Did AI Just Kill Marketing Tracking Attribution?

Stewart Gandolf (Healthcare Success): Welcome again to the Healthcare Success Podcast. Today we are going to have a technical discussion about digital marketing, attribution, how the world has changed. And in fact, we're going to call this episode, Did AI Just Kill Marketing Tracking Attribution? So joining me to discuss this topic are some of my favorite tech people out there, Richard Wong, who is currently an SEO and performance marketing consultant, a leader in this field, who I, by the way, Richard, feel like I know just based upon your relationship with Brandon and the rest of the people from our search agency background. This podcast was actually inspired by one of Richard's articles, AI Broke the Search Funnel.

And then Brandon Schakola, of course, some of our listeners will recognize as our Senior Director of Digital Services for Healthcare Success. Welcome, Richard and Brandon.

Richard Wong (The SEO Consultant.ai): Yeah, thanks Stewart. Pleasure to join you guys.

Stewart Gandolf (Healthcare Success): Great. So today we are going to talk about... Brandon and I talk about, are we going to talk about AI and SEO again? Yes. Yes, we are. It's changed the world for digital marketing up and down across the board.

And we don't talk as much about attribution in some of our other podcasts and some of our other articles. I want to start off, Richard, and this is going to be pretty free-flowing, but your article talked about something that we believe in entirely, that AI search hasn't killed the funnel. It's just moved it into a black box that we can't see.

Before we go to that discussion, let's just talk about the rise of AI, Google AI Overviews, ChatGPT, Perplexity, Claude, all those. What are you seeing today? Because we can mark it month to month with our business. Back in March of last year, 2025, immediately we saw the Google AI Overviews changing website traffic immediately. But what's the latest? What are you seeing out there?

Richard Wong (The SEO Consultant.ai): Yeah, I think people behaviorally have changed their web behavior and browsing behavior over the past two years. I don't know how long it's been since ChatGPT became popular and then the other LLMs, but I think as people have gotten more accustomed to prompting and getting conversational responses back, I think the overall behavior in terms of Google has shifted with it too.

As you mentioned, the funnel may have changed. I'll give you some examples. I was shopping for baseball bats for my son since he plays Little League baseball. And I actually did very little Google search, Google research to find websites to read. I started off with ChatGPT. I told it my situation, how old my son was, what he does, what he needs. Then I used that to narrow down models, types of bats, or even what to look for before I started seeking out websites where I would get it. But I got most of my base information from ChatGPT first. Then that helped determine the types of searches I may do in Google for a more transactional type of search.

So that is an example of how my own behavior has changed. And I would imagine for other people they're doing similar things for other topics and other things in life too.

Stewart Gandolf (Healthcare Success): For sure. I know Brandon and I talk about this a lot offline. It's changing my search, and it's with an increasing rate. And I feel like you've probably seen the Diffusion of Innovations model, which talks about people being either innovators, early majority, late majority, laggards, that kind of thing. I tend to be on the early side of that, but the rest of the world is catching up fast. And I think the more people work with AI, the more they are hooked to it and tend to think of it first.

Brandon, from our clients, any examples you can think of, whether it's website traffic or talking to clients about this change and how it's impacting the world of search?

Brandon Schakola (Healthcare Success): Well, I think we talked a little bit about this in a webinar sometime last year, which is Google Analytics, for instance, just weeds out bot traffic by default. And so a lot of where you would see referral traffic coming in that you could group as a source, and then you can see what activity is happening, just isn't there.

And while they said, I think maybe about a week or so ago, that they're looking to fix that, there's no timeline on when they're rolling it out. There's also no understanding of what exactly they're fixing.

In the meantime, we're stuck looking at log files as a direction. This is what these bots are interested in. This is what the user is interested in. Here's the difference and the overlap. And we're kind of sticking our thumb in the air trying to figure out how much of this actually converted downstream but just got logged into a different channel, like uncategorized or direct.

