GEO & SEO content strategy: planning content that gets found and cited

An SEO & GEO content strategy decides which topics to cover, in which format and on which page, based on what your personas are actually searching for. It turns your editorial strategy into a concrete plan: intents, pillars, clusters and facets, prioritised by probability of gain. One requirement now comes on top: being found on Google, but also being cited in the answers of generative engines. This guide shows how to move from keyword to intent, build the plan and measure what it delivers.

What an SEO & GEO content strategy is

The SEO & GEO content strategy turns the "what" of the editorial strategy into topics and formats driven by search data. It does not redefine objectives, audiences or tone: it starts from that framework and tests it against the demand observed on Google and in AI answers.

How it differs from the editorial strategy

The two levels complement each other without overlapping. Each answers a different question and produces a different deliverable:

• the editorial strategy sets the why and the what: objectives, audiences, positioning, thematic pillars and tone;
• the SEO & GEO content strategy translates these pillars into specific topics, each tied to a search intent, a format and a target page;
• the editorial plan organises execution: topics, formats, owners and dates, with the editorial calendar as its view over time.

When a content strategy starts redefining targets or tone, it is a sign that the upstream framing is missing. In that case, go back to the editorial strategy before planning.

What it actually produces

The main deliverable is a content mapping linked to a prioritised backlog. Each line combines a keyword or question, an intent, a level in the architecture (pillar, cluster or facet), an existing page or one to create, and an estimate of the expected gain. This mapping then feeds the editorial plan, which assigns owners and dates.

Content marketing and SEO are designed together from the outset. Every piece of content must bring value to the reader and answer measurable demand; otherwise it stays invisible or does not convert.

From keyword to intent: choosing the right format

A keyword says nothing about the content to produce until you have identified the search intent behind it. The same expression can call for a guide, a comparison or a demo page, depending on what the reader wants to achieve. The table below links each intent to the format and page type that serve it.

Reading intent in the search results

The most reliable way to qualify an intent is still to look at what Google already displays for the query. Several signals guide the decision:

• the nature of the pages ranking on the first page: guides, product pages, comparisons or directories;
• the presence of an AI Overview, which signals a question Google considers worth answering directly;
• related questions and related searches, which reveal the sub-intents to cover;
• the wording of the query: "how", "best", "price" or "alternative" point to different formats.

Splitting intents by persona rather than by quota

There is no ideal split between informational and commercial content. A B2B software vendor whose buyer researches at length will need many guides; an offer its target already knows will favour comparisons and solution pages. The split is decided persona by persona, based on the queries actually observed and the stages of the buying journey the site covers poorly.

The pillar, cluster and facet plan

The three-level architecture organises topics to cover a semantic territory without pages competing with each other. It draws on the semantic cocoon: a pillar page covers the broad topic, cluster pages go deeper into each major variant, and facet pages answer the most specific queries.

A B2B SaaS example: project management software

Take a vendor selling project management software to mid-sized teams. Its content plan could be organised as follows:

The pillar targets the broadest and most contested query: it gains legitimacy as the clusters and facets are published and reinforce it. Facets, being more specific, usually rank faster and provide the first signals.

Connecting the levels through internal links

The architecture only works if the pages connect with each other. Internal linking follows a few simple rules:

• each cluster and each facet links to its pillar with descriptive anchor text;
• the pillar links to all its clusters, which link to their facets;
• pages at the same level link to each other when they genuinely complement each other for the reader;
• a given intent is carried by one page only, to avoid cannibalisation.

Editorial calendar and SEO content plan: two distinct tools

The two documents are often confused, yet they do not answer the same question. The SEO content plan says what to produce and why; the editorial calendar says when and by whom. The table below sets out what separates them.

Key components of an SEO content plan

A workable plan brings together at least the following elements. The guide to the editorial plan with an example calendar then details how to lay it out week by week.

• the prioritised inventory of keywords and questions to cover;
• the target page for each topic, existing or to be created;
• objectives for each piece of content: target ranking, presence in AI answers, expected conversion;
• the type of production: in-house writing, external writer or supervised AI-assisted production, always reviewed;
• the tracking KPIs to collect after publication.

Seasonality matters too: a topic tied to a key moment in your market, such as a budget close or a trade show, must be published early enough to have time to rank.

AI citability as a planning criterion

Generative engines such as ChatGPT, Perplexity, Gemini or Google's AI Overviews answer by citing sources. A brand is cited when it is named or linked in the answer. At the time of writing, AI Overviews and AI Mode have been available in France since summer 2026, and Search Console counts them in the Performance report without distinguishing them.

Organic ranking remains the gateway: AI Overviews cite the organic top 10 in 99% of cases (Squid Impact, 2025). The SEO content strategy and the GEO strategy are therefore not planned separately; further benchmarks are gathered in the SEO statistics.

Starting from the questions asked to AI

Users phrase long, contextualised questions, often closer to a conversation than to a keyword. At the planning stage, each topic in the plan benefits from being enriched with:

• the full questions your personas ask about the topic, including comparisons between solutions;
• the current AI presence on the keyword: brand named, domain cited, domain as source or absent;
• the sources the engines already cite, to understand the level of proof expected;
• the third-party platforms that weigh in citations, such as LinkedIn, YouTube or Reddit.

