What we do
Search Engine Optimisation
Search optimisation is mostly a research discipline. The technical work matters, but it is the part that stops a good page failing — it cannot make a redundant page succeed.
Search engine optimisation is the work of understanding what people are trying to find and building the best available resource to satisfy that demand, then making sure it can be crawled, understood and trusted. Ferguson Media Group approaches SEO as research and information architecture first, with technical quality as the foundation that stops good work being wasted rather than as the point of the exercise.
Understand demand
Everything starts with what people are actually trying to find out, which is rarely what a keyword list suggests. A search phrase is the compressed form of a question, and the question is what has to be answered.
The useful research is therefore less about volume than about intent and gaps: what is being asked, what stage of a decision the person is at, what the existing results fail to cover, and where the good answer exists but is buried in something nobody will read.
Build the information architecture
Once demand is understood, the structure follows: which pages exist, what each is responsible for, and how they relate.
The most common structural failure is a set of pages that overlap. Four pages each half-answering the same question compete with each other, split whatever authority they earn, and leave a reader unsure which one to read. One page that answers it properly, supported by pages that genuinely cover different ground, outperforms all four.
Answer better than what is already there
A new page has to justify its existence against everything already ranking. "We wrote about it too" is not a reason for anyone to rank it.
The improvement has to be real: more accurate, more complete, better structured, easier to act on, better sourced, or covering something everyone else has skipped. Pages produced because a keyword had volume — with nothing new to say — are the bulk of what search engines have spent the past decade learning to demote, and language models are, if anything, less interested in them.
Establish evidence
Content that makes claims should show where they come from. Named sources, dated information, stated methodology, and figures that can be traced.
This has become more valuable rather than less. When systems are deciding which page to trust and which to quote, verifiability is one of the few signals that is hard to fake at scale. It is also simply the right way to publish.
Build authority coherently
Internal linking should reflect how the subject genuinely fits together, not how a template distributes links. A page that answers a question should link to the pages that answer the next question, in the sentence where that question arises.
Across a site, this produces topical coverage that is coherent rather than scattered: a set of resources that together demonstrate the site knows its subject. That is what "authority" means in practice, and it is built by covering a subject properly rather than by acquiring links to a thin page.
Maintain technical quality
The technical layer does not win rankings. It loses them.
- Crawlability — pages reachable, robots directives correct, no accidental blocking.
- Indexation control — one canonical URL per page, no duplicate paths, no parameter sprawl.
- Page speed and Core Web Vitals — fast rendering, stable layout, minimal JavaScript.
- Semantic HTML — correct heading order, real landmarks, content in the markup.
- Structured data — accurate JSON-LD that describes what is genuinely on the page.
- Internal linking — deliberate, contextual, and free of dead ends.
- Entity consistency — the same organisation, person and subject described the same way everywhere.
- Sitemaps and status codes — accurate, current, and honest about what exists.
Measure and improve
Search work should be reviewed against evidence: what is being found, what is being read, what is being acted on, and what should be rewritten, merged or deleted.
Deletion is the underused option. A site carrying two hundred pages that nobody wants is harder to understand — for a reader, a crawler and a language model — than the same site carrying forty that people do.
Common questions
How long does SEO take to work?
In a competitive category, months rather than weeks, and the first movement is usually in what a site is found for rather than in how much traffic it receives. Work that produces immediate dramatic gains is usually exploiting something temporary.
Is SEO finished now that AI answers questions directly?
No, but it is changing shape. Fewer people will visit a page to get a fact they can be told directly, which mostly harms pages whose only function was repeating easily available facts. Work that produces genuinely useful, well-structured, verifiable resources is if anything more valuable, because that is what both search engines and AI systems draw on.
The practical response is not to abandon search work but to stop producing the kind of page that was only ever winning by default.
Is optimising for AI search different from SEO?
It overlaps heavily. Clear structure, direct answers, factual accuracy, visible sourcing and consistent entities serve both. The differences are in emphasis — answer-first writing, extractable facts and internal consistency matter more when a system is summarising rather than ranking. See LLM and AI search optimisation.