And looking at shifts and changes in there that may or may not match up. And even in order to do that, you'd have to be a pretty damn good statistician to work out those sorts of variance models. So Richard, I'm probably sure you've seen similar things with your clients.

Richard Wong (The SEO Consultant.ai): Yeah. Direct. Almost every website I've worked on over the past few years, direct has been skyrocketing. And for the most part, the key events, or what they call conversions these days, have been going up with it. Not necessarily at the same rate, but there's been a huge increase in direct to a point where it's like fifty percent, hundred plus, two hundred plus percent year over year for direct on all these metrics.

Meanwhile, some of the more traditional channel metrics like organic search or even paid or email and those other channels, sometimes they may be flat-ish to down a lot.

But then you have to start digging into why that might be. And you may not even necessarily have a problem. I know this was something we discussed before, but maybe SEO is not even really a problem there. I have a client that I'm working on right now where it looks like their SEO traffic is down. But then if you look at some of their other channels, they're running SMS a lot, email, and doing a lot more PMAX, whereas a year ago they weren't doing any of those things.

And what's interesting is the business trajectory has improved a bit since then. So if you start digging into the click paths, I know that we're talking about clicks and other things, but it's still important to see what it may look like. What I'm seeing is with the rise of all those other channels, organic is going down. However, it's the first touch in all those conversion points for all those other channels. So that's an example of the funnel changing just by doing additional marketing.

And so that's how I see AI in this too. It's kind of like a branding and billboard type of channel where if you keep seeing your name mentioned on AI, that naturally shifts how you would search for something in Google. So you wouldn't expect the same baseline or same type of SEO traffic if you're getting those impressions and brand impressions from other places.

Stewart Gandolf (Healthcare Success): So if you're talking about... we talked a moment ago about it being a black box and the funnel is changing now, which I think you just alluded to, Richard. Before AI search, the typical customer left a lot of different touchpoints that you could track depending on your analytics. And now a lot of times you're just seeing the crescendo at the very end, right? Like when they're doing the LLM conversion.

Are marketers that you see out there still in denial? Where are people adapting? What are you guys seeing out there in the world, given this is such a profound change?

Richard Wong (The SEO Consultant.ai): I think at least the whole time I've been working in what they call performance marketing now, and SEO, which I guess is for the past twenty years or so, we've always built our projections and definition of what's good based on traffic. But I don't think that's really a good indicator of success.

The reason for that is, in the past with SEO, people would just build out what they call top-of-funnel content. And that's kind of a broad term. For some people, top of funnel may just be informational content based around what they actually sell or offer. Whereas other people's top-of-funnel definition might be the HubSpot model, just write about anything you want and hope that getting all SEO traffic somehow wins your business.

That approach doesn't work. It doesn't work anymore.

So my point in that is, yeah, all those different touchpoints... I think you need to go back to common sense, really. In the past, before we had GA4 and performance marketing data where people think they can measure every single click, before that marketers actually had to think critically. Does this campaign make sense? Am I speaking to the right person? What's my messaging? All these things.

And so I think we're kind of getting pushed back into that where people working in this field have to also think back and become regular marketers again too, and don't obsess over what a single dashboard tells you.

Because GA4 is always going to tell you the last-click stuff and break channel by channel. But if all your exec team looks at is a GA4 dashboard, that dashboard by default is making these channels compete against each other. One looks good and one looks bad.

That's not going to be a good outcome in terms of how to budget and building a strategy. So I think people really need to think holistically. And if you're building a marketing strategy, let the marketer figure out where to spend that money. But in the end, it's either going to grow your business or it's not.

I think that's what's the most important, not what a dashboard tells you.

Stewart Gandolf (Healthcare Success): So let's talk about that a little bit more. With the seventy percent of traffic being stripped out and just called direct by GA4, that's obviously significant, right? It's like you can't trust the data at all anymore. And all you're seeing is the conversion.

