One number is never the whole story
People often ask for “the average” as if it were a single, obvious figure. In practice there are several ways to summarize a list of numbers, and they can disagree, sometimes by a lot. Each one answers a slightly different question, and each one can hide something. Understanding the differences helps you pick a sensible summary and spot when someone else has picked a misleading one.
Three common summaries
- Mean. Add all the values and divide by how many there are. This is what most people mean by “average.” In a spreadsheet it is usually the average function.
- Median. Sort the values from smallest to largest and take the one in the middle. If there is an even number of values, take the mean of the two middle ones.
- Mode. The value that occurs most often. It is most useful for categories or for numbers that repeat.
A small made-up data set
Imagine, hypothetically, that a small shop recorded how many minutes seven deliveries took:
20, 22, 25, 25, 28, 30, 120
The list is already sorted. Here is how each summary comes out.
Mean. The total is 20 + 22 + 25 + 25 + 28 + 30 + 120 = 270. Dividing by 7 gives about 38.6 minutes.
Median. With seven values, the middle one is the fourth: 25 minutes.
Mode. The value 25 appears twice and nothing else repeats, so the mode is 25 minutes.
Notice what happened. Six of the seven deliveries took between 20 and 30 minutes, yet the mean says 38.6, a time that not one delivery actually took. The single 120-minute delivery pulled the mean upward. The median and mode both stayed at 25, which describes a typical delivery much better.
What happens without the outlier
Remove the 120 and the six remaining values add up to 150. The mean becomes 150 divided by 6, which is exactly 25. The median is the mean of the third and fourth values, (25 + 25) / 2 = 25. So without the unusual value, all three summaries agree.
That comparison is a quick and useful test. If removing one or two extreme values changes the mean a lot but barely moves the median, the data is skewed or contains outliers, and the mean on its own is not telling you much.
Why the median resists outliers
The mean uses the actual size of every value, so an extreme number has a large pull. The median only cares about order. Whether the largest delivery took 120 minutes or 1,200 minutes, it is still just the largest, and the middle value stays the same. This is why medians are often used for things like household incomes or home prices, where a few very large values could otherwise dominate. You will often see that choice explained in reports, though it is worth checking which one is actually being used.
When the mean is still the right choice
The median is not automatically better. The mean has real advantages:
- It uses every value, so it reflects the total. If you want to estimate total time for 100 similar deliveries, the mean multiplied by 100 is the relevant figure, because the extreme values really do add to the total.
- It behaves nicely in further calculations and is widely understood.
- When data is fairly symmetrical with no extreme values, the mean and median will be close, and either works.
So the question to ask is what you want to know. A typical single case points toward the median. A total, or a rate across many cases, often points toward the mean.
Outliers are not always errors
An unusual value deserves a look before you delete it. It might be a typing mistake, such as an extra zero. It might be a real event, such as a delivery that was delayed by a genuine problem. Real events can be the most important part of the story. Dropping them quietly to make a summary look tidy is a way of hiding information rather than revealing it. A better habit is to report the summary you chose, mention the unusual value and say what you did with it.
Cleaning obvious entry mistakes is part of data cleaning, but decisions about real but unusual values are judgment calls, and it helps to write them down.
Averages also hide spread
Two data sets can share the same mean and still look very different. A set of values all close to 25 and another set swinging between 5 and 45 might both average 25. The mean does not show how scattered the values are. Useful companions include the smallest and largest values, the range between them, and a simple chart that shows every point. Good data visualization often reveals clusters, gaps and outliers that a single summary number would flatten.
A grouped average can also mislead. An overall figure might combine very different groups, such as new and returning customers, and describe neither well. When you have information about groups, looking at them separately, perhaps with a summary table, is often more honest than one blended figure.
Quick checks before you trust an average
- Look at the sorted data, or a simple chart, before summarizing it.
- Calculate both the mean and the median. If they are far apart, find out why.
- Check the smallest and largest values for mistakes or real events.
- Note how many values you have. A summary of five numbers is much less reliable than one of five thousand.
- State which summary you used when you share the result, especially with survey results or customer feedback where a few strong opinions can pull the mean.
None of this requires advanced mathematics. It requires a habit of asking what the number is really describing, and a willingness to show more than one summary when the data is uneven.