The median says what most people got, which is almost always the question being asked.
Report both when it matters
Median next to average tells you something neither does alone. A large gap between them means a tail worth investigating, and that tail is usually where the unhappy customers are.
Averages of averages are wrong
Averaging each agent's average handling time treats an agent with three tickets the same as one with three hundred. Aggregate from the underlying data rather than from other averages.
Percentiles beat both for targets
Where you have a service commitment, "ninety per cent within four hours" is what the commitment says, and neither average nor median measures it.
Say which one the chart shows
On the report itself. A number labelled "response time" that is silently a mean will be compared with somebody else's median, and the argument that follows is unwinnable.
Comments
0 comments
Article is closed for comments.