I'd like to expand a little bit on, and this could be both Brandon and Richard, how are people convincing CEOs or the boardroom when they're just... because I'm seeing now, this is obviously really important. We need to take action, but we don't want to panic, right?

So what are you seeing out there? And how would you, if you're walking into a boardroom, explain the upside and downside beyond what we've already discussed?

Richard Wong (The SEO Consultant.ai): Do you want to go first, Brandon, or you want me to?

Brandon Schakola (Healthcare Success): Yeah, sure.

I think the first thing is just understanding the nature of visibility and retrievability. It's not that SEO is going away. We already know that, right? These AIs are using search as part of their mechanisms in various ways.

I think the most recent models, for instance, are using a method where, as it starts to decompose the question and starts the query fan-out, it doesn't just grab that fan-out and go. It's actually grading the quality of the fan-out and causing other ones to get spawned off.

So it's not even just one retrieval moment. It's actually multiple.

And I think one of the reasons, Richard, you mentioned this directly in the article, that this is breaking attribution, is it's causing the funnel to get compressed because they understand intent so well from all of those various fragments.

While as a marketer we can break things out into phases of the funnel, what we need to be building is the connective tissue between it.

So in healthcare marketing, let's say you're working with a substance use or behavioral health clinic. There you need to really be focusing on the continuum of care. What people are looking at in that instance and what they need to know is how long does this last? How long does it need to last? How many different types of therapies am I going to need? What do those look like? Is there a sequence? Is there aftercare when I get out?

That's the kind of stuff you need to steer toward. And it's also the kind of stuff that allows the uniqueness of your brand as an entity and as a voice to come through.

Because one other thing I'm seeing is the more unique facts you have that you can make claims against is exactly what these machines are hungry for. They already know the generic commodity content that's in the field. And so you need ways of filling that in.

What I'm seeing on the board is there definitely is a thirst for how to measure. But I think one of the ways we need to be measuring that is the shifts in intent over time. Because that's something that's also very present in the data.

You can see it in the variants of GPTBot coming through and impacting the overall traffic and impressions landscape. So we do have a way of stitching together part of the story.

But I do think, like Richard, we're having to retrain the C-level in how to think this through.

Stewart Gandolf (Healthcare Success): Okay. Richard, any other comments? I have a comment as well, but I want to let you jump in.

Richard Wong (The SEO Consultant.ai): Yeah. I agree with everything Brandon said.

I wrote about something recently too. I was thinking through how you may measure the value of SEO. Given that clicks are not necessarily a great indicator of value anymore these days, and maybe never were, but especially nowadays with zero-click search, and certainly social media has always been this case. It's always been a black box, but most people never question how much social media can influence behavior, otherwise it wouldn't be a societal issue, right?

But then it's something you could almost never measure.

So I've been brainstorming ways how we may measure the value of different activities. And I came upon the concept of media mix modeling. That's something I discovered during my performance marketing corporate days.

That's a very sophisticated statistical model that traditionally takes paid media spend and revenue or lead generation to determine, hey, if I do this combination of activities, this will lead to this much incremental lift or not, and at what point are you spending too much?

So I started building out my own custom model that actually has SEO traffic and SEO activities in there to see if I were to do this, if I were to build content here, if I were to build links here or whatever, does that result in more incremental lift or not?

So I've been experimenting with different ways to measure that's not click-based.

I think that's the way marketers need to think too, beyond just literal dashboards. I think they literally need to think about what is the overall mix of marketing activities that they're doing here, and is that actually incrementally lifting your business or not?

Stewart Gandolf (Healthcare Success): So I think I'll add a comment and then a question.

As an agency owner, most of our clients are interested in bottom of the funnel anyway. At the end of the day, they kind of feel like... no offense guys, but SEO is the vegetables and the meat is going to be the paid search because they can show that. They can prove that.

But all these things we've all been arguing for a long time have been important, right? The top of the funnel, mid-funnel, all the different attribution. And especially with the LLMs who are looking for authority up and down, they want to see mid-funnel, top-of-funnel.

