Name three brands of olive oil. Most people manage one, maybe two, then stall. Now think of three brands you saw advertised in the last week, in any category at all. That list comes faster and longer, because you are not retrieving those names from memory, you are recognising them from a cue you just saw. The names people pull from memory most easily do two things, they break their category's pattern and they are built from words the mind already stores.
That gap, between the names you can produce on your own and the names you can recognise when prompted, is the difference between recall and recognition. A brand name does two different jobs, and they are not equally hard. Being recognised on a shelf or in a search result is cheap, because the name is right there to jog you. Being recalled, named out loud from nothing when a friend asks what you use, is expensive, and it is the job that drives word of mouth and the typed-in search. The names people call unforgettable are the ones that clear the expensive bar.
Why some names clear it is not luck. Two forces decide most of it, and both can be measured. A name that breaks its category's pattern is recalled better than one that blends in, and the mind retrieves real words it already stores far more reliably than invented ones it has never seen. Read together, those two forces are not contradictory advice. They are a budget, and the rest of this post is how to spend it.
Recall is not recognition
Memory is not one thing, and naming advice almost always optimises the easy half while calling it the hard one.
Recognition asks whether you have seen something before. It is what happens on a category page, in a search result, on a shelf where every option is in front of you. The name only has to feel familiar enough to pick. Recall asks you to produce a name with nothing in front of you. It is unaided, with no list to choose from, and it is what happens when someone asks a friend for a recommendation, or when a buyer types a brand straight into a search bar instead of browsing. Recall is harder because the mind has to produce the name from nothing.
The distinction matters because the two failures look nothing alike. A name that wins recognition but loses recall does fine in any test where you show people the name and ask whether they like it, then quietly underperforms in the real world, where nobody is holding the name up at the moment a buyer wants to mention it. Most naming research, and most of the advice built on it, measures recognition, because recognition is easy to test. The expensive failures are recall failures, and they hide from the tests that only check recognition. Everything that follows is about the harder half, the part that decides whether a name travels on its own.
Distinctiveness buys recall
The first force is one of the oldest findings in the science of memory. In 1933 the psychologist Hedwig von Restorff showed that when a set of items is mostly alike, the one that stands out is recalled far better than the rest, an effect that still carries her name. Put a single red word in a list of black ones and people remember the red word. The brain spends more attention on the thing that breaks the pattern, and attention at the moment a memory is formed is what makes it retrievable later.
A name that breaks its category pattern gets its own slot in memory and is recalled by name. Names that sound alike share one slot and come back only as an approximation.
The branding consequence is direct. A name that sounds like the rest of its category is the black word in the list. It is encoded with less attention because nothing about it is surprising, and it is recalled worse for the same reason. This is not oddness for its own sake. The useful kind of distinctiveness is being the item the mind has a reason to file separately from its neighbours. On a shelf where every name reaches for the same sound, as an analysis of five thousand DTC names shows, the name that does something different is the one that gets its own slot in memory rather than sharing one with four competitors.
This is the recall side of distinctiveness, which is a separate payoff from the preference side. A polarizing name can win on preference by breaking the pattern, but even setting preference aside, distinctiveness earns its keep at the level of plain retrieval. The name that breaks the pattern is easier to pull back out of memory, and on a crowded shelf that is the difference between being named and being approximated.
The cost of a word the mind has never seen
The second force pushes the other way, and it is the number the conversation about distinctive names usually leaves out. In a 2002 study, Lerman and Garbarino tested how well people recalled brand names that were real words against names that were invented. Real-word names were recalled correctly 68.8 percent of the time. The invented ones were recalled 38.1 percent of the time, less than half as reliably from unaided memory.
From unaided memory, real-word names were recalled 68.8% of the time and invented ones 38.1%. Shown the name instead of asked to produce it, the gap shrinks sharply.
The reason is the same mechanism that makes distinctiveness work, seen from the other side. Recall depends on what the mind already stores. A real word arrives with a memory trace already built, years of prior exposure the brain can reach for when it grasps for the name. An invented word has no such trace. The first time anyone meets it is the day they meet the brand, so there is nothing to reach for, and the name has to build its own trace from scratch through repetition.
Two parts of that study keep the finding from becoming a blanket case against invented names. First, the gap is specifically a recall gap. When the same people were tested on recognition, shown the name and asked whether they had seen it, the difference shrank sharply, because recognition leans on the cue in front of you rather than on retrieval. An invented name is hardest in exactly the unaided situation, the word-of-mouth moment, and far less penalised on the shelf. Second, meaning is not what drives recall. Arbitrary real words used as brand names, an Apple or a Grey Goose, were recalled about as well as names that describe the product. A real word does not have to relate to what you sell to carry its memory advantage. It only has to be a word the mind already holds. The separate cost of a name that is hard to say, rather than hard to recall, is its own processing-fluency problem.
