How To Build a Full-Funnel AI Strategy, Part 3: Bottom of the Funnel

Michael 00:00 – 00:23

So what predicts whether your company gets named is how much of the rest of the internet talks about you. Hey, welcome back to the Search Signal. I'm Michael Transon.

Michael 00:23 – 00:48

I'm the founder and CEO of a search marketing agency called Victorious. And here at the Search Signal, our goal is to help marketers understand and also take action on what is happening in the big world of search marketing. So we have been at Victorious running search campaigns for well over 13 years. We work on hundreds of websites, honestly, every year for some of the biggest and also some of the fastest growing brands out there.

Michael 00:48 – 01:13

And we bring what we learn from those experiences plus a lot of our own first party data and research into this podcast to help you with your own work. So this week is part three of our four-part series on building a full funnel AI search strategy from scratch. So last episode we covered the top of the funnel, which is the informational content that's on your actual website.

Michael 01:13 – 01:40

And we also talked through why the reason that you produce that content has effectively changed. Today we're going to pivot. We're going to be talking about the bottom of the funnel. And this is where the question that we're going to answer today moves from whether the AI systems trust your company as a reliable source of information to actually being whether they recommend and use your company's name when those same users start to ask for recommendations.

Michael 01:40 – 02:09

So before we get into today's episode, let's take a little bit to talk about what is happening this week in the world of search marketing news. Okay, let's kick things off with Google's AI mode and the day and a half where it actually stopped showing any of its sources. So if you did not see this, Google put a new model into AI mode at the start of last week, I believe. And within about a day, people all over the internet were posting screenshots of AI mode answers that just didn't have any links or sources in them at all.

Michael 02:09 – 02:32

There was no citations, no sources, just like the answer. It was mostly from what we can see showing up on informational questions. And it seemed to also be limited to the paid tier of users who had access to the newest model. So Google's head of search product said, I believe it was on X that you know this wasn't obviously working as intended and that a fix was upcoming.

Michael 02:32 – 03:04

And I think by the next morning, those links and citations were all back. Now, you know, this was a bug. Google says so, and I have no reason to doubt it, but. for about a day and a half, arguably the biggest AI search product in the world was answering questions without showing any of its sources. And what that tells me is you know, whether an AI system shows a link to your website or a citation could be the property or a property of the model that happens to be running or you're running on that specific day.

Michael 03:04 – 03:30

And, you know, models get swapped out all the time. We might find that there are future models that are behaving differently and how they choose or don't choose to show their citations. And point being, you know, trying to improve your citation rate, which just as a reminder, it's how often your site is being used in an actual answer. Using that as a means to get like AI referral traffic or or other traffic may not be sustainable.

Michael 03:30 – 03:55

Instead, we talked about this actually in a prior episode. Improving your citation rate as a brand should be seen as a way to influence the AI's system's understanding of the brand. And the authority that you have in your industry. So as marketers, I think that we should be maintaining some interest in how these AI systems are and how these models, these new models are surfacing their sources as newer models ultimately get released.

Michael 03:55 – 04:24

And then the second update this week to talk about is also from Google, and it is about the ads that are showing up inside of AI mode. So if you have not seen Google confirmed, it's testing regular search ads, which is the kind that you run with exact match and also phrase match keywords inside of AI mode answers. So up until now, if you wanted an ad in AI mode, you had to be running one of Google's automated campaign types.

Michael 04:24 – 04:50

This is something like performance max, for example, or broad match keywords with automatic bidding. So Google's ads liaison said that this test is " limited to cases where there is," again, I'm quoting here, like "explicit and direct user intent." And they're calling it a small experiment. But here's why I thought this was worth talking about.