And again, if people are doing all this research before they buy, just like you said with your baseball bat, you did all your homework offline. That was a critical part of your decision-making.

And so to say that's worth nothing... it's difficult because you just can't tell what's working.

So going back to the models, tell me more about that. What do you see other marketers doing? How are they modeling this? What are the areas of hope? And Brandon, I don't know, with Google I/O, what are the tea leaves from that event showing up?

Brandon Schakola (Healthcare Success): We’re going to have to talk about that. Richard and I need to dish a little bit about that. You know, Richard’s been talking about measurement and how like, the bottom of the funnel is broken. But even now, with Google, basically said, “you know, the days of the ten blue links is officially over and we’re rolling out this generative UI and you’re all gonna love it.”

We don’t even know yet how badly that’s going to break every single rank tracking platform on the planet. So I think that’s something that we’re gonna see coming out of this core update that we’re still in the middle of is just how radically different those results are. And number two, it falls right in line, whereas last year they were definitely struggling to figure out any kind of moetmization strategy. Now they’re really starting to use new ad formats.h recrods marketing, you know, media mix modeling, all of that could blow up because you’re in a new world and in a new normal where you can’t make week over week, month over month, quarter over quarter, year over year comparisons.

Stewart Gandolf (Healthcare Success): Yeah, Richard, any other comments on that?

Richard Wong (The SEO Consultant.ai): Yeah, I think you know, with a shift in behavior that may shift the amount of time it takes for someone to make decisions. So in the past, right? Like would do a model based on, hey, if someone sees or does something maybe 14 days later, they’ll buy or generate a lead with a shift in different types of or then the shift in the funnel, maybe that’s done in seven days now or three days now.

So I think you need to… you can’t just apply old dashboards you’re doing or methodology you did before and expect it to still look right. So I think you really need to look at a lot of different things and figure out, hey how do I adapt? And measurement for what you’re currently doing. And you know one idea might be if you’re doing prompt tracking or AI visibility tracking, you may tie correlations with that and term and do like a time decay on that. And does that correlate with an increase or decrease in your KPIs, for example.

Stewart Gandolf (Healthcare Success): Going back to what Brandon just said, I just, while we're talking here, asked Google, how much money does Google make from pay-per-click each year? And it came up with total, including YouTube and different things, $294 billion.

I think unless you've been in the business, a lot of people forget just how much money Google has at stake here. It's not in making phones. It's not Google Home devices. It's all in this environment. So their entire franchise, like eighty percent of the revenue plus, is probably tied to this business.

So they certainly have a vested interest in not becoming obsolete here. And so they're figuring it out as we are, right? Everything is changing so quickly.

Going back to something you just said, in terms of how do you measure this, Brandon and I are working on some things that we're very excited about that we'll be announcing shortly in the next few weeks about a specific AI tool that can help. But if you don't have AI tools, if you're just starting from scratch, or some of the leading ones, what would you say? How do you get your arms around what's going on out there to present, even if it's sort of scrappy or more advanced?

Richard Wong (The SEO Consultant.ai): For myself, I think it was about a year and a half ago. I wasn't paying for any LLM trackers at the time since it was all so new. But I did some demos on those tools and I figured out, okay, I could try building these out for myself.

The first version of this I built for myself, I used Google Colab with Python scripts. Then I used APIs to all the LLMs, fed all those prompts into Google BigQuery, then I used Google Looker Studio to visualize all those into the same types of reports you get off SaaS tools today.

More recently, I built out another version of that that uses a Streamlit front end. It feeds all that data into a Supabase. Then I'm feeding that Supabase back to that Streamlit front end to do Python visualizations using entity extraction, UMAPing, and other things, and time-series charts to visualize the same thing.

If you really want to do it yourself, it's certainly a lot cheaper than paying five hundred to one thousand to two thousand dollars a month on prompt tracking. And you can automate all that stuff too with cron jobs if you really want to.

That would be a scrappy way of doing it.

