This week on Local Search Tuesdays, I’m calling out the nonsense around AI rank tracking. Yeah, AI search is shaking things up—but that doesn’t mean we can slap old-school SEO metrics on it and call it a day. The truth is, traditional rank tracking just doesn’t work with AI systems, no matter what some tools are trying to tell you. I break down why it’s not possible right now, how AI results actually work, and what you should be paying attention to instead.

VIDEO TRANSCRIPT

Welcome back to another episode of Local Search Tuesdays. This week, I’m talking about AI search again, and I’m sharing some bad news… You can’t do traditional rank tracking for AI search systems.

If you’ve been online at all in the last year, you’ve seen how much AI is exploding. The digital marketing world is in a state of extreme flux. There are tons of people talking about how to optimize for showing up in AI search results, and they all claim to know the secret sauce.

And unfortunately, there’s a lot of misinformation out there, along with more than a few shady operators who are just trying to make a quick buck.

Today, I wanted to talk about one of the areas that’s really frustrating, since so many claim to have the solution figured out - when actually, it’s just smoke and mirrors. I’m talking about AI rank trackers.

I’ve been trying to educate people on the inaccuracy of rank tracking for years, especially in local searches. I even did a whole episode back in early 2021 about why rank tracking isn’t a KPI that should be used to judge SEO success.

It was a flawed metric, but marketers and business owners still clung to it for years. Now, with the advent of AI search systems, people are still trying to force old-school thought processes and KPIs on an entirely new paradigm.

While it was flawed, rank tracking still gave you a high level picture of your visibility. It was possible because Google’s algorithm was consistent, so a search for “blue widgets” would always return the same search results page. 

With AI search systems, that concept simply doesn’t apply. AI models don’t return “results” in a traditional sense. The answers you get from AI systems are dynamically generated based on multiple unique factors. Responses vary based on phrasing, the context provided, the past history of that individual’s use of the AI model, and can even vary based on the AI model being used. You don’t get a list of ranked results that are the same results anyone would see if they did the same search.

But - that’s also a big part of the problem. Search behavior is changing because of AI, and people don’t type simple short keywords anymore. When we talked about the flaws in traditional rank tracking, everyone always pointed out that 20% of the search queries entered in Google have never been searched before. With AI models, nearly every query has never been searched before - so how are you supposed to have a system that can track those results?

Then there’s the concept of query fan-out. In a traditional search, a single query returns a list of ranked websites. With query fan-out, the original prompt is broken down into multiple sub-queries. In other words, a single question is broken down into multiple related questions, which the AI system then rewrites dynamically to the final results that are returned.

Then there’s also the fact that click-through rates are getting destroyed by AI systems. Rank tracking was important because we knew that the top-ranked sites would get more clicks. Now, if no one needs to click because the answer is provided in an AI overview or AI mode… you can’t do the “position one gets 23% of all clicks for a query” math.

That doesn’t mean there aren’t some valuable tools that give you some insight into how you might potentially show in AI search results. Personally, I think the best tool out there so far is Waikay - but it’s not a rank tracker in a traditional sense. Instead, it looks at the individual facts that the AI systems know about your brand, so you’re able to find any incorrect information, figure out where it’s coming from, and then correct it at the source. It also allows you to check multiple AI models and even evaluate how those models view you versus your competitors on a single query.

But it’s not marketed as a rank tracking system - and that’s the key. Some really big tools are out there telling you that they can track your visibility in AI systems, so just be aware of the fact that while some work a little better than others, you’re only getting a single snapshot of how you’d appear for a single query - and it’s not likely that your customers are searching for you with that same query.

I’m not saying there won’t be a way to figure this out in the future. But right now, when things are changing on an almost daily basis, and for all the reasons I went through already in this episode, there simply is no possible way to do rank tracking in AI models.

That’s not a bad thing though - rank tracking was flawed from the start. Just remember - the number of people looking to buy what you’re selling hasn’t changed. The only thing that has changed, or will change in the future, is how people are looking for the answers to their questions.

That’s all the time we have left for this week’s episode, so you know what that means. Put your hand on the screen right here: We totally just high-fived ‘cause you learned something awesome. Thanks for watching, and we’ll see you again next week for another episode of Local Search Tuesdays.