If you've read about "query fan-out" and nodded along without changing anything on your product pages, this one's for you. It's a useful concept that stays abstract until you actually map it against a real PDP so let's do that.
What fan-out actually is, mechanically
When someone asks an LLM-backed search or shopping assistant a question, the system rarely treats it as one lookup. It decomposes the question into several sub-queries, retrieves passages for each, and synthesizes an answer from whatever it found. "Best running shoes for flat feet under $150" isn't one query it's closer to four: shoe recommendations for flat feet, price filtering, comparative ranking language, and possibly a fit/sizing sub-query if the assistant is trying to reduce return risk.
Your PDP is being evaluated against all four sub-queries independently. It doesn't matter how well-written your headline product description is if none of your content actually resolves the price or fit sub-query in a retrievable way.
Map your own PDP against its likely fan-out
Take one real product and one real customer question about it, and manually decompose it the way an assistant would. Not "what does this product do" go a layer deeper. A customer question like "is this good for sensitive skin and does it work with retinol" fans out into: ingredient/formulation sub-query, skin-type suitability sub-query, and a product-combination/compatibility sub-query.
Now check your page section by section. Does your ingredients list actually state formulation details in prose, not just an ingredient table? Is there a sentence anywhere that directly addresses skin-type suitability instead of implying it through lifestyle photography? Is compatibility with other product categories addressed anywhere at all, or is that information sitting in a completely separate blog post the assistant is unlikely to retrieve alongside this specific PDP?
Most PDPs pass the first sub-query and silently fail the other two. That's not a content quality problem it's a coverage problem. The information might exist somewhere on the site. It just isn't attached to the passage that's actually being evaluated.
Practical structure: answer sub-queries as adjacent, self-contained blocks
The fix isn't rewriting your hero description. It's adding short, direct blocks lower on the page, each one built to resolve a single likely sub-query on its own:
A formulation/specification block that states facts plainly, not persuasively
A suitability block ("works well for a skin type / a use case") stated as a direct claim, not inferred from an image
A compatibility or comparison block addressing the "does this work with/instead of z" question directly
Each block should be readable and complete in isolation — assume it might be the only passage retrieved from this page for a given sub-query, because increasingly, it is.
Why this differs from traditional on-page SEO
Traditional on-page optimization mostly cares about topical coverage and keyword presence somewhere on the page. Fan-out coverage cares about passage-level completeness for each distinct sub-intent. A page can rank well under the first model and still fail most of the sub-queries under the second, because ranking doesn't require every section to independently resolve a distinct question retrieval for a specific subquery does.
If you're auditing PDPs for AI visibility, don't start with "does this page cover the topic." Start with "if I decompose the questions a real buyer would ask about this product, does every resulting sub-query have its own resolvable block on this page." That's a different audit, and it usually finds different gaps than a standard SEO content audit would
Top comments (0)