In this Whiteboard Friday, Garrett discusses AI Overviews and how to optimize your SEO strategy for this new super-charged AI-powered search ranking snippet.

Click on the whiteboard image above to open a high-resolution version!

Hey, everybody. Garrett Sussman here of iPullRank, and today, I'm going to talk to
you all about AI Overviews. The most controversial thing outside the Google leak in the world of SEO. AI Overviews are this supercharged, AI-powered snippet that's appearing at the top of search rankings, and it has been a messy journey.

Okay, so if you're not familiar, basically, there are three types of AI Overviews that appear in search results. You'll see informational, shopping and local.

The informational AI Overviews consist of a range of different content formats, right? So you'll see paragraphs, you'll see breakouts, you'll see bullet points, you'll see citations, you'll see videos, not necessarily in every single type of AI Overview, but basically, Google will pick and choose to provide the best summary of whatever the query is that your searcher's searching for.

Now, shopping AI Overviews are fascinating, because what they're doing is, they're trying to provide products for commercial intent types of queries. So you'll actually see a range of different products, some recommendations from product review sites, some different tags specific to your query.

And then local is very similar to the existing local pack that you'll see, citations, listings, and the map pack, but it's driven by AI search.

And so, how does that work? Well, in essence, what it does is, in theory, Google tells us it just uses the basic search algorithm, right? So, we've got the search algorithm, and it's looking, you take the query, and then it's going through the entire index, finding information, generating an AI output, and then supporting that with specific passages from the content, right? It's taking those with those links, generating a new version, and then producing your AI Overview. All of this happens in the blink of an eye.

Initially, it took a minute, and there were concerns about whether or not people would actually pay attention to it or scroll down, but now it's getting really, really fast.

So, the reality of it is it doesn't seem to actually use the normal, organic search algorithm. It's been tested for the past year and it's in the wild, and it seems to only consider relevance. So the things that we're used to when it comes to AI or to any sort of SERP result is, authority and links and backlinks, and this, that, and the other thing. No, it's just relevance in reality.

And how does that work? Well, Google understands words in mathematics. Every word has a mathematical value, and so what it actually does is it takes your query and it compares it to passages that have a same relevant content mathematical value. Yeah, how does that all work? Well, don't worry, I've got a tool for you that you can try and use to see the relevance to your target queries of your own content. It's called the AI Overviews Visualizer. That's produced by Market Brew. It's a free product, and what you can do is you take your query, you take your content, you put it in, and it gives you percentage points, and you wanna get really high percentages, so your content is very relevant to what you're targeting.

Then, in terms of optimizing for the AI Overview, there was actually a fascinating study done by some folks at Princeton University and University of Delhi that you can use specific techniques to make yourself more likely to appear in these AI Overview super summaries, right? So you can take your content, first off, reverse-engineer, and make sure that the content is relevant mathematically. You use the AI Overviews
Visualizer, check that, that's your starting point, it frequently updates.

Then, you can start to use an authoritative tone of voice, you can use quotes, and you can use statistics. They found that those are statistically more likely to show up in large-language model outputs. Now, we'll get to the caveats in a minute.

It's slightly different for shopping queries. Shopping queries tend to focus only on the Shopping Google Graph. So we're talking information that you provide to your merchant feed, product reviews, subreddits, like Reddit and forums where people are talking about your product, or classic affiliate listicles, where people are saying, "Oh, this is the best type of shoe for running." You are more likely to show up in a commercial search, kind of standard shopping results, focusing on those.

And then, finally, local. Local is typical. Directories, proximity, making sure that you're actually in the area, and then showing up, actually, surprisingly, in places like geographical subreddit. So, I live in Northern Virginia; there's a Northern Virginia subreddit. You're likely to see restaurants recommended in that subreddit show up in your local query.

Now, as I mentioned, there are caveats when it comes to optimizing. It is appearing. It's like the featured snippet where you want to appear in there, but you're not necessarily going to show up in organic down below. There's that risk that if
you mess with your relevance, that might impact your rankings.

So the other caveats, they've been rolling it back. There's no reliability in terms of how Google is playing this out. Initially, they rolled out all over the place; they got all sorts of
really controversial, AI-generated results, they got a lot of bad press, and they rolled it way back, like all the way to 1% of queries. Now, over time, I do anticipate them rolling out more, but right now, it's not there.

The second part is limited appearances. Like I said, it only shows up in 1% of queries at this point. So, is it worth spending your time on? That's part of your SEO strategy and how valuable those keywords are to you. If you are not ranking in the top five, maybe it's worth it for you to try to optimize if you're going against the big brands that you really have no chance of getting up there.

And then you also have the changing features. We've seen things like the follow-up questions, which used to exist. They were removed. We haven't seen them. I expect them to come back because that was a big selling point, because it's all about the context of the journey, taking you from one AI Overview to the next, but they're gone. They might come back, we don't know.

There's also volatility. Just because you show up one day doesn't mean you're going to be there tomorrow. It's AI-generated, so it's every time is going to be unique. In theory, you'd be the one, but it's a little risky in that respect.

And as I mentioned, the unintentional consequences of messing with it you could be messing with other aspects of your website.

So, AI Overviews, it's tricky for SEOs. We always want to rank number one and show up at the top, and there's a ton of value when it comes to brand visibility, but if it ultimately hurts your rankings, and they don't show up, and they take it away, and they give it back, is it worth it? That's for you to decide to play with to test for yourself.

My name's Garrett Sussman. Find me on Twitter or LinkedIn, wherever you prefer your socials. I'm not on Instagram, so let's stick to LinkedIn. Peace.