Most naming guides are written by people who have never had to name anything real under real pressure. They give you frameworks that feel strategic but produce names like "Nexify" and "Veloxa" — the naming equivalent of a stock photo. Technically a thing. Utterly forgettable.
This guide is different because it starts with honesty: naming is hard, the rules are genuinely contradictory, and there's no formula that guarantees a great name. What there is, instead, is a set of approaches that have worked at scale, a filter you can run any name through, and a clear-eyed view of what "sticky" actually means in practice.
The Five Structural Types (And What Each Costs You)
Every startup name in history falls into roughly five structural categories. Understanding the category helps you understand what you're trading against what.
Invented words are the purest form of branding. Kodak didn't mean anything when George Eastman coined it in 1888 — he specifically wanted a word that was short, memorable, and could be pronounced in any language. Spotify is a blend of "spot" and "identify," but the compound is so tight the etymology disappeared. Häagen-Dazs is a completely made-up phrase that sounds vaguely Scandinavian (it's from the Bronx). These names are maximally ownable — you can trademark them completely, they don't carry baggage, and no competitor can claim the same territory. The cost is that you have to build all of the meaning from scratch. Kodak meant nothing until it meant cameras, film, and memory. That's an expensive educational journey.
Compound words take two known things and smash them together. Facebook is a face and a book. YouTube is you and a tube. Salesforce is sales and force. These feel obvious in retrospect because the component parts do real semantic work — you understand roughly what's happening without being told. They scale well because the component words are neutral enough not to trap you. The cost is that you're starting with borrowed meaning rather than creating new meaning, which limits how much the name can transcend its literal description over time.
Metaphors reach for something bigger. Amazon chose the largest river on earth because Jeff Bezos wanted to signal scale — and he chose it before the company was a bookstore, because he was naming the ambition, not the product. Apple chose a fruit associated with knowledge, creativity, and simplicity (there's a reason it was the fruit in the garden). Metaphors give you infinite headroom because they're not describing what you do — they're describing what you aspire to be. The cost is that metaphors require strong brand execution to make the connection land. If you call yourself "Summit" and nothing about your product experience feels like a summit, the name becomes ironic.
Clipped names compress something more literal into something more brandable. Intel was Integrated Electronics. FedEx was Federal Express. Reddit was Read It. These names are smart compression — they hold onto the meaning of the longer phrase while becoming pronounceable and typeable. They require a small upfront explanation ("what does that stand for?") but the explanation actually helps — it gives people a story to hold onto, which aids memory. The cost is that the clipping sometimes loses precision. FedEx no longer means just federal express delivery; it means anything fast and reliable. That's actually fine.
Founder and place names are the oldest naming convention and still the most underrated. Dell. Ford. Patagonia. Chanel. These names carry personal accountability in a way that invented names never can — you're putting your face on the product. They're also deeply authentic, which matters more now than it ever has. The cost is real: a founder-named company can struggle to feel bigger than one person, which affects exits, acquisitions, and leadership transitions. And "Patagonia" works because there's a genuine connection to the place; fake geographic names are just another flavor of made-up.
What "Sticky" Actually Means
Naming people talk about "sticky" names the way diet books talk about "eating clean" — it sounds like a real concept until you ask what it means.
Here's a working definition: a sticky name is one that a stranger can hear once, spell correctly, say aloud to a third party, and remember the next morning.
Notice what that definition does not include: cleverness, symbolism, founder intention, or brand strategy. Stickiness is a purely practical property. It lives in the gap between what you intended and what a stranger actually experienced.
The things that make names slippery — easy to hear, hard to retain:
The things that make names sticky:
The Grandma Test
This is the filter. Run every serious candidate through it.
The grandma test is simple: say the name out loud to someone who has no context for your company. Ideally someone outside your industry, ideally someone who isn't trying to be helpful. Tell them you're starting a company and the name is X. Then ask three questions:
The third question is the most interesting one. You're not looking for a correct answer — you're looking for a reasonable one. "Amazon" doesn't tell you anything about e-commerce, but it tells you the company is big and ambitious. "Stripe" tells you something clean and precise is happening. "Banana Republic" tells you something about adventure and exoticism. These aren't exact descriptions. They're vibes — and vibes are transferable.
If your grandma's answer to question three is "I have no idea" or "it sounds like a software company" (which is just a polite way of saying "I have no idea"), you have a name that's doing no semantic work. That's not necessarily fatal — Kodak did no semantic work either — but it means your brand has to do all the lifting.
If her answer to question one or two is wrong, you have a practical problem to fix before you have a brand problem to solve.
The Real Enemy Is Consensus
Here's the thing nobody tells you about naming: the names that feel most comfortable to a team are usually the ones that will disappear in a crowded market. They're comfortable precisely because they don't stand out. "Nexify" sounds like a startup because it sounds like every other startup.
The grandma test and the five frameworks above will help you generate and filter names. But the last judgment call — the one where you look at the finalists and pick one — requires accepting some discomfort. The name that wins is often the one that makes someone on the team say "is that a bit much?"
Stripe. Slack. Figma. All of these made someone nervous before they made everyone else jealous.
Pick the one that makes someone nervous. That's the one worth keeping.
