mode
The most frequent value in the data; the only measure of central tendency that works for categories.
The mode is the value that occurs most often in the data. There is hardly a formula: count how many times each value appears (its frequency) and pick the one with the highest count. Data can have one mode, two (a bimodal distribution), several tied modes, or no useful mode at all when every value is unique.
The mode's natural home is categorical data. Flavour, traffic channel, device, plan tier, cancellation reason: you can't take the mean or the median of "salmon" and "chicken", but you can count them. That makes the mode the only measure of central tendency that works for categories. It also fits discrete scales with only a few possible values, such as 1–5 star ratings or the number of items in a basket.
For continuous numbers such as order value or session length, the mode is usually the weakest summary. Precise values rarely repeat, and one extra observation can move the mode somewhere else entirely. If you need it there, group the values into bins first, as a histogram does, and talk about the most common range (the modal bin).
The mode has a useful side effect: a histogram with two clear peaks tells you that you are probably mixing two different groups, for example households and cat cafés. No single typical value, whether mean, median or mode, describes both. That's a cue to segment rather than to average. On right-skewed numeric data the three measures often line up as mode < median < mean, though that is a tendency, not a law.
Example
Categories. CatChow tags each of 20 orders with its main flavour:
The mode is salmon, with a frequency of 8 out of 20, or 40%. Note what it does not say: 60% of orders were something else, so "most customers want salmon" would be wrong.
Ratings. A new grain-free recipe gets 12 reviews: 1, 1, 1, 1, 3, 4, 4, 5, 5, 5, 5, 5. The mode is 5 (five times), but 1 is close behind (four times): opinions split into fans and people whose cats refused it. The mean, , describes nobody, and the median, , hides the angry group. Here the mode plus a look at the second peak tells the real story: find out why a third of the reviewers gave one star.
Common mistakes
- Reading the mode as "most customers". The most common category can still be a minority: salmon at 40% means 60% chose something else.
- Taking the mode of raw continuous values. $31.47 and $31.48 count as different values; bin the data first or use the median.
- Ignoring a second peak. Two modes usually mean two segments. Report them separately instead of forcing one typical value.
- Averaging categories coded as numbers. If channels are stored as 1 = organic, 2 = paid, 3 = email, an "average channel of 1.8" means nothing. Use the mode and the frequency table.
- Treating the mode as stable in small samples. With ties or near-ties, one extra order can flip it; show the counts, not just the winner.
Learn it in the course
- Typical value: mean, median, mode · Mean, median, mode and trimmed mean on real-looking CatChow orders: what each one answers, why one bulk order can fool the average, and how not to average averages.
- Samples, populations and the normal distribution · Why we study samples to learn about all users, and why the bell curve shows up everywhere.