Naming advice
A practitioner's playbook
This is the short version of what I have learned watching founders name things, naming a few things badly myself, and reading more naming research than is healthy for one person. There are deeper articles linked at the bottom. This page is the playbook.
1. Stop describing what your product does
The descriptive name feels safe in the room where you pick it. It is the most expensive safety you will ever buy. AltaVista, Infoseek, Ask Jeeves — all descriptive names, all dead. Google, Stripe, Vercel — none describe what the product does, and all are now category-defining. A descriptive name is a borrowed name. You are renting credibility from your category and ceding the right to ever become something else.
2. Pick the embarrassing name
If the name feels obvious and comfortable, it is probably not the right one. The names that age well were embarrassing on day one. Sonos sounded clinical. Azure sounded dumb. Windsurf sounded playful for a serious coding tool. Each of those teams overrode their own discomfort. Each of them is now defining a category.
3. Optimise for the second customer, not the first
The first customer will adopt your product because of what it does. The next 999,999 customers will adopt it because of what it feels like to recommend. Pick a name you would be happy to say out loud in a meeting with your most conservative-looking colleague. If the name fails that test, the word of mouth fails too.
4. Compound real words to multiply meaning
Two real words combined create three things. Wind plus surf creates wind, surf, and the experience of being expertly carried. Black plus berry creates black, berry, and the addictive quality of a precious found object. Compounds outperform invented words consistently. They are the highest-leverage naming pattern available in English.
5. Choose the opening phoneme deliberately
The first sound of your name does some of the brand work for you. V is the most energetic opener in English. Z cuts through noise. B signals trust. X signals frontier. K signals sharpness. M and N signal warmth. Pick the phoneme that matches the feeling you want the brand to create, then build the name around it.
6. Test the phone test
Can you say your domain over the phone to a stranger without spelling it? If not, you will lose conversions forever. Hard-to-spell names cost direct traffic. Hard-to-pronounce names cost word of mouth. Both costs are small per interaction and enormous in aggregate.
7. Test the trochaic rhythm
English speakers prefer two-syllable words with stress on the first syllable. Apple. Google. Facebook. Twitter. Netflix. The pattern is the rhythm of nursery rhymes, and a child can learn a trochaic name before they can read it. Names that fight this rhythm are working against the language.
8. Get to a thousand names before you evaluate any of them
Most naming processes produce 50 candidates and immediately start evaluating. That is too few and the wrong instinct. Generate widely, in different phonetic territories, with different lengths and shapes. The right name is usually inside the second thousand candidates, not the first fifty.
9. The polarisation test beats the consensus test
If everyone in the room loves your shortlist, the shortlist is too safe. The names that win the market are the ones half the team hates. Polarisation is signal. Consensus is the absence of signal.
10. Do not register a domain you found through a search tool you do not trust
If you searched a domain on a free tool and it is still available 24 hours later, you got lucky. Register it now, before the tool's owner registers it for you. Use search tools that do not own domain inventory. There is a longer version of that argument on the no-front-running page.