Michael 04:50 – 05:24

Explicit and direct user intent is Google's way of describing a person who is just simply ready to buy. That is the Bottom of the funnel. And that's what today's whole episode is going to be about. And what Google is building is a paid way to get your company's name into an AI answer at exactly that moment. So as you're thinking about how much to invest into earning your way into these AI answers, which is again something we're going to talk about today, I would keep in mind that the alternative, if you don't do it,

Michael 05:24 – 05:59

is going to be buying your way in. And I would expect that price to go up and up as more advertisers and brands find and ultimately utilize it. And then the third update this week to talk about is also about Google and the EU and who shows up first in search results over there. So Google told Reuters this week that the changes that Europe's new digital markets law forced on its search results are the biggest drop in quality in the 29-year history of Google search.

Michael 05:59 – 06:24

So here's what changed. So, for example, when somebody in Europe is searching for a hotel or a flight or like a restaurant, the first thing at the top of those pages is now going to be a booking site like Expedia or Tripadvisor or Booking.com. Then they'll have at least two more of those types of sites. And then you'll see the hotel or the restaurant or the flight itself.

Michael 06:24 – 06:49

Like the actual end business. Those are gonna show up below that. And Google took the live pricing also out of those results. So Google's argument is that this ruling favors the middlemen over the actual businesses themselves. Now, two things I would say. One, this only applies in Europe. And then two, you know, Google has a pretty obvious reason to describe regulation as being bad for users.

Michael 06:49 – 07:14

So I would take the The "worst drop in quality ever" as Google's opinion, not necessarily as a fact. But the reason why I picked this one is that the layout Europe is forcing on Google in this is already the way that AI answers work everywhere in all their categories. The thing that that gets ranked is going to be in AI answers things like a directory or a review site or list of the best companies, right?

Michael 07:14 – 07:49

Your company is a name inside of those sources. If you are in the EU or you operate there, these middlemen aggregator websites just became a lot more important in your user acquisition strategy and definitely something that I would take and I would strongly recommend taking some time to understand a lot further. So I picked each of these three updates. There was a lot of news this week, but I picked these three on purpose because all three of them are the platform deciding what shows up around the answer and you know whether the links or citations appear at all, who

Michael 07:49 – 08:16

gets to pay to be named, or like which middleman ultimately sits on top. But the thing that you control in all of these is whether your website and your brand name is being mentioned on other companies' websites on pages that talk about what you do, because that is what earns citations, and that was what earns brand mentions as we get into the commercial, down-funnel part of building an AI strategy.

Michael 08:16 – 08:43

So with that being said, let's get into today's episode and we're going to start with a quick catch up if you missed the last episode. Last time we talked about the top of the funnel, which is all of the informational content that is on your website, like the blog posts and the guides that your buyers are going to read when they're still trying to figure out and understand what their problem is. And the case that I made that in that episode was that this content now has two different readers.

Michael 08:43 – 09:10

Okay? One of them is has been and still will be a person, but the other one is going to be the AI system itself, which is going to read your content when it is updating its model and learning about the world. And then also again when it goes out and it searches the web to help build and source an answer. And because fewer and fewer people are clicking through to a website when they get an answer, the job of that content has materially changed.

Michael 09:10 – 09:39

Its job now is its primary job is to teach Google and these AI systems that your company knows this subject well. So that when your buyer and their question turns commercial, when they're asking like who they should hire or where they should buy from, your company is one of the names the system considers because that informational content that you publish on your website has influenced its opinion of you as a, I would say like a premier authority in the space.

Michael 09:39 – 10:02

So if you missed the last episode, I would strongly recommend go back and listen because today is going to be building on top of that. And there's also something that I want to build on last episode, today, which I had said in the last episode that building authority, top of the funnel makes your company a candidate. But what it does not do is make you the actual pick.

Michael 10:02 – 10:35

Those are two different things, and they turn out from our research to be a lot further apart than a lot of people assume. So we did a whole episode on our Q2 research for AI search earlier this summer. And the short version of that is that AI can very accurately describe most brands that you name. And they can also be pretty accurate at describing what they do and who they serve, but it still leaves the vast majority out of them in an answer when someone asks who to work with.

