For fifteen years, content marketing and SEO lived in different meetings. The SEO owned keywords and technical health. The content team owned the calendar and the brand voice. They traded briefs and occasionally argued about headlines. That arrangement is now a liability, and the thing that killed it has a name: query fan-out.
I have watched it happen in our own data. One article I wrote about query fan-out earned 5,824 impressions inside Google’s AI features in 28 days, up from 1,514 the month before, with no new links and no promotion. It did that because it answers a cluster of sub-questions completely, not because it ranks first for a head term. That is the new job, and it belongs to one team.
What Query Fan-Out Actually Does
When a buyer asks Google’s AI Mode a complex question, the system does not run one search. It breaks the question into many related sub-queries, runs them in the background, and assembles one answer with citations. Google describes AI Mode as doing “a dozen searches for you in the time it takes to do one.” Each of those background searches is an opportunity to be cited, and each one is judged on how completely a page answers that specific sub-question.
The consequence shows up in the citation data. Ahrefs found that in July 2025, 76 percent of pages cited in AI Overviews ranked in the top ten for the query. By March 2026, across 863,000 keywords and four million cited URLs, that figure had dropped to 38 percent, with nearly a third of citations coming from pages ranking beyond position 100. The engine is no longer citing the page that ranks. It is citing the page that answers.
Why This Merges Two Disciplines
Under the old model, the SEO picked a keyword and the writer wrote to it. Under fan-out, the question is which sub-questions a buying committee will ask and whether any page on your site answers each one cleanly enough to be lifted into an answer. That is a content strategy question and a search question at the same time. You cannot brief it across a wall.
Old division of labor | What fan-out requires | Who owns it now |
|---|---|---|
SEO picks the head keyword | Map every sub-question a buyer asks | One team, one map |
Writer optimizes one page for one term | Each page answers one sub-question completely | Writer and SEO on the same brief |
Measure rankings and traffic | Measure AI impressions and citations per page | Shared scorecard |
Publish to fill the calendar | Publish to close a gap in the sub-question map | Strategy decides the calendar |
Pew Research measured what happens to the click when an AI summary appears: users clicked a traditional result in 8 percent of visits, versus 15 percent without one, and clicked a link inside the summary just 1 percent of the time. The reward for ranking is shrinking. The reward for being cited is the only visibility left in a growing share of sessions. An SEO content marketing strategy that still optimizes for the click is optimizing for the smaller prize.
The SEO Content Marketing Strategy I Run Now
This is the process I use for every hub we build, and it is the same process behind the fan-out post that outperformed the rest of our site.
- Start from the buying committee’s questions, not a keyword export. Keyword tools show you what people typed. Fan-out rewards what people mean. I list the ten to fifteen questions a champion, a finance lead, and a technical evaluator each need answered, then map keywords to those questions rather than the reverse.
- Give every sub-question its own complete answer. One pillar owns the head topic. Each spoke owns one sub-question and answers it in the first sentence of its section, in language a model can lift without editing. Thin pages that gesture at five questions get cited for none.
- Write capsule paragraphs. Question in the heading, answer in the first sentence, evidence in the second, implication in the third. This is what AI systems quote, and it is also what a VP skims on a phone.
- Corroborate with data the model cannot get elsewhere. Original numbers from your own programs, client results, and benchmarks are the strongest citation magnet I have found. Restating a vendor’s statistic earns nothing, because the model already has the vendor.
- Link the cluster the way the engine reads it. Pillar to every spoke with the exact sub-question as anchor text, every spoke back to the pillar. This tells both the model and the buyer that you understand the whole problem.
- Measure impressions and citations per page, monthly. If a spoke earns rankings but no citations, its answer is buried. Move it to the top. If it earns citations but no rankings, the topic is real and the page needs authority. Both are fixable once you can see them.
The full version of this framework, including how it fits alongside audit, measurement, and authorship, is in the SEO content marketing strategy section of my B2B content marketing guide.
What to Stop Doing
Stop publishing to a weekly quota. Stop writing one page per keyword variation. Stop letting the SEO and the writer work from different documents. And stop reporting rankings to leadership as if a position still guarantees a visit. Google’s own guidance says no special files or markup are required to appear in AI features, just indexable, text-first, well-structured pages. The fundamentals did not change. What changed is that they now have to be applied to sub-questions instead of head terms, by one team that owns both the search map and the words.
That is how we structure content marketing inside our search practice: the strategist who maps the sub-questions is the same person who owns the brief, and the scorecard tracks citations alongside rankings. It is also why our clients’ content earns AI visibility months before it would have earned a top-three ranking under the old model, and why we run search and content as a single engagement rather than two line items.
Frequently Asked Questions
Is SEO still relevant if AI answers dominate search?
Yes. Indexing, crawlability, structured data, and topical authority decide which pages are eligible to be cited. What changed is the target: pages now have to answer specific sub-questions completely rather than rank for one broad term.
How does query fan-out change keyword research?
Keyword tools still show demand, but the unit of optimization becomes the buyer’s question rather than the typed phrase. Group keywords under the sub-questions they represent, then write one complete answer per question.
Should content marketing and SEO report to the same person?
In practice, yes. The team that maps sub-questions must be the team that writes the answers and measures citations. Splitting them recreates the gap that fan-out punishes.
Want to know which of your pages are being cited and which sub-questions you are missing? Request a free marketing audit and I will map your fan-out gaps against your buyers’ questions.
1 views