Is .com Name Length Lognormal? The Distribution Explained
The name length distribution of .com is close to lognormal. We fitted a curve to all 167 million names on 9 October 2026. About 5% of names would have to change length for a perfect match. A normal curve misses by about
The name length distribution of .com is close to lognormal. We fitted a curve to all 167 million names on 9 October 2026. About 5% of names would have to change length for a perfect match. A normal curve misses by about 11%.
What is a lognormal name length distribution?
A lognormal curve is a bell curve drawn on a stretched scale. If you take the logarithm of each length, the results follow a normal curve. The shape is lopsided: it rises fast, peaks, and trails off to the right.
Names fit that pattern in three ways. The most common length is 11 characters, which holds 8.3% of the zone. Half of all names are shorter than 12, so 12 is the median. Meanwhile the mean is 12.8. In a lopsided curve, the mean sits above the median, and the median sits above the peak. That is exactly the order we see.
You can test this yourself. Count the names at each length, then plot the lengths on a log scale. If the bars form a bell, the data is lognormal. Our fit does the same thing with arithmetic instead of a plot.
How well does the curve fit?
We fitted the curve to the count of names at each length from 1 to 63 characters. The centre is a geometric mean of 11.8 characters, and the spread is 0.40 on the log scale. Then we compared the curve with the real counts.
To measure the gap, we added up how far each bar sits from the curve. The total is the share of names that would need a new length for a perfect fit. For the lognormal curve, it is 5.3%. For a normal curve, it is 10.9%. So the lognormal fit is about twice as close.
Where the name length distribution breaks the rule
Real names bunch up at the short end. The table compares the real share of the zone with the share the curve predicts.
| Length | Real share | Lognormal fit |
|---|---|---|
| 4 characters | 0.66% | 0.68% |
| 5 characters | 2.61% | 2.01% |
| 6 characters | 5.03% | 3.96% |
| 8 characters | 6.95% | 7.71% |
| 10 characters | 8.11% | 9.10% |
| 12 characters | 8.06% | 8.28% |
| 15 characters | 6.08% | 5.57% |
| 20 characters | 2.40% | 2.11% |
| 30 characters | 0.13% | 0.22% |
Five and six characters hold more names than the curve allows, and eight to ten hold fewer. More names sit at 15 and 20 characters than the curve predicts. From 25 characters on, the curve predicts more names than exist. We do not claim a cause. However, two guesses are easy to state. People want short names, and that demand would push names toward five and six characters. Besides that, many owners join two or three words, which fills the long end.
What the shape tells you
First, a typical name is not an outlier. A name of 11 or 12 characters sits at the top of the curve. Second, a short name is rare for a clear reason. There are few of them, and owners hold nearly all. Our post on taken names by length shows how full each short length is.
Finally, the shape gives a quick test for your own list of ideas. If every idea is 20 characters long, you are far from the crowd. The Naming Patterns tool shows how many names start or end with a word you like, and the Short Domains tool lists the short ones that are missing.
How we measured it
We read the daily zone file of the registry, which approved users get from ICANN's Centralized Zone Data Service. Then we counted each domain name by the number of characters before the .com. Only totals leave our system. The State of .com page keeps the numbers current, and the average length post explains the simple mean and median.
One caution applies. A fit is a description, not a law. It tells you the shape of the zone today. It does not tell you why people name sites the way they do.
Originally published at namedaemon.com/com-name-length-distribution-lognormal/. NameDaemon has 18 free tools for .com names, built on the daily zone file, and the numbers are on the State of .com page.
Originally published by Dev.to WebDev. Aggregated on AIWithGhost for educational purposes — full credit and traffic to the original publisher.