For client work, I have them pay for stuff. The reason why we're using Scrunch AI and SEMrush One is because they're in need of certain dashboards that can be sent easily to different execs.

So that's why we use those tools. But you certainly don't need to spend loads and loads of money doing this stuff because, from building these myself, it's really not costing me any money to do this on my own projects.

Stewart Gandolf (Healthcare Success): Brandon, any other comments about some of the tools that you're seeing out there, or do you want to wait until we come back with some reveals in the next few weeks?

Brandon Schakola (Healthcare Success): No. It's a really interesting time because everyone's trying to figure out how do we measure this stuff and how do we... do I have to build it myself? Should I get by using a third-party tool?

I've looked at a bunch of them. Profound. Scrunch, I still haven't checked out yet. There's Waikay from the InLinks folks. There's AirOps.

I tend to find that using multiple, sort of in the way we used to do old-school SEO, right? We would have SEMrush, we would have Majestic, we would have all of these tools to try to actually triangulate what is the source of truth.

I think on my end, prompt tracking gives you a directional sense of what's happening. But the reality is, I think the thinking is somewhat fundamentally flawed if you're not approaching it in the right way because these tools in and of themselves are hyper-personalized. They're hyper-localized. That creates a different dimension for how the agents are going to be wandering around whatever latent space they're in.

I think the power that we have right now is we can actually use these tools to interrogate themselves in new and interesting ways.

In general, they all have something that's worth looking at. Waikay is really great for understanding gaps in the topic space and it's really helpful on that front. I think AirOps is kind of good for general citation visibility over time, and they've got a really good sentiment analysis, which is actually really helpful.

But when we're dealing in healthcare, we've also got these other problems that are multipliers. It's much more difficult to handle in the sense of what's the relationship between location and the main brand entity? What's the interaction between, let's say, maybe an individual clinician? And then you blend in the topic space.

If you tried to map that out, you'd be spending like eighteen thousand dollars on API calls for a twenty-location business. That's a month.

So in some ways, the measurement is almost impractical.

I also think that a lot of the recommendations that come out of these systems aren't really that great to begin with. You really do have to sit there and stew in it.

What's going to happen to a hospital that's just decided to track six thousand prompts? Is anything other than noise going to come out of that?

I'd be curious how your experiments have gone on that front, Richard, but that's at least some of the problem space that I'm seeing.

Richard Wong (The SEO Consultant.ai): Yeah. And the point you mentioned earlier about how things are hyper-personalized, I know this is not perfect science, but how I've handled this is I've been using LLMs and also training data to classify my content and search data and analytics and determine what my ICPs and personas are for these different businesses I work on.

Then I take those core personas, at least the ones that make sense, then I feed those into my prompt tracking. I'll have some that don't have any personas attached to the same prompt tracking. Then I'll apply personas to another version of it. Then you can see, okay, with the personas attached to it and without personas, how does my visibility and performance look there?

Again, like you said, this prompt tracking, that's all we have now. But directionally speaking, if you apply personas to this, does that fundamentally change your visibility over a long trend of time?

If you're looking at the same sites across a broad enough range of data for a long period of time, that's enough to at least give you a sample size of what's going on there. So at least it gives you directionally where you may look.

At that point, you also need to use some discretion too. If you have a persona and it's showing, hey, I'm not visible at all compared to this other company, and I never have been for the past six months across a broad range of stuff, then you go back through and ask, what am I missing on my actual website and third-party information?

And you may see, yeah, I can see what it means because I'm not even talking to that sort of person there. I think it just gives you directions as to where you can go, but you shouldn't take that literally as a  performance dashboard at this point.

Stewart Gandolf (Healthcare Success): The question I have before we wrap up here, number one is just... and I've seen conflicting data and fuzzy data on this, but just experientially I can tell you for sure we've seen it. The people that are coming in inquiring, especially on a B2B basis, are wildly more qualified when we reach them from LLMs than we are from Google.

