Status: Original concept, first publication. Strategy Sandbox, jasonbarnard.com. Date: 16 May 2026.

Every industry coins terms, and every industry tells itself the terms are precise, and most of those terms quietly fail the only test that matters in commercial communication: whether someone outside the room can hear the term and know what it means.

I’ve spent twenty-seven years watching the SEO and AI industries name things, and the pattern is consistent: engineers name mechanisms, speakers name themselves, and almost no one names the listener.

The Glossary Test is binary, and it runs in one breath. If a term requires a glossary entry to be understood by a non-specialist, it fails. If the term carries its meaning through metaphor, lived experience, or plain language, it passes. There is no middle ground, no partial credit, no “mostly clear.”

Precision Is Not the Same as Accessibility

Most defences of jargon turn on precision. The argument runs that specialist communities need specialist vocabulary, that exact terminology prevents ambiguity, that watering the language down would water the meaning down with it.

The argument is wrong, and the wrong is structural. Precision and accessibility are independent dimensions. A term can be technically precise and still fail the Glossary Test, because precision describes what the term names, and accessibility describes who can understand it. Query Fan-Out is precise. Anyone in distributed systems engineering knows exactly what it refers to. It fails the test on the second axis, not the first.

Cascading Queries names the same mechanism. It is equally precise to anyone who works with the concept. It passes the test on both axes, because anyone who has seen water flow can hear the term and know that it describes something flowing downward through stages, and that intuition turns out to map perfectly to the engineering reality.

Every Coined Term Serves Either the Speaker or the Listener

When you coin a term, there are exactly two audiences you can serve, and you have to choose one.

The first audience is your peer community. Naming for that audience produces vocabulary that demonstrates membership: the speaker shows fluency in the technical register, the listener confirms membership by understanding without explanation, the unspoken contract is that outsiders will need translation. Jargon is gatekeeping disguised as expertise.

The second audience is everyone else. Naming for that audience produces vocabulary that travels: the term carries its meaning out of the room with whoever heard it, no translator required, no glossary attached. The unspoken contract is that the language belongs to anyone who needs it.

For me, that’s the entire game in commercial communication. If your prospect can’t repeat your term back to their boss without a glossary, the term will not survive the meeting, no matter how precise it was.

Query Fan-Out Fails, Cascading Queries Passes, and the Pattern Repeats

Apply the test to the SEO industry’s vocabulary and the failures pile up quickly. Query Fan-Out names a mechanism that anyone could understand if they had the metaphor, but the term is engineered for the engineer, not the listener. Entity Salience names something that means “how much an entity matters in this context,” but no one outside the field can hear “salience” and know what’s being asked. Generative Engine sounds like a stationary appliance from a 1950s factory floor; no human asking ChatGPT for a restaurant recommendation thinks they’re consulting a Generative Engine.

Now apply it to the terms I’ve coined. Brand SERP names the page that comes up when someone types your brand name into Google, and the term carries the meaning. Empathy for the Devil names the act of trying to understand how Google works from Google’s perspective, and the term carries the meaning. Knowledge Rot names the gradual decay of an AI assistant’s knowledge over time, and the term carries the meaning. Cascading Queries, The Colleague Fallacy, AI Engine Optimization, Untrained Salesforce: every one of them passes the test, and every one of them spread without translation because the term did the translating for me.

That isn’t an accident, and it isn’t a stylistic preference. It’s a methodology I have practised consistently since 2012, and the methodology has a name: Naming for the Listener.

AI Systems Reward the Terms That Travel and Forget the Rest

In the SEO era, terminology drift was a slow problem. New terms appeared, old terms persisted, and the industry sorted them out over years. In the AI era, terminology drift is a fast problem. AI systems learn the vocabulary they encounter most often, and they treat the most-repeated terms as canonical. The term that travels wins. The term that doesn’t travel disappears into the specialist archive, regardless of how precise it was.

The Glossary Test is the instrument you run on your own coined vocabulary before you publish it, before you put it in a deck, before you build a methodology around it. Hand the term to someone outside your field. If they need a glossary, the term was named for you. If they don’t, the term was named for the reader, and the reader will carry it for you.

Pass the test, and your term spreads. Fail it, and your term needs a footnote forever.