Ask someone for "the average" and they'll almost always hand you the mean -- add everything up, divide by how many numbers there are. It's the most familiar statistic and the easiest to compute, which is exactly why it gets used in places where it's the wrong tool for the job. Mean, median, and mode aren't interchangeable synonyms for "typical." Each one answers a different question, and picking the wrong one can make a dataset look nothing like what it actually represents.
The mean gets dragged around by outliers
The mean is sensitive to every single value in the dataset, including the extreme ones. That's its strength when data is evenly spread, and its weakness the moment it isn't.
Imagine ten people in a room, nine of whom earn a modest salary and one of whom is a billionaire. The mean income in that room is enormous -- but it describes almost no one in it. The billionaire's single data point pulls the average so far up that it stops representing the group at all. This is exactly why national income and house-price statistics are usually reported alongside a median: a handful of extreme values can distort a mean badly enough that it tells you about the tail of the distribution, not the center of it.
The more skewed or outlier-prone your data, the less trustworthy the mean becomes as a stand-in for "typical."
The median only cares about the middle
The median sorts every value and picks the one sitting exactly in the middle (or averages the two middle values, for an even count). It doesn't care how large your biggest number is or how small your smallest one is -- only about rank order.
That's why the median survives outliers that wreck the mean. In the salary example above, the median income in that room of ten is just the 5th or 6th value once sorted -- the billionaire doesn't move it at all, because their position in the ranking doesn't change the middle. This is also why the median is the standard choice for anything prone to extreme values: home prices, incomes, response times, wait times.
But the median has its own blind spot: it throws away magnitude. A dataset where the top half is just slightly above the middle value looks identical, median-wise, to one where the top half is astronomically higher. If you only report the median, you can hide exactly how extreme your outliers are.
The mode is the only one that works on categories
Mean and median both require numbers you can rank and do arithmetic on. Mode doesn't -- it just finds whichever value shows up most often, which makes it the only one of the three that works on non-numeric data: the most common browser in your analytics, the most frequent support ticket category, the most popular shirt size ordered.
For numeric data, mode becomes useful again when a distribution has more than one cluster. A dataset can have no mode (every value is unique), one mode, or several -- and when it has several, that's usually a sign the data is actually made up of more than one underlying group mashed together. A bimodal distribution of "time spent on page" might mean you have two very different kinds of visitors, not one messy average between them.
Why spread matters as much as the center
Any of these three numbers can hide how spread out the data actually is. Two datasets can have the same mean and still look completely different -- one tightly clustered, one wildly scattered. That's what variance and standard deviation measure: how far, on average, values sit from the mean. A small standard deviation means most values sit close to the center; a large one means the "average" value is doing a poor job of representing any individual data point.
This is why a single number -- mean, median, or mode -- is never the full picture. It's a summary, and summaries lose information by design. The question is whether the information they lose happens to matter for what you're trying to decide.
Check which one you're actually looking at
Before trusting an "average" in a report, a dashboard, or your own data, it's worth asking which of the three it is and whether that's the right one for the shape of the data. If you're working with raw numbers and want all of them at once -- mean, median, mode, range, variance, and standard deviation -- rather than computing each by hand, the Statistics Calculator will run them all on a pasted list of numbers so you can compare what each one is telling you, side by side.