And it makes sense. Google is saying, “hey, here's a bunch of ideas.” At least currently. That may change next week. But hey, here's a bunch of ideas.

Whereas ChatGPT or Perplexity or whatever is saying, “this is the one you should talk to. This would be my top choice.”

So we find that people are calling us as an agency pre-sold. They say, my gosh, I'm so excited to talk to you. It's just a different vibe entirely.

So how do you take that reality with trying to think through your marketing budget? I know there is no singular answer, but I'm curious what you guys think. Knowing that the LLMs are so important and it's where the puck is going, to steal a Gretzky metaphor, how do we allocate our budgets going forward?

Richard Wong (The SEO Consultant.ai): I think you need to look at all of your channels individually. First of all, what part of your funnel does each channel contribute to the final result? Then you look through, okay, have I tapped out all available resources for this? Because that's the point where you should no longer invest in it.

For example, paid search. Maybe you've maximized your share, or maybe you're at a point where you're cannibalizing everything that you should be doing. That should, in theory, free up budget to do other things with it.

So at that point, if you've determined that AI or SEO is at the top of your funnel, that drives everything downstream and you're underinvested there, I think you should put money there.

Where companies start falling short is if they start dictating their marketing strategy based on what the CFO says. I think that's where they start going wrong. In the end, they're bean counters. They're not marketing strategists. They're there to tell you this financially makes sense or not.

Your job as a marketer is to lead strategy. As a marketer, you need to determine where you're underinvested. If you've invested in all the other stuff and you're underinvested in the top of your funnel, I think you need to invest there. Then you'll see what the results are.

Even if you can't perfectly measure every single click and marginal ad, is your business growing? Are you changing your KPI trajectory or not?

That's the end goal, and I think that's what we really need to keep our sights on and not be so locked in on what a single dashboard tells you.

Stewart Gandolf (Healthcare Success): Brandon, I'd love for you to expand on something we've been talking and writing about a lot lately. The way I look at AI is that it's turned the tables upside down. Everybody's website is out of date. The bad news is your website is completely out of date for the AI era. The good news is that everybody else's is too. That creates opportunity.

Brandon Schakola (Healthcare Success): Yeah, I think it's what we're all figuring out right now. I mean, hilariously, Google came out with their GEO, whatever guidance, that completely contradicts everything that the developers keynote was talking about, right? They were talking about LLMs.txt, the documentation for their things says use it, things like Web MCP and the universal context protocol and the commerce protocol and eighty protocols that we all have to figure out. How do I make my site more available to agentic search? That's kind of an important thing that we're all figuring out and experimenting with.

There's also things that we're experimenting with on the edge, right? So on the CDN layer, or your content delivery network, are there different caching mechanisms? Or could you have something like Cloudflare and an edge worker translate your site rapidly into some sort of cached form of markdown syntax? All of those things are important.

I think the other thing—and there's always lots of controversy about this in our field—you have the yay-sayers and naysayers that schema has an impact. We've known it's had an impact for years. It's a harness, and we can use that and wield that in different ways. While it may not parse it directly, it's still part of the data stream that the LLMs are getting through. But the LLMs are also still dependent on search, so thus they're dependent on knowledge graphs in some way, shape, or form.

So I think on-site, it's really about getting your harness in place so that all that works properly. But then you also have to figure out what are all the other next steps for when, instead of a person coming through and filling out a form on your site, the agent is filling out all that information and passing it through an API directly into the system. And in healthcare, can it do it in an encrypted way that doesn't break the data stream or expose it to additional risk? I wanted to—

Stewart Gandolf (Healthcare Success): So Richard, that makes sense. And then Brandon, you talked about on-site versus off-site. Part of your article, Richard, talked about the importance of Wikipedia, Reddit, YouTube, as being so essential for the LLMs. What are your thoughts there? Expand upon that for our audience so they understand how these things—and obviously Brandon, you can chime in too.

