x
Ranking put you in a list of ten and let the reader choose. Citation puts you inside a sentence, and there is no list.
That sounds like a narrowing, and it is. But it changes what wins, and that matters more than how many slots there are.
Ranking rewarded scale. Citation rewards clarity — and clarity has no economies of scale. Why the new mechanism quietly favours the one-person business.
A list of ten rewards whoever can sustain the effort of being on lists: more pages, more links, more domain authority accumulated over years. It is a volume contest, and volume contests are won by whoever can spend the most for the longest.
A sentence rewards whoever is easiest to describe accurately. That's a clarity contest. And clarity, unlike volume, has no economies of scale — a fifty-person marketing team cannot be fifty times clearer than one person who knows exactly what they do.
If anything the advantage runs the other way, and this is the part worth dwelling on.
Because specificity survives fewer approvals.
Large organisations produce vague copy for structural reasons, not stupid ones. Every specific claim on a large company's website has passed legal, brand and three stakeholders who each removed the bit that exposed them. "We help businesses transform" is what's left when everyone has protected themselves. It is unfalsifiable, which is the point — and unfalsifiable claims are exactly what a model has nothing to do with.
A one-person business has none of that friction. You can write "I do monthly bookkeeping for creative agencies with five to twenty staff, and I'm not the right fit under about £200k turnover" on a Tuesday afternoon, because there is nobody to ask.
That sentence is worth more in a citation economy than a year of content marketing, and it costs an afternoon. The GEO research found the strongest visibility levers were statistics, source citations and quotations — all of which are forms of committing to something specific enough to be checked. Committing is precisely the thing a committee cannot do.
So the structural read is: the mechanism that replaced ranking happens to reward the one advantage a solo operator actually has. Not for anyone's benefit — it's a side effect of how these systems assemble answers — but real nonetheless.
It reorders what's worth your time, mostly downward.
Volume goes down the list. The reflex on hearing "AI can't find you" is to publish more. But if the lever is describability rather than surface area, five more pages nobody can lift a clean claim from add nothing. One page rewritten to answer a real question, with a number and a limit in it, adds something. This is a relief if you take it seriously: it means the thing you can't do — sustain a content operation alone — is no longer the thing that matters most.
Positioning goes up. Not positioning as a branding exercise, but the plain sentence: what you do, who for, what changes. If you can't say it without adjectives, no page structure rescues it, and that's Sunday's post.
Saying who you're not for goes right to the top. It's the most quotable line most businesses could write and almost nobody writes it. It's also the same move as disclosing what AI does in your work before anyone asks — volunteering the limit is what makes everything else credible, to a reader and to anything reading on their behalf.
It's worth being careful here, because there's a comfortable story available and it isn't entirely true.
The bad trend is real. Click-through to publisher sites fell hard — Ahrefs measured 58% for top-ranking pages where an AI Overview appears — and the losses fall hardest on small publishers, who have the least ability to absorb them. Nobody should pretend a mechanism that removes clicks is good news for people who needed clicks.
The honest position is narrower: if you were never winning the volume contest, its replacement is a better contest for you. That's not the same as saying the change is good. For a business that had built a real search channel, this is straightforwardly a loss.
There's also no guarantee the current shape holds. Assistants are non-deterministic, engines change without notice, and the incentive for platforms to keep users inside the answer is not going away — which could mean fewer sources cited, not more.
Which is why the only sane version of this advice is the version that pays off either way. A clearer sentence, a sourced claim, a stated limit and a page that answers first are all things worth having if none of this materialises. If your marketing plan only works when the forecast is right, it isn't a plan.
Ranked was a competition for attention. Cited is a competition for accuracy. Small businesses have always been better at the second one. It just never counted for much before.
For the audit that finds where your week actually leaks — free, 5-min read: The 10-Hour Audit.
Related Decodes
AI the Brief (@aithebrief) — AI advice you can act on and trust, for time-starved founders.
Not marketing advice. Assistant behaviour changes without notice; figures are dated and sourced so you can re-check them.
TL;DR: Ranking put you in a list of ten and let the reader choose; citation puts you inside a sentence with no list. That changes what wins. A list rewards volume — more pages, more links, sustained spend — which is a contest won by whoever spends longest. A sentence rewards being easy to describe accurately, and clarity has no economies of scale. Large organisations write vague copy for structural reasons: every specific claim has passed legal, brand and three stakeholders who each removed the exposing part, leaving unfalsifiable phrasing that a model has nothing to do with. A solo operator can write a specific, checkable sentence on a Tuesday because there is nobody to ask. So volume drops down the priority list, positioning rises, and stating who you're not for goes to the top. But be honest about the trend: a 58% CTR drop is a real loss for anyone who had built a search channel, and nothing guarantees the current shape holds — which is why the only sane version of this advice is the version that pays off even if the forecast is wrong.
