I used to stare at certain short letter domains and feel like I was missing something.
Cars.com makes sense. Hotels.com makes sense. Even a solid two-word .com is easy to picture, because you can see the buyer standing right there. Then a three-letter .com sells for six figures, sometimes seven, and it isn't a word at all. It looks like three letters somebody shook out of a Scrabble bag.
So where's the value?
Start with scarcity. There are only 17,576 three-letter .coms in existence. Sounds like plenty until you remember that every company, fund, startup, agency, and acronym on earth is fishing out of that same little pond. A big city has more restaurants than the planet has LLL.com domains. Once they're registered, that's it. You can't hand-reg one anymore, and you can't manufacture more. You either buy one off somebody who already has it or you go without.
That's the floor the whole market stands on.
Four-letter .coms give you more room, 456,976 combinations, but the pool shrinks fast once you start asking for quality. Pronounceable, repeating patterns, strong letters, the Chinese premium stuff, real acronym potential. Filter for any of that and the big number gets small in a hurry.
Which brings up the second piece, and honestly the one that took me longer to get. Optionality.
A three-letter .com isn't valuable because it means one thing. It's valuable because it can mean almost anything. Take a decent LLL.com and you can usually find a dozen companies, a few funds, some product names, a stock ticker, somebody's initials, an internal system somewhere, all circling the same three letters. One buyer reads it as their company. The next reads it as a product. Somebody else just wants something short and clean for a rebrand and couldn't care less what it "means."
That's the whole game, really. The buyer pool.
A long two-word domain usually has one obvious type of buyer. A good LLL.com might have hundreds. That doesn't mean they'll all pay, but it means the name has somewhere to go. To you and me it looks random. To the market, HSM or GLD isn't random at all, it's a stack of possible industries and tickers and companies waiting for one of them to need it.
The prices back that up. GLD.com went for a million. NAS.com did $1.25 million. HSM.com sold around $550,000, TXT.com at $502,250. And those are just the public ones. A lot of LLL.com deals happen privately and never get reported, which tells you something on its own about how much real money is moving under the surface.
Then there's the international layer.
In the West a lot of us like letters we can say and spell, ideally something that bends into a word. Vowels help. A CVCV four-letter, consonant-vowel-consonant-vowel, can sound like a real startup even when it's totally invented, because it passes the ear test.
China built a whole separate layer of demand on top of that. For a good stretch, Chinese premium letters were one of the biggest categories going. CHIP is the term, short for Chinese Premium, and it usually means short names with the vowels stripped out, no A, E, I, O, U, and no V either, because of how the letters line up against pinyin. Here's the part that trips up Western investors. We tend to flinch at letters like X, Y, J, Q, Z. In China some of those are the good ones, because X and Y and J map to a pile of pinyin words. Same letter, weak on one side of the world, strong on the other.
That's one of the genuinely strange things about letter domains. The value of a single letter changes depending on who's looking at it.
A Western buyer wants it to sound like a word. A Chinese investor wants the no-vowel pattern. An acronym buyer doesn't care how it sounds at all, because it just has to match the company name. A startup mostly cares that it's short and carries some authority. You're not pricing one market, you're pricing several at once, and they don't agree with each other.
Five letters is where it gets murky. Five letters on their own buy you nothing. There are millions of combinations and most are worthless. But a clean, pronounceable, brandable five can be excellent. The real line is whether it sounds like a company or looks like a license plate. Something you could see on a fintech app or a fashion label has a shot. Something nobody can spell after hearing it once doesn't.
Patterns carry weight too. Repeats help. Balance helps. Easy syllables help. The absence of "wait, how do you spell that" helps a lot. A name somebody hears once and types right the first time beats a name that needs a footnote every time you say it out loud.
This is where newer investors get hurt, and I've done my own version of it. You hear "short domains are valuable" and you start thinking any four or five letters are money. They're not. Short is the starting line, not the prize. Bad letters, ugly patterns, no acronym pull, no way to pronounce it, any one of those can gut the value while the name still technically counts as "short."
So the short version, if there is one. LLL.coms hold value because they're scarce and flexible and understood everywhere. LLLL.coms hold value when they've got a strong pattern or acronym demand or they're easy to say. Five-letter names hold value the moment they stop reading as random and start reading as a brand.
The letters themselves aren't magic. The value is whatever the market can build out of them. Closest thing I've got to a mental model is an empty lot in a good location. One person sees a law firm, the next sees a fund, somebody after that sees a media company. The lot itself is nothing special. It's the location and the room to build that people are paying for.
Which is also why I still can't tell you with a straight face which random four-letter name is going to pop and which one's going to sit in my account collecting dust for years. I just know the good ones were never really random, and learning to feel that difference is most of the job.