Richard Wong (The SEO Consultant.ai): Yeah, I'll give you an example. This predates Google AI Overviews for the most part, but I previously worked for a company where Reddit—they had a loyal following of customers that were very vocal on Reddit, to a point where they criticized or complimented every single thing the company did.

And it got to a point where there was a vocal minority that kept saying, "I hate this, I hate this, I hate this." Then it started influencing literally almost every single person out there that would type in their brand name because Reddit was present on all the brand searches.

Reddit's not something you can really do direct attribution on and build budget for, but I think it's common sense. If your brand name is a brand name and every single search has Reddit right under your website saying how much they hate this stuff, that's going to color and influence all the other behaviors moving forward.

So I think your strategy shouldn't just be about literally your website. Now with AI Overviews out there—and again, I don't work for that company anymore, nor have I paid attention—but I would imagine with Reddit being so prominent and so much of that negative sentiment being on Reddit, I would imagine all the AI Overview stuff and ChatGPT are either not recommending that company as much anymore compared to competitors, or if people are asking for feedback, it's probably pulling in a lot of that negative feedback, which could certainly influence behavior.

So yeah, that's Reddit, right? Then Wikipedia is obviously another thing in its own right. But I think those are all examples of why you need to look beyond just your website because it's even more important now.

In the past, somebody could put something out there and maybe someone would read it, maybe they wouldn't. Now AI is also reading it and summarizing it and basing its response on what someone else said too.

So you're losing control of the narrative. If you're not addressing those concerns, the vocal minority may become the majority at some point.

Brandon Schakola (Healthcare Success): Yeah, especially at speed. And this comes down to the interplay between technical SEO and the retrievability aspect.

You probably saw the article about Burger King and AI recently, which is pretty hilarious, but I think dead on point. If your website is supposed to be seen as the authority on you and you're not retrievable in that moment when the language model is having a grounding moment, then it's going to whatever other sources it can grab information from as fast as humanly possible so it can come back with a milliseconds response.

If you're not training it both on-site and off-site, you're essentially losing the consensus layer.

So along with Wikipedia, you might find things like Wikidata. In the healthcare space, we've got things like Vitals and Healthgrades and other niche sites that can have an impact that's sometimes even stronger than Reddit.

It's a really interesting beast for us to be looking at right now and observing those results. And that actually is where I think the power in these tools for prompt monitoring comes from, because that's telling you the precise direction to go in and what to prioritize first.

Stewart Gandolf (Healthcare Success): So I have to tell you a sort of fun—not so fun—anecdote on this topic that ties all this together.

I did a podcast recently where we recorded a bunch of podcasts at a private equity law conference. We did twelve podcasts in two days. And the first day, I spent the whole time messing with my mics. I had DJI mics and they weren't working right. So I tried to troubleshoot with AI and it gave me exactly the wrong advice.

I feel like somebody probably went to Reddit and purposefully gave misinformation because it said, "No, that is not the on button. This is the on button." And I was like, that can't possibly be.

Meanwhile, for that particular mic, YouTube had a much better explanation of how to use the DJI mic than DJI had. So it's really crazy that that can happen.

Let's talk about the future as we're wrapping up here. I would love to get your sage comments, both of you, on what we think is most likely to happen over the next 12 to 18 months. Beyond that I think is total speculation. But when it comes to attribution and data, how do you think things are going to change?

Richard Wong (The SEO Consultant.ai): Well, for myself, at least on the SEO side, I think clicks will continue to decline for most businesses, especially established ones.

Of course, if you're starting from nothing, you can certainly grow. But if you have a long-established history of relying primarily on SEO traffic, I think that will continue to decline.

But that's not necessarily a bad thing. Is your business declining with it? Things are changing, right? So I think your expectations for how to measure a channel should shift with it.

Traffic—if your peak traffic was three years ago, it's not coming back. It's never coming back. But your business should certainly grow from there. So I think you need to adapt with it. Your measurement needs to grow with it. Otherwise you'll just think SEO is worthless now and then you'll forget about it. And then, yeah, your business will eventually go down because you underinvested in it.

