Content Strategy

Start from demand

We use search volume and competition data alongside your own Search Console to find queries that are both valuable and winnable. High volume with impossible competition is a waste of a year. Low volume with a high value per click, which is common in B2B, is often the best place to start.

Demand also has to be read differently in 2026. On informational queries Google often answers with an AI Overview, and clicks to the pages below it fall sharply. Commercial and comparison queries, and questions specific enough that a general answer does not satisfy them, still send clicks. The demand map marks each query with whether an AI answer appears, so the plan favours queries where a visit is still likely and treats the rest as citation targets rather than traffic targets.

Clusters, not one-offs

A single article rarely ranks on its own. A cluster does: one hub page on the core topic and several supporting pages on specific questions, all linked to each other. That is how a site demonstrates depth on a subject, and it is the structure both Google and AI assistants reward when they pick a source.

A typical cluster for a service business is one service page, one pricing or cost guide, one comparison page, and three to five pages answering the questions buyers ask before they enquire. The order of publishing matters: the hub first, then the supporting pages, with links added in both directions as each goes live.

How the work runs, step by step

1. Content audit. Every existing page is scored on traffic, queries, links and age, and marked keep, improve, merge or remove. Most sites with more than a hundred pages have a third that should be merged or removed.
2. Demand map. Queries grouped into clusters with volume, difficulty, value and AI-answer presence, checked against what you actually sell.
3. Calendar. Clusters ordered by expected value and effort, with publishing dates and owners.
4. Brief. Each piece gets a brief: the target query and intent, what currently ranks and why, the questions the page must answer, the internal links in and out, and the structured data it needs.
5. Draft and review. Written by us on Growth and Scale, or by your writers from our briefs on Starter. Your subject expert checks every factual claim before publishing.
6. Publish and link. The page goes live with its links, schema and a date, and is submitted for indexing.
7. Measure and refresh. Each piece is reviewed monthly. Pages that stall are improved; pages that overlap are merged.

What we need from you

• Search Console access and, if you have it, GA4 with conversions, so value per query is grounded in your numbers.
• Access to a subject expert for about an hour a month: the questions customers ask, the objections they raise, and the facts only your team knows. That is the material assistants and readers cannot get elsewhere.
• Any original data you can share: anonymised job figures, prices, timelines, survey results. Original data is the most reliable way to earn links and citations.
• Brand and legal rules: claims you cannot make, words you avoid, regulated content.
• Sign-off within a week of receiving a draft, so the calendar holds.

What content strategy can and cannot achieve

It can build visibility for the queries buyers use before they enquire, make the site the obvious source on its topics, and give AI assistants something specific to cite. Over two or three quarters a well-built cluster usually becomes the largest source of organic leads on a small business site.

It cannot produce results in the first month, because new pages need to be crawled, indexed and evaluated, and it cannot win queries where the site has no authority yet; those need links and mentions alongside. It also cannot bring back the clicks on informational queries that AI Overviews now answer directly. We say which queries are traffic targets and which are citation targets before we write anything.

How it is measured

Per piece: the queries it appears for, impressions, clicks and position in Search Console; leads or sales it assisted in GA4; whether it is cited in AI answers on the prompt panel; and links or mentions it earned. Per cluster: total clicks and leads against the plan's estimate. The monthly report shows what each published piece earned and what we are doing about the ones that did not move.

Common failure modes

• Writing about the company instead of the buyer's question. The page ranks for the brand name and nothing else.
• Publishing a supporting page before the hub exists. Nothing to link to, so the cluster never forms.
• Scaled pages that differ only by a city or product name. Google's scaled-content policy names this pattern, and it puts the whole site at risk.
• Volume over depth. Ten thin pages on one topic lose to one thorough page.
• No refresh. A page written in 2024 with 2024 facts stops being cited as soon as a fresher source appears.
• Unreviewed generated text. Fluent, generic, occasionally wrong, and indistinguishable from the thousand other pages the assistant already has.

What we will not publish

Scaled pages that differ only by a city or product name, content that restates what already ranks without adding anything, and anything that makes a claim we cannot support. These are the patterns Google's spam and helpful-content guidance targets, and they put the whole site at risk. We use tools to research and draft, and a person edits and checks every page before it goes live.

Deliverables

How it runs

1. Audit and researchExisting content scored; demand and competition mapped to your business.
2. PlanClusters chosen, ordered and scheduled.
3. Brief and produceEach piece written to a brief and reviewed by your expert.
4. Measure and refreshPieces tracked; underperformers improved or merged.

Which plan includes it

Compare all plans →

Questions

Related