Planning answer pages and proof

A citable page answers quickly, then develops. Structure matters: pages with an H1-H2-H3 hierarchy are 2.8x more likely to be cited (State of AI Search, 2025). So does format: 80% of pages cited by AI use lists (State of AI Search, 2025).

For each piece of content, the plan therefore specifies the direct answer expected at the top of each section, the proof to gather (sourced data, examples, in-house expertise) and the extractable formats to include, such as tables and lists. The GEO content strategy guide details the citability checklist before, during and after writing.

Prioritising by probability of gain

A complete content mapping always contains more topics than the team can handle. Prioritisation decides the order, and it is based on expected value rather than search volume alone.

Estimating the expected gain of each topic

The expected gain combines three dimensions: the value of the targeted traffic for the business, the probability of ranking given the competition and the site's authority, and the effort required. A modest but accessible topic close to conversion often comes before a broad query where the site has no chance in the short term.

Fanny Magréault, Product Manager at Naturalforme, describes this logic: "The solution makes it easier to research trends and lets us choose the best keywords according to the probability of gain."

Reading each page's signals before deciding

Optimising existing content is a decision in its own right: you compare the expected gain of an update with that of a new page. Search Console and AI presence data guide the trade-off:

Adapting the content strategy to your sector

The method stays the same from one sector to another, but the weight of formats and pages changes with the acquisition model:

• B2B services and software: guides, comparisons and case studies dominate, because the buying committee researches at length before contacting sales;
• e-commerce: category pages and facets carry most of the commercial intent, complemented by buying guides;
• local business: one page per area served, with content specific to each location rather than duplicated text;
• media and publishers: a news-driven calendar, backed by evergreen content that stays relevant over time.

In every case, AI citability applies: comparison and selection questions are precisely the ones buyers ask generative engines.

Measuring what the content strategy delivers

A content strategy is judged by its effects, page by page and topic by topic. Measurement must be planned from the outset, with a baseline recorded before publication.

Sources and KPIs to track

Three sources cover most of the tracking:

• Google Search Console: impressions, clicks, average positions and queries per page, AI Overviews included but not distinguished;
• Google Analytics 4: engagement rate, conversions and session sources, including chatgpt.com and perplexity.ai for traffic from AI answers;
• AI presence tracking: citations per question and presence status per keyword, recorded at regular intervals.

Comparing before and after

Each update or new page is compared with its baseline over an equivalent period: rankings, clicks, citations and conversions. This comparison shows a correlation, not causation, since other factors change in parallel. It is nonetheless enough to decide whether to continue a cluster, rework a page or reallocate effort.

The lessons then feed back into the content mapping: a topic that progresses justifies its facets, and a page that stagnates goes back into the backlog with a new diagnosis.

Depending on your situation

Your starting point depends on where you stand:

• no editorial framework yet: first set objectives, personas and pillars with the five-step editorial strategy method mentioned in the introduction;
• a content mapping ready for production: organise editorial content production, from the brief per persona to review;
• writers to align with Google's expectations: share the rules of SEO copywriting in every brief;
• pages struggling to rank despite good content: check the technical and authority fundamentals described in the SEO guide;
• good Google visibility but few AI citations: apply the citability checklist presented above to your best-ranked pages.

To estimate the probability of gain keyword by keyword and page by page, and decide between updating and creating, the Incremys opportunity analysis module ranks the topics in your backlog by expected gain.

FAQ on SEO content strategy

SEO content strategy vs editorial strategy: what is the difference?

The editorial strategy sets the why and the what: objectives, audiences, positioning, pillars and tone. The SEO & GEO content strategy turns that what into topics, intents, formats and pages, based on search data from Google and AI answers. The former rarely changes; the latter evolves with each measurement cycle.

How many pages does a cocoon need?

No number applies to every site. A cocoon has as many pages as there are distinct, useful intents in the targeted territory: one pillar, one cluster per major variant, and one facet per specific query that deserves its own answer. Start with the pages offering the best expected gain, then expand based on results.

Do you need one pillar per keyword?

No. A pillar covers a broad topic and groups all queries sharing the same general intent. Close variants are handled on the same page, distinct variants in clusters. Creating one pillar per keyword multiplies pages competing for the same intent and dilutes the site's authority.

How can you tell whether a topic will be cited by AI?

No method guarantees it, but there are clues. Ask generative engines your personas' questions, and note the sources they already cite and your presence status on the keyword. A topic where your page ranks well, is well structured and is backed by proof has a better chance of being picked up.

Update or create?

Compare the expected gain of both options. A page already close to the first page, outdated or poorly aligned with the intent usually comes before a new page. A topic missing from the site, or a page that would have to serve two different intents, justifies new content instead. The backlog mixes both types of task.

How long should SEO content be?

The right length is the one that fully covers the intent, without padding. Look at the pages already ranking for the query to estimate the level of detail expected, then aim for a clearer, better-supported answer. A short, precise page can beat a long text that scatters the reader's attention.

How often should you publish?

The pace depends on your capacity to produce quality content that is reviewed and tracked. A regular, sustainable cadence beats peaks followed by long pauses. Also set aside part of your time for updates, which often pay off faster than new pages on an established site.

Do you need a separate strategy for GEO?

No, GEO fits into the content strategy as an additional criterion. Generative engines rely heavily on well-ranked pages, so the same plan serves both objectives. It is enriched with the questions asked to AI, a clear answer structure, proof and a presence on the third-party platforms that get cited.