The mean adds values and divides by their count. The median describes the middle of a sorted series. They answer different questions.
01STORY BRIEFThe idea in three screens.3 SCREENS
The average loves an overly good story.
The mean adds values and divides by their count.
Fictional example: 100, 100, 100, 100 and 10,000 views. The mean is 2,080; the median is 100. The first describes theoretically spread volume; the second identifies the central post. Keep both alongside the distribution to distinguish an isolated hit from broad improvement.
In the example, removing the 10,000-view result brings the mean down to 100. This sensitivity check does not mean the hit should be discarded. It reveals dependence on one observation. A useful report keeps the complete result and context, then asks about repeatability rather than inventing it. The exceptional post may matter greatly while still telling us little about what an ordinary next post will achieve.
Make the ordinary result visible
Imagine an illustrative batch with nine modest posts and one enormous outlier. The average answers a legitimate question about total reach per post, but it describes an experience most posts never had. The median answers a different question: what did the middle result look like? Neither deserves to win by default. A business financing a whole portfolio may care about totals; an operator debugging a repeatable publishing routine may care about the typical result. Put both numbers next to the actual number of posts and explain which decision each supports.
Two summaries, two questions
- 01Mean
Total outcome divided across all included posts.
- 02Median
The middle post in an ordered set of results.
- 03Distribution
The shape and concentration that both summaries hide.
Editorial model: a way to reason about the process, not measured platform results.
Then inspect the distribution instead of arguing about one summary. Sort the posts, mark accounts and formats, and look for a small group carrying the total. A higher average with a flat median may mean the upside improved while the ordinary post stayed unchanged. A higher median with a lower total may still be welcome if the operation became cheaper or more predictable. Do not remove an inconvenient hit just because it is inconvenient; show the result with and without it and explain why the comparison matters.
“The average loves an overly good story.”
02FIELD KITFrom concept to practice.2 SCREENS
Show mean and median together.
Adapt this to your audience, budget and test scope.
Put it to work.
- 01Show mean and median together.
- 02Inspect extreme values.
Notes & sources.
The limit. The median can also hide the economic importance of a few very large outcomes.
Further reading is linked throughout this article. Worked examples explain a process; they are not campaign results.
THE ORGANIC CLUB is an editorial project initiated by TokPortal. That relationship is separate from the evidence behind claims. Editorial signatures and image credits are explained in our policy. Our editorial policy →
