Business name recall test

1. Type up to three candidate names, one per line.
2. Read the list once when it appears, the way a customer glancing at a van would. Do not write anything down.
3. Do the arithmetic task while the timer runs. This is the part that makes the result honest.
4. Type back every name you can remember. Guessing is fine — near misses are informative.
5. Read which of your names came back exactly, which came back approximately, and which did not come back at all.
6. Then send the link to five people who have never heard the name. Their results are the ones that mean something.

If you take it yourself the result is heavily contaminated — you already know your own candidates, which is precisely the advantage a customer does not have.

Recognition is answering "have you seen this before?" It is easy, it is fast, and almost everything is recognised. Multiple-choice name tests measure recognition, which is why they produce flattering results.

Recall is producing the thing from nothing. It is much harder, and it is what a customer actually does. Three days after someone glanced at your van, or a neighbour mentioned you over a fence, that person sits down with a phone and has to reproduce your name from memory well enough to type it into a search box.

If they cannot, what happens next is not neutral. They search for your trade and your town instead, and land in a list of your competitors. Your recognition score does not help you there, because nobody is showing them a list with your name on it.

That gap is why this test uses free recall with an interference task, and why the interference matters — a recall test taken immediately measures nothing but the last thing you read.

Your names are shown among realistic competitor names — Apex Home Services, Meridian Trades, Pinnacle Pro Group, and similar. They are deliberately the kind of name a real business in your category might have.

That makes the test one of discrimination rather than raw memory. The question is not whether you can remember nine things; it is whether YOUR name is one of the ones that survives when it is competing with the sort of names it will actually compete with.

It also produces the most uncomfortable and most useful finding available here: if a filler name we invented stuck better than yours did, that is real information. The decoys were generated to be plausible and generic. Losing to one is not bad luck.

• Concreteness. A name you can picture outlives one you cannot. Ironclad, Redwood, Anchor and Kestrel are all imageable; Synergy, Optimal and Premier are not. The brain stores pictures better than abstractions and this effect is large.
• Distinctiveness. In a list of nine trade names, the one that is not built from Apex, Premier and Solutions is the one that survives. Being unlike your competitors is a memory advantage before it is a marketing one.
• Brevity. Recall falls off sharply past about three syllables. Longer names get shortened by customers, and you do not choose the abbreviation they land on.

What is not on that list is anything about typography, colour, or how the name looks in a logo — which is what most naming decisions are actually made on. A name is chosen looking at it and used by being said, heard and remembered.

This is a demonstration of the right test, not a study. A single trial with a single participant has enormous variance: someone who happens to be good at this will recall six names, and someone tired will recall two, and neither result says much about your name specifically.

It is also heavily contaminated if you take it yourself. You already know your candidates, you have been thinking about them for days, and you are motivated for them to do well. Every one of those works against the validity of your own result.

The useful version costs nothing: send the link to five people who have never heard the name. Count how many recall each candidate exactly. Three out of five is a strong signal; zero out of five is a strong signal in the other direction. That is a real if small experiment, and it is more evidence than almost anyone has when naming a business.

And one more honest caveat: recall is one property among several. A forgettable name attached to excellent work still gets referred — just by description rather than by name, which costs you the direct search and not the customer.

The test separates exact recall from approximate recall, and the approximate column is the interesting one. Someone typing "Riverside Plumbers" when your name is "Riverside Plumbing" has remembered you in every sense that matters emotionally and in no sense that matters commercially.

What happens to that person is that they type the wrong thing into a browser and get a search results page rather than your website. At that point they are looking at a list that includes your competitors, with their attention already committed and your name only approximately available.

Near misses cluster around specific features: names with a spelling ambiguity, names where a word could plausibly be singular or plural, names using a symbol that could be a word. Those are the same features the name strength test flags as spelling traps, which is a useful convergence — two different instruments finding the same problem.