Michael 10:35 – 11:02

So point being, being known and being recommended are almost unrelated. And the gap between those two is what we are going to talk about in today's episode. And as in prior episodes about AI search, we're going talk a little bit about citation rate. And we're also going to talk a little bit about mention rate. So I just want to quickly remind us of what they are. A citation is when an AI system uses a web page as a source for its answer.

Michael 11:02 – 11:27

And you will usually see it as a link off to the side, or it might be like a small footnote, or in some of them, also it it's a hyperlinked text within the actual answer. Okay, that's a citation. A brand mention is when the system says a company's name inside of the answer itself. And at the top of the funnel, we are mostly earning citations.

Michael 11:27 – 11:50

And that is totally fine because our thesis, again, is a correlation between citation growth and perceived brand authority. At the bottom of the funnel, a citation really does not do much for our website unless your company's name is in the text of the answer. So from here on out, when I talk about mentioning, I'm mostly going to be talking about mentioning your name in the answer.

Michael 11:50 – 12:12

And that is what we are chasing at the bottom of the funnel in AI search. So if the last episode we talked about was about getting these systems to trust your company in your industry on your subject. Today is going to be about that, really one thing that the trust is supposed to be able to buy you. Which is your company's name showing up when someone asks who to hire or what to buy.

Michael 12:12 – 12:41

And the first thing to understand about this is where this all gets decided, because forming these answers is not just about what is on your website. Okay, here's the thing about the bottom of the funnel in AI search. It really only comes down to one single material outcome. When somebody is asking ChatGPT or Google to recommend a company in your category, the only thing that counts is whether your name comes back in that list.

Michael 12:41 – 13:04

It does not matter whether your website got cited. It doesn't matter how you rank or how much traffic you are getting. I want to be as unambiguous as I can here because there's a lot of noise in this space. And this is the entire objective at this stage. Get your company's name into the answer. So the natural question is well, what influences that, right?

Michael 13:04 – 13:33

What influences it? And the answer based off of what we can measure, in my opinion, is probably not what most people would guess. So what predicts whether your company gets named is how much of the rest of the internet talks about you. So when we ran our Q2 research around we did like 150 different brands and we were looking into this, trying to find an answer.

Michael 13:33 – 14:06

The two strongest signals of getting named and AI answers were first the number of other websites linking to you. And then second, the number of other websites mentioning you with or without a link. And those two were close enough to each other that we essentially called it a tie. The brand's own historical search rankings or how much organic traffic, those were sitting well down the probability levels of what was influencing it.

Michael 14:06 – 14:32

And then also Knowledge Graph entries, which is the panel that Google shows confirming like who a business is, a lot of people assume this is a thing that matters for branded mentions. There was no reliable relationship that we could connect. Also, as it relates to being mentioned across the web, brands with fewer than 2,000 mentions across the web got named and AI answers about 3% of the time.

Michael 14:32 – 15:04

But brands with more than 30,000 mentions got named around two-thirds of the time. So basically, what we're seeing is the AI systems are consistently naming companies that it keeps running into on other people's websites. And I mentioned this in passing, but there are also two variables in these mentions that I would say I was surprised have a very low material effect.

Michael 15:04 – 15:29

The first is the link itself. So a page that names your company in a paragraph with no backlink at all still counts towards the volume. You can literally have an unlinked mention on somebody else's website, and it's going to influence the system. Which is crazy. The second one is the actual authority of that website that is doing the mentioning.

Michael 15:29 – 15:58

So in traditional SEO, we've spent you know 20 plus years going after, you know, links from big high authority websites. And in our data, the quality of the referring or mentioning site added like a modest amount of impact, but not much. So just being honest here, and you know, we don't currently see an association between a high quality website and improved mention rate.

Michael 15:58 – 16:25

And I wanna also be clear that might change. And I think it probably will change, but that is the objective honest current state of what's happening as this episode is recording. Now, what does matter, what is important is whether the mentioning website's topic is associated with your category or your industry.

