How to Improve Rankings for AI Query Fan Out

Hi, today I’m joined with Luis Zalifa Herado and today’s topic is how to rank better for AI query fan-out queries. Luis Zalifa Herado (0:15): Hi everybody. Thanks James for having me. Let’s jump on in. James Dooley (0:22): So let’s jump straight into it. How do you rank better for AI query fan-out? Luis Zalifa Herado (0:30): Firstly, you have to define the entity you want to rank for. Let’s define entity. An entity could be an organization, a person, a thing — anything on the internet. And then you have to find the right attributes for that entity, because the game of just ranking for keywords no longer works that well. Now Google ranks entities. Then you have to structure the content in order to answer the question related to that entity. So let’s use an example: James Dooley. Let’s say you want to ask Google a question about James Dooley and Google is going to give you the answer. You can request “James Dooley” and then Google is going to gather all the information they have across the internet and give you the answer. But the answer could be a snippet, could be an AI overview — it depends on how the user triggered the query. It depends on the interface the user uses. It could be voice search, text, AI mode — anything. And then Google is going to get all that information. The one which is easily structured to answer that question — “what is James Dooley’s age” — is going to come up on top probably. So how to rank better for query fan-out in AI is very simple: You trigger a bunch of questions or queries on Google, analyze the SERPs, analyze how the data is structured and which content serves that answer. And all you have to do is come up with a better structure — cover more attributes about the entity and answer more questions about that entity than your competitors. That would be the summary. James Dooley (2:48): So with regards to query fan-out, is there a way of knowing… if I was to search for me — when I started looking into it further — let’s say one of my businesses is called Fatrank, right?

So: “is Fatrank legitimate?” or “is Fatrank a good company?” What I started to realize was the query fan-out terms were like “Fatrank reviews,” “Fatrank testimonials,” and it was all different attributes.

How do you determine what attributes you need to cover? Is it just stuff in your head? Do you prompt AI? Do you look at People Also Ask, related searches, articles? How do you find all the different fan-out queries to fully answer the question? Luis Zalifa Herado (3:54): Okay. First of all, let’s define the entity: Fatrank. And then the first thing I do is analyze the search — you trigger the query “Fatrank” on Google and you analyze as many interfaces as possible. Let’s say first Google, second AI mode, and then you trigger ChatGPT, Perplexity, or any AI interface you want about that entity. You look at the SERP and you look at the information Google extracts. The first things I would do is look at the data on top, and then the second would be the related searches about the entity — in this case Fatrank. Once I have that data, you create a procedure to keep searching and keep expanding about those related terms. With each related term you search, you’re going to get more data and you start kind of a game — chasing the rabbit hole — to solve the puzzle. Each step opens new avenues of information and new research. You gather all the data and when you have a general idea, you create a strategy — a war plan — and you analyze either with a screenshot or with SEO tools. Then you define the entity (Fatrank) and the attributes: “Fatrank reviews,” “Fatrank testimonial,” “is Fatrank legit,” and you get all those queries related to the entity.

Then you think: how can I expand, how can I augment the queries related to that entity — the brand.

That process is how you get the ideas and then implement SEO optimization to keep growing. James Dooley (6:27): Well, I’m going to throw a curveball. Let’s say you’ve got a brand new brand. You generally work with an existing brand that’s got search volume. But specifically for LLMs — you’ve got a new brand now and you want to make certain you’re ranking for all possible query fan-out terms. You can’t search it because there’s nothing there yet. Let’s say they’re called Herado Dooley Semantics — a brand new company we’re setting up, doing semantic SEO.

People might want to know: we’re $20,000 a month, so we’re expensive — why should they use us?

How do you think of the fan-out queries when there’s no Google data for that entity? Luis Zalifa Herado (7:31): Okay. Let’s assume Google has no data about the new brand that is going to launch.

We have defined the brand: “Herado Dooley Semantics.”

Let’s split the brand in semantic terms — three words — and then we analyze on Google if Google has previous data about these three terms. “Semantics” is easy. “Dooley” is a surname but could be attributed to something else. “Herado” is another surname but could be attributed to different meaning. Now you launch a new brand and you have to provide accurate information about these three terms. All the information you provide on the website is going to feed the bot and create a relationship between these three terms and all the content around that brand. When you launch a new brand, you want your customers or potential customers to search in a certain way about your product or services. Then you structure the content related to that product or services. So when a bot crawls your site and gets all that information, it creates relationships between the brand name and all the semantic terms and meanings related to that brand. That way you create a connection. And when the user types on Google…

A semantic search engine can only group three things: queries, users, and documents.

Queries: what the user types in Google

Users: the human searching

Documents: any information on the internet — a document, image, video, podcast, whatever

You create a connection, and the connection is triggered at the moment that Google realizes there is a real human searching for a query and it has no information about this brand. So it triggers a bot, visits that brand site, crawls it, extracts the data, and generates a map of information between the brand, what the user searched, and the document. James Dooley (10:20): I love that terminology — users, documents, queries.

So with regards to that: if you just do the articles and documents on your own website — how important is it to corroborate that on third-party sources?

Because you saying “this is who I am, this is what I do, this is why I’m great” — that’s you saying it. How important is third-party sources repeating that same message for who you are and what you do, to get better rankings on query fan-out in the LLMs? Luis Zalifa Herado (11:00): This is a very important thing. If you don’t have external source validation, you are an isolated entity in the universe — nobody knows about you, nobody talks about you. How important can you be if nobody mentions you, talks about you, creates a link to your site, or has ever heard of you? Third-party validation sources are the glue that puts all the pieces together. If you’re not popular, if nobody knows you, if nobody talks about you, if nobody searches for your brand, if you have no backlinks — how important are you going to be? You’re an isolated thing not connected to anything. So third-party validation like social media, reviews, press releases, citations and so forth are really important — but only as consolidation and corroboration signals when everything is strategically planned and designed from a brand perspective. James Dooley (12:23): Yeah, for sure. I used to be obsessed with backlinks — now I’ve changed the terminology internally from backlinks to corroboration. My guest posts and press releases used to be there just for the link juice. Now I’m semantically trying to expand my topical map externally on third-party sources, and I’d never done that previously. Is there anywhere people can follow you or contact you to learn more — like your newsletter? Luis Zalifa Herado (13:41): Yes. They can find me at seotenico.com. They can search for my name and find everything I share on LinkedIn. And if they’re interested in semantic SEO, they can subscribe to dailysemanticseo.com. Every single day I write an email about semantic SEO. Check it out. James Dooley (14:06): So I hope you like this video with regards to how to rank for query fan-out terms. It’s been here for a long time with query augmentation and query networks. If you want to know more about query augmentation, check out the link in the description — we did another video covering the difference between query fan-out and query augmentation. It’s been a pleasure having you, Luis.