104 names hold more of the register than the 8,205 rarest combined

This site counts real registrations — licences and shelter intakes published as open data — and it counts a lot of them.

Spread across the catalogue they'd give every name a few dozen. They aren't spread. Half of every registration sits on a few hundred names, and the lower half of the catalogue, six thousand names, divides about a sixteenth of the register between them.

The four bands are not four quarters

The catalogue sorts names into popular, common, rare and very unique.

The band names describe how crowded a name is, and it's easy to read them as slices of similar size. They're nothing like it: 104 names are popular, 399 common, 3,487 rare and 8,205 very unique.

Weigh them by registrations rather than by names and the order turns over. The 104 popular names hold 160,590 registrations between them. The 8,205 very-unique ones — nearly eighty times as many names — hold 147,159. Less. A hundred names outweigh eight thousand.

The largest share belongs to neither end. The rare band holds 318,544 registrations across 3,487 names, more than any other band, because it's where a catalogue this size keeps most of its usable middle.

This is what naming always looks like

The shape isn't a quirk of pet naming. It's what happens to almost anything people choose freely and independently.

Human first names do it. Words in a language do it — a small number of words account for most of everything said, and the dictionary's long tail is used almost never. Town sizes, book sales, the popularity of websites: the same steep curve turns up wherever many people make the same kind of choice without coordinating.

The mechanism is usually some version of the same thing. A choice that's already common is easier to encounter, easier to think of, and comes pre-endorsed by everybody who made it before you — so being chosen makes a name likelier to be chosen again. Nothing has to be organised for that to produce a very lopsided result.

Which means the concentration on this page isn't evidence that pet owners are unimaginative. It's evidence that they're doing what people always do when picking from a large open set, and the interesting question is what the tail looks like rather than why the head is crowded.

What the average hides

Averages per band tell the story cleanly enough — 1,544 registrations for a popular name, 566 for a common one, 91 for a rare one, 18 for a very-unique one.

Across the whole catalogue the average is about seventy. The median is sixteen.

A mean four times its median is the signature of a distribution with a long tail, and it's why the average is the wrong number to plan with. Almost no name gets seventy registrations. Most get a handful; a few hundred get thousands.

The practical version of that, if you ever read a statistic about names anywhere: ask whether it's a mean or a median. On a distribution like this the two describe different worlds, and the mean is the one that flatters whatever point is being made.

What the tail is made of

Six thousand names dividing a sixteenth of the register is a strange object, and it's worth asking what's actually in it.

A large part is ordinary words used once or twice — the objects, foods, places and jokes that a household arrived at on its own and nobody else happened to. These are the genuinely good finds, and they're indistinguishable by count from everything else down there.

Another part is spelling variants. The same name written four ways is four entries in a catalogue that reads spelling, and three of them will be thin. Nothing here merges them, because deciding that two spellings are the same name is a judgement rather than a measurement.

A third part is names from languages the registers cover thinly — perfectly ordinary where they come from and rare here purely because of which cities publish their data.

And some of it is noise: a typo entered once, a name that was never a name. That's the honest cost of counting what's there rather than curating it, and it's a small share of a very large tail.

What a catalogue this size is for

If a few hundred names carry half the register, it's fair to ask why anybody would publish twelve thousand.

The answer is that the head of the distribution is the part you already know. Anybody can list the top twenty pet names; you could probably write most of them from memory, and every naming site has them. What's hard to find is everything else, and everything else is where a name that suits your animal and isn't shared with three others actually lives.

A catalogue earns its size in the filters rather than in the ranking. Twelve thousand names sorted by popularity is an unreadable list; twelve thousand names you can cut by length, ending, sound, species and rarity is a tool. The long tail only becomes useful when you can slice it.

It also earns it in honesty. Publishing only the crowded end would mean publishing only the part of naming that's already visible, and quietly implying that the rest doesn't exist. Most animals are called something no list features, and a catalogue that can't show you those is describing a much smaller world than the real one.

Why the site does not rank on this

It would be simpler to order every list by raw registrations, and this site deliberately doesn't.

A raw total ranks a name common in a large city above one common in a small one purely because the large city published a bigger file, so popularity here is a share measured per source group instead. The concentration on this page is a fact about the registers, not about the ranking — it's what the ranking was built to correct for.

Both things are true at once. The registers are steeply concentrated, and the ordering you see refuses to inherit that steepness from whichever city happened to publish most.

What it means for choosing

If you want a name nobody else is calling across the park, the tail is enormous and cheap to reach: 8,205 names sit in the very-unique band at 18 registrations each.

If you want a name people recognise without it being the third one at the vet, the rare and common bands between them hold most of the register and most of the catalogue.

And if a name you like turns out to be popular, that's 104 names out of more than twelve thousand. Being in that group is unusual in itself.

The most useful consequence of the steepness is the one that's easy to miss: because the curve falls so fast, you don't have to go far to escape the crowd. A few places down from the top is already a fraction of the company, with none of the cost in familiarity — which is a better trade than most people realise is available.

It also means the word rare is doing less work than it sounds like. On a distribution this lopsided, most of the catalogue is rare by any ordinary reading of the word, and the band label is marking a degree rather than a kind. A name with fifty registrations and a name with five are both names you will almost certainly never meet twice.