Michael 16:25 – 16:57

And there's a term for this. We've talked about it before. It's called co-occurrence. And all it means is your company's name, your brand name showing up on a page about the type of work that you do. And oftentimes next to other companies that also do it. The reason this matters is that the AI system keeps finding your brand name and your industry details together in content that somebody else wrote, and that repetition is what is going to build the association.

Michael 16:57 – 17:20

So we saw brands in our research too that validated this that had links from very authoritative websites, yet still rarely got named because those sites had nothing to do with the questions in the industry that were being asked. So we know the outcome that we are effectively after here.

Michael 17:20 – 17:44

And we also know, based off of our research, it's decided by how many other websites say your name in the right context. So the obvious next question becomes which websites? And the good news is that if you did the work from the first part of this four-part series, you've already got that list.

Michael 17:44 – 18:14

So let's talk about how we can pull it. Okay, our list of websites that we want to get contextual brand mentions starts with the prompts that you are already tracking. So for this, you want in your system the commercial and the transactional prompts that you are tracking only. Okay. commercial prompts are things like short list questions. These are would be like your prompts that are tracking best competitor for this or alternatives to a certain competitor, top firms to use, certain use cases.

Michael 18:14 – 18:40

Okay. And then transactional prompts are going to be, I would say, like ready-to-act questions. These would be like pricing and quotes and how do I get started? Leave the informational prompts out of this completely. Those are the top of the funnel set. They almost, like we've talked about a lot in this podcast, they almost never actually name a company. And so if you mix them in, you're going to be building an outreach list from websites that we're probably never going to be used to recommend anybody.

Michael 18:40 – 19:06

If you do not have a prompt set yet, the method from part one of this series still applies. And I would re-listen to that. So in that we talk about we essentially take existing search keywords with real search volume. We map them to entities your company makes money from, and then we write out how a buyer would ask each one of those as a question. And I had said that I start typically with 10 to 30 commercial prompts per entity.

Michael 19:06 – 19:32

Once you have those, here's what I recommend that you do. First, for each prompt, pull the last few times that it ran in whatever software that you are using to track your prompts. Then for each of those runs, pull and organize the full citation list of sources that were used for the answer. Okay. These are the citations, the websites that the system pulled from to cobble the answer together.

Michael 19:32 – 20:07

Third, and you can do this this third pair. You can do it manually, or you can ask an AI tool to check. go to each one of those sites and see which of those citations do. And also do not currently mention your brand name. Every one of those sources that does not mention your company is going to be one of our outreach targets. And I would say this beats any cold outreach prospecting that you can come up with because we already know for a fact that the website in this list is already being used to shape the answers that your buyers are getting.

Michael 20:07 – 20:36

The sources that you pull, I would say, are also going to look different from the ones that you would get if you were pulling the top of the funnel prompt citations. You should expect a lot of directories, a lot of review platforms. You'll see ranking roundups. You'll probably see a lot of comparison articles. Those pages are there for a reason because their whole purpose is to help list companies in a category.

Michael 20:36 – 21:02

There's also a lot of research that backs up why this is the case. So if you look at all AI answers combined, lists are actually the most cited type of page on the internet. And you'll see that stat used to argue that your company should be publishing lists. But when you split it by what the person was asking, it tends to flip. So on informational questions standard articles get cited about twice as often as lists.

Michael 21:02 – 21:25

And then on commercial questions, the lists win by a really big margin, like about eight, I think eight in ten citations are lists and they're usually going to be most often somebody else's third party rankings that names brands. So the type of page that gets your company named at the bottom of the funnel is most of the time going to be a list that you don't own.

Michael 21:25 – 21:48

And I do want to say some of these cited websites, and we'll talk about this later, like review platforms, can be actioned without having to do any outreach. You can just like simply go up and sign up for an account. We're gonna talk more about that in a second, about the types of websites that might be in your citation list and how to go about each of them. I would say also expect these sources to move.