A name is anchored, not remembered alone
A name is rarely recalled by itself. It rides on everything attached to it, the colour, the logo, the sound, the short form, the shape of the package. Work from the Ehrenberg-Bass Institute on what Jenni Romaniuk calls distinctive brand assets describes how those consistent cues build the retrieval network a name lives inside. Each cue is another route back to the name. The more of them a buyer has met, consistently, the more ways their memory can arrive at the brand when it matters.
This is how an invented name pays down the recall cost it starts with. A made-up word begins with no trace, but every consistent exposure lays a little more of one, and every distinctive asset anchored to it adds another handle to grab. A coined name with five years of consistent colour, sound, and a clean short form is no longer a word the mind has never seen. It has become one the mind has seen many times, in one steady shape. The short form matters on its own. A long name almost always gets shortened in everyday speech, and Federal Express became FedEx in conversation before the company made the nickname official. If buyers are going to compress the name anyway, the one that survives compression cleanly keeps its memory equity, and the one that gets mangled loses it.
The memory budget
These settle into one decision. A distinctive name is recalled better than a conventional one. An invented name is recalled worse than a real one. Distinctive assets and repetition can pay down an invented name's recall cost over time. The question a founder actually faces is not which rule to obey but how to spend a budget, because the rules pull against each other.
The two forces are a budget. A real word starts with recall and may not be ownable. An invented word is ownable but spends recall, which repetition, a clean short form, and distinctive assets pay back down over time.
An invented name spends recall. It is harder to retrieve on day one, and it will cost real time and repetition to build the memory trace a real word came with for free. What it buys is ownership. An invented word can be trademarked cleanly, defended without a fight, and owned outright on a crowded shelf in a way no common word can be. So the trade is recall now for ownership later, and whether it is worth making depends on the category. Where every plain word that fits the product is already taken or too generic to defend, which is the normal state of tech and much of consumer software, the invented name is the only ownable option and the recall cost is the price of admission. Vercel and Pentium are not accidents. They are names built where no real word was available to own.
Where a distinctive, ownable real word is still on the table, paying the coined-word recall cost is a choice that needs a reason. And if you do spend it, spend the rest of the budget paying it down on purpose. Repeated, consistent exposure builds the trace, the long-documented effect of mere exposure on familiarity and liking. A deliberate short form keeps the name intact when buyers compress it. Distinctive assets, held consistent from the first day, give recall more routes home. None of that happens by itself. It is the work that earns back what an invented name borrows.
What this means for your shortlist
The transferable move is to decide recall against ownership on purpose, rather than letting an available domain decide it for you. Three checks make the trade visible before you commit.
Ask whether a stranger, shown your name once, can produce it unaided a day later. If they cannot, you are buying an invented name's recall cost, so make sure you are also buying its ownership upside. Ask whether the name breaks its category's sound or echoes it, because echoing the shelf forfeits the recall that distinctiveness would have earned. And ask whether there is a clean short form and a distinctive asset to anchor, because those are how the name pays its cost down once it is live. A name that is distinctive, ownable, and anchored is spending its budget well. A name that is conventional, unownable, and bare is paying full price for nothing.
Frequently asked questions
Why are some brand names more memorable than others? Two measured forces decide most of it. A name that breaks its category's pattern is recalled better than one that blends in, the isolation effect first shown by von Restorff in 1933. And the mind retrieves real words it already stores far more reliably than invented ones it has never seen. Distinctiveness and familiarity drive recall, not luck.
Are made-up brand names harder to remember than real words? For unaided recall, yes. In one study, real-word names were recalled about 69 percent of the time against 38 percent for invented ones. The gap shrinks sharply for recognition, when the buyer can see the name, so a coined name is penalised most in the word-of-mouth moment and least on the shelf. In exchange it can be owned and trademarked cleanly.
What is the von Restorff effect in branding? The isolation effect, described by Hedwig von Restorff in 1933, is that the item which stands out in a uniform set is recalled best. Applied to names, a brand that sounds like its whole category is encoded with less attention and recalled worse, while one that breaks the pattern gets its own slot in memory. The lever is attention at the moment the memory forms, so a name that is usefully different is not the same as one that is merely strange.
Does a brand name have to mean something to be remembered? No. Arbitrary real words used as brand names, like Apple or Grey Goose, are recalled about as well as names that describe the product. What helps recall is that the word already lives in the buyer's memory, not that it relates to what you sell. Description is a separate question from retrieval, which is why a meaningless but real word still outpaces an invented one for unaided recall.
BrandNames weighs recall against ownership on a founder's real shortlist, and reads each name against the category it has to be remembered in. The doors open soon. Drop your email to get one note when they do.