We hold more names than we use. In the accounts they look like an expense: a line item, renewed every year, attached to nothing. If you judge them as purchases, most of our domains are bad ones.
But they aren't purchases. When you register a name, what you actually buy is the right to build something under it later, at a price you fix today. The name does nothing on its own. The right to use it does.
The cost of waiting
Every option has a premium, and for a domain the premium is the renewal. You pay it every year so that you can decide later. The maths is simple:
The cost is small and known. The value is unknown and could be anything.
Where \(n\) is years held, \(r\) the annual renewal, and \(V\) whatever the name turns out to be worth to whoever eventually needs it, whether that is us or someone building the thing we once described to ourselves and never built. The left side is a rounding error that grows slowly. The right side is either nothing or a lot.
That imbalance is the whole decision. The question isn't is this name worth what we pay for it but can we imagine a world where \(V\) beats ten years of renewals. For a name that is short, plain and free of baggage, the answer is usually yes. For a clever one, it usually isn't.
What a name buys you
The more useful thing a good domain does is narrow your options. A vague name allows anything, which means it points at nothing. A specific one has a product inside it already: a scope, an audience, sometimes a way of making money. Owning it means that direction stays cheap to come back to.
We have started projects because we held the name and got tired of looking at it unused. That sounds like a bad reason to build something, but it turns out to be a decent one. The name had survived several years of us not needing it, which is a longer test than most ideas ever get.
When we let one go
Most of them never become anything. That is what the model expects. But you have to act on it, or a set of options quietly turns into a pile of renewal invoices.
So once a year we go through the list and ask one question about each name: if it expired tonight, would we buy it back tomorrow at the going rate? If not, we let it go. The ones that survive are always the ones we can't think of a replacement for.
We're not domain investors. We don't buy to resell and we've never had one valued. We're not collectors either, because a name we wouldn't build under is just a subscription. What we keep is a short list of names we expect to use, and a small yearly cost to keep that possibility open.