Stewart Gandolf (Healthcare Success): Well, that's also a scary thought because, as we talk about a lot, the SEO basics are the underpinning of all this. You can't just skip straight to the AI SEO part. You need to get the SEO fundamentals down.

And I could see the knee-jerk reaction being to start cutting budgets, but that's potentially disastrous because then you're not even in the conversation anymore. And you don't know it until it's really difficult to retrieve.

Brandon, any comments from your standpoint?

Brandon Schakola (Healthcare Success): Yeah. I think while bottom-of-the-funnel is going to continue to be a bit of a challenge for the SEO space, what I do think is that eventually we'll get better connectors. At some point, the models are going to have to produce data.

Sort of like Google Search Console—"Hey, here's all of the different query breakdowns we looked at." Not that it's important to get back into trying to reverse engineer the model, because there are different ways of doing that. What I think they're going to need to start doing, if they want this to be functional in e-commerce and larger spaces, is they're going to have to give us that data.

Someone came in. They performed this query or this prompt. Here's how we broke it down. Here's how you made money off of these. Again, it's not that SEO went away. It's just getting more complicated across more surfaces.

Richard Wong (The SEO Consultant.ai): Yeah. And thinking about ChatGPT, if they want to start monetizing through ads or however they monetize, maybe the reporting isn't literally Google Search Console with search queries. Maybe they cluster prompts into topics. Then they cluster them by intent—traditional commercial, navigational, informational. Then you can see attributed conversions or assisted conversions based on each of those buckets.

So maybe if they don't want to give away every individual prompt, which I can understand, they could at least give marketers clustering and other labels to associate with value. Because if your goal is to get more informational visibility because you're in healthcare, you want to see how that's helping versus somebody like me who might be more interested in commercial or navigational queries.

So I think providing more performance data around those, whether it's paid or organic, would be super valuable.

Stewart Gandolf (Healthcare Success): So I'll just add my final two cents on this.

I think the trend away from keywords is really important to understand. And I think I understand this more deeply now. Brandon's been telling me about this for a long time. I understood it at the beginning, but now I understand just how fundamentally the machines understand context. Keywords are kind of irrelevant. They're a bad marker because people aren't asking questions in keywords anymore. They're asking things that have nuance.

And I think you just headed to where I was going to suggest, which is categorization. Because if there are a million different forms of a search, how actionable is that exactly?

So if we start thinking about categories—top funnel pricing, top funnel this, bottom funnel that—some way of categorizing things so the data is actionable.

And then the other part of it is how exciting and how scary it is because you're really driving now at a hundred miles an hour and you can't see the street. You think you're going the right direction, but you don't really know until you hit something or you win the race.

We're making decisions with our own business and with our clients. We're constantly looking at best practices, but you still have to use judgment based on years of experience to make calls that aren't necessarily answerable. Nobody has the answers right now.

But it is fun for our own business because we've been doing this stuff for 20 years, so fortunately we do well with the authority part of it and it seems to be really helping us.

And that's where I would say, if I were looking at this as a listener, I'd be thinking about the basics of SEO and certainly establishing authority because, for the long run, that's got to serve you. I'd certainly rather have authority.

And I'd think about building out the mid, high, and low funnel because at the end of the day all you can really do is follow best practices, listen to what the machines say they want, and navigate accordingly.

Any final comments, guys? That's my last word. Any other comments before we wrap up here?

Richard Wong (The SEO Consultant.ai): No, I think I've offered all of my thoughts.

Stewart Gandolf (Healthcare Success): Very good. Well, thank you, Richard. Thank you, Brandon, for participating. And thank you for listening, everybody.

This is certainly an exciting time to be a marketer, that's for sure.

Thanks, guys.

Richard Wong (The SEO Consultant.ai): Thanks, Stewart.

Brandon Schakola (Healthcare Success): Cheers.

Stewart Gandolf (Healthcare Success): Good job. Thank you.