The short answer
The average spreads total response across all measured posts. The median is the middle value after sorting. When a few posts dominate, the median often gives a more useful description of an ordinary observed post. Keep both and show the sample size.
What you’ll take away
- A large outlier can lift the mean while the median barely moves.
- A measured zero belongs in the sample; a missing count does not.
- Small or selectively collected samples can mislead with either statistic.
Ten posts, two very different headlines
Use these illustrative reactions-plus-comments counts: 2, 3, 4, 4, 5, 5, 6, 7, 8, 96. The total is 140. Divide by ten and the average is 14. Sort the values and average the fifth and sixth positions: the median is 5.
Nine of the ten posts received fewer than 14 responses. The mean is mathematically correct, but “my normal post gets 14” would be a poor description. A more faithful sentence is: “The median was 5 across ten posts; one post received 96.”
| Statistic | Value | Useful question |
|---|---|---|
| Total | 140 | How much response did the sample receive? |
| Mean | 14 | What is total response per measured post? |
| Median | 5 | What was the middle observed result? |
| Largest post | 96 | What stood out enough to investigate? |
Change the outlier and watch what stays still
In the interactive example, replace 96 with 960. The mean rises to 100.4. The median remains 5. Most posts have not changed at all; the summary headline has.
Now add several genuinely measured quiet posts. Both the mean and median may fall, because the observed distribution changed. That is not evidence that the calculator is unfair. It is evidence that including or excluding ordinary activity affects the story.
Do not turn unknown counts into zero
A post with zero measured reactions and zero measured comments has a response of zero. A post whose comments were unavailable has an unknown combined response. Treating that unknown value as zero would bias the summary downward.
Wrapped.li’s richer response distribution requires at least eight posts with both counts measured. That threshold is a product safeguard, not a statistical guarantee. It does not eliminate collection bias or make eight posts representative of a year.
A middle value still needs a range
Two sets can share a median while behaving differently. The values 4, 5, 5, 5, 6 and 0, 0, 5, 20, 100 both have a median of 5. The second set is much less consistent. Look at the spread and the original posts before deciding what is repeatable.
A middle-50% range describes where the central half of a measured distribution sits, using a stated percentile method. It is not a prediction interval for your next post. Small samples make such boundaries especially sensitive to individual observations.
Write a sentence you can defend
Use: “Across ten collected posts with both counts available, median reactions plus comments were 5 and the mean was 14. One post received 96.” Add the date range and any collection limit if you publish it.
Avoid: “My content improved because my average increased.” The increase might be one outlier, different observation ages or a different collected sample. A better next step is to read the standout and an ordinary post, identify a specific difference, and test a follow-up.
Sources & product references
Product details reflect the publication date. External features and interfaces may change.