Michael 21:48 – 22:15

The same prompt from this week to next week, it can pull a completely different set of sources. And oftentimes it's for no reason that we can really see. Maybe the system like decided that the source wasn't good enough, or maybe it was testing something, maybe the model changed. Nobody outside of the platform, sometimes no one even in the platform itself knows why. In some of these systems, you know, more than half of the websites being cited can swap out from one week to the next.

Michael 22:15 – 22:45

So the reason why I bring this up is because you can't just like put a check mark next to this work and see it as like a one-time task. We really need to stay up front and ahead of the changes to a lot of the sourced citations and constantly be outreaching. So this is similar to SEO in that this is a never-ending race against the competition. okay, so you know, taking a step back at this point, we've got a list of websites that are actively being used to answer your buyer's questions.

Michael 22:45 – 23:10

And we also have a subset of that list that doesn't say your company's name. So the question now is, how do you get onto those? And the way that I would probably start by splitting this up is what costs money or not. But let's talk about how we would actually work this list because this is the part that I see a lot of people just skipping past. They see a big list of websites and they talk about, you know, how do I get a bunch of mentions everywhere?

Michael 23:10 – 23:35

And then the answer that I would say to that is it starts with a much narrower job. So these websites in this list are the ones that we're already answering buyers' questions and your company's not on them. And the first thing I would do is sort this list by what kind of page each source is. Because, like I mentioned, the approach for getting on these sites is going to be different for each kind.

Michael 23:35 – 24:03

When you do this, you're gonna find that almost all these sites are gonna fall into what I would call four major buckets. Let's talk through each of these. Bucket one is what I would call directory and review platforms. This is like G2, Captera, Clutch, any industry-specific ones that are in your space. These are the easiest to get onto because there's going to be no publishing process.

Michael 24:03 – 24:29

And also no one, no one has to agree to include you. You can claim the listing, you can fill out your profile all the way, you can pick your category. And you know, some of them have specialization tags and you get your customers to leave reviews there. All of that is very much completely under your control. most of them also do sell a paid tier, and those work in different ways. Some sell positions in the list outright.

Michael 24:29 – 24:53

Clutch, for example, orders its directory pages by sponsorship level first, and then everybody else comes after. There are other ones that compute your position from your reviews and how many there are, how recent, how happy they were. And what they sell you is the program that kind of like helps you collect more reviews. And some of them sell like the data for who's looking at you. G2 has said directly that it doesn't sell placement.

Michael 24:53 – 25:18

Nobody has really measured how much of that paid placement spend reaches the AI answers. What we know is that the AI treats the resulting page, though, as a source, and that a complete and well-reviewed profile is a much better page for it to read than one that is much more thin. That's bucket one. Bucket two is what I would call the ranking roundups and the comparison articles, right?

Michael 25:18 – 25:41

This is the like "10 best companies for this" post on like a trade pub. Or like an alternative to a competitor article on someone else's blog. These are the ones where you're gonna actually have to reach a person. We would need to find out who wrote the piece or who maintains the piece, which is, you know, could be the byline. or you might find that information on an about page.

Michael 25:41 – 26:07

And then we need to make a specific ask. You'd like to be considered for inclusion. Here's why your company belongs on a list about this topic. Here's what makes it easy for them. Like the, you know, give them the facts about what you do, proof that you've done it, data points they can use. A lot of these lists get refreshed about once a year. So you could also, if you want, ask when the next update is and then just try to get in front of that.

Michael 26:07 – 26:30

And if it's a commercial site, you can also just straight up plainly ask whether they offer paid inclusion, because many of them do and they're, you know, don't always volunteer it up front. I'll get to what you can do in terms of offers in a little bit. That's bucket two. Bucket three, I would call forums and community sites, you know, Reddit and the like. You cannot typically buy your way onto those.

Michael 26:30 – 26:59

And I wouldn't recommend you trying to. What does work on these types of sites and in this bucket is when somebody from your company is just being a real participant where your buyers can, where your buyers are asking questions and doing so over, you know, a long period of time. And whatever you do on those sites has to be above board. These are real people from your company saying who they are, where they work, and then also adding something actually useful to the conversation.

