For a restricted category, organic is often the only channel a policy update cannot switch off. Technical foundation, topical clusters, and pages written to be lifted verbatim into an AI answer — because a growing share of searches now end without a click, and the cited page wins them.
Why this matters more here
Paid media is rented and organic is owned. An ad account can be switched off on a Friday for a
policy reason nobody could have anticipated; a domain that ranks cannot. For a clinic, an
immigration practice or an exchange, organic is frequently not a supporting channel but the only one
that cannot be taken away by somebody else's decision.
It is also slower than anything else we do, and we would rather say that at the start than
explain it in month four. A domain with existing authority can move within a couple of quarters. A
new domain in a competitive category is a multi-year project. Anyone quoting a timeline before
looking at the domain is quoting a sales figure.
Technical foundation
Crawlability, indexation, architecture, internal linking, speed, and correct handling of multiple
languages and markets. On sites that grew in layers the usual finding is not an exotic bug but
thousands of near-duplicate pages competing with each other and a crawl budget spent on parameters.
- One canonical version of every page, and hreflang that is reciprocal and self-referencing where
a site is multilingual.
- Structured data that describes what the page actually is, so a machine reader does not have to
infer it from markup.
- Server-rendered content. Anything that exists only after JavaScript runs is a bet on a crawler
that may not execute it — and the AI crawlers are markedly worse at this than Google.
Topics, not keywords
We build clusters: one page that answers the broad question properly, supported by pages
answering the specific ones, linked so the relationship is obvious to a reader and to a crawler. The
research starts from what people actually type — including the unglamorous long tail where most
categories still have room — and from what competitors rank for that you do not.
GEO: the half that is new
A growing share of searches now end in a generated answer rather than a click. That is not a
variation on ranking, it is a different competition: the question is no longer whether you appear in
a list of ten links, it is whether your sentence is the one the model repeats. Ranking first and
being cited are related but not the same, and sites regularly do one without the other.
What gets lifted into an answer is specific and self-contained: a direct sentence that answers the
question, detail after it, on a page a crawler can read without executing scripts. In practice that
means:
- The question phrased the way people ask it, and the answer in the first sentence — a paragraph
that builds to its conclusion gets skipped.
- Concrete numbers, named conditions and stated dates rather than hedged generalities. A model
quoting you needs something quotable.
- Structured data confirming what the page claims to be, and an author or organisation a machine
can identify.
- Being reachable at all — a robots.txt that blocks GPTBot, ClaudeBot and PerplexityBot is a
decision to be uncitable, and a surprising number of sites made it by accident.
None of this is at the expense of ranking. The discipline that makes a page quotable is the same
one that makes it useful to a human in a hurry, which search has rewarded for years.
Links
The part of this work with the worst reputation, largely deserved. We are conservative: digital
PR, genuine partnerships, and placements on sites with real traffic. We do not buy from link farms
or private blog networks. Those are the cheapest links available and the ones most likely to leave a
domain harder to rescue than it was to build.
What you get
A technical audit you can hand to a developer, a content plan tied to actual demand, work
delivered monthly, and reporting that separates what moved because of the work from what moved
because the market did.