Michael 26:59 – 27:30

Fake accounts, no. Planted recommendations, no. Those get caught. And when they do, they are going to damage the brand that you are working very, very hard to build. So that's bucket three. The fourth and final bucket is what I would call news and editorial coverage. And these are typically articles that have already run Usually you cannot get added to a piece that's already published in a news or editorial publication, but that publication is now potentially a target for the next piece that it's going to write about your category or industry.

Michael 27:30 – 27:53

I can talk about, and I will be talking about how you can do that in just a minute. Now, how to make the quote unquote ask for mention, right? Because this is where outreach fails a lot. Whoever you're writing to, there's gonna be, you know, I I think there are about five things that you can offer them. And, you know, we can talk about offering a couple of these things or just one, but let's talk about what they are.

Michael 27:53 – 28:16

The first one in this list is not surprising. It's going to be money. Oftentimes you can pay to be included. You would be surprised how many websites operate in this way. Some, once you get to this conversation, could quote you on a fixed cost for the placement. Some of them might quote you on a recurring rate to keep you on the page.

Michael 28:16 – 28:45

It can range from 10 or 20 bucks for a small site up to hundreds or even thousands of dollars per placement for a bigger one. Whether that's worth it is a decision that you're gonna have to make site by site, The case for it, though, is that you get the mention and then the mention exists for as long as that page does. The case against it is that that site or like that specific page that might stop being cited by the AI systems at some point.

Michael 28:45 – 29:09

But I would say even then, it's still a mention. And it's a mention of your brand on the internet on a page that ideally is about your industry and your category. And as we talked about earlier, that counts towards the overall volume that gets you named. So I think that paid mention does have value beyond just a specific prompt that might be using it as a citation today.

Michael 29:09 – 29:38

That's the first option. The second option is an exchange. And this is very simple. You can link back to their site or you can mention them in a piece of content on your site and they include you on theirs, right? Straightforward. Third option, you can try to make their page better. And what I mean by that is give them something that only your company can provide. If you have unique data or research, or you've got a unique insight from your company's experience or an example from a customer.

Michael 29:38 – 30:06

You're offering to improve what they have already published. And that is, you know, in my opinion, probably easier to say yes to than just asking for a favor, which we'll get to in a second. But the fourth option would be like you can also explore a broader relationship between the two businesses. Maybe a marketing relationship, a referral agreement, some sort of economic relationship between your company and theirs and mentions and being included on their site could be like one piece of it.

Michael 30:06 – 30:34

And the fifth last one is just simply goodwill, which means like we make just a clear, strong, specific argument for why our company belongs on that page. We talk about what we do, why a reader of that list would be worse off not knowing about you. We can make that argument, keep it short, make it about their readers rather than about you, whatever offer you ultimately end up leading with, the mechanics are pretty much the same.

Michael 30:34 – 30:58

We need to keep it clean, short, reference the page, it's clear, and then we follow up, you know, follow up a week later, follow up two weeks later. And then most importantly, we need to move on. And move on is the right way to think about it because to be honest, you're gonna have to send a lot of these. Response rates on outreach like this can be very low, as low as like a half a percent to two percent.

Michael 30:58 – 31:23

So the goal here is going to be volume. You want as many mentions as you can get, relevant mentions as you can get, which means you're reaching out to far more sites than will probably ever say yes. That's also why we need to be able to track it very well. There are tools that are built for tracking this type of email outreach, or you can also do it manually inside your own email client. You can get a spreadsheet put together, either way.

Michael 31:23 – 31:51

We need to be able to write down our sources. We need to know what prompts it's showing up in. We need to know what kind of page it is. We need to be able to figure out who we contacted, what we offered, when we offered it, and then what happened. We're going to run this source poll at least once a month. And the record here is how we know whether that we've got a mention that ended up landing and whether the AI answer ended up moving to a new citation or not.

Michael 31:51 – 32:14

Now there needs to be a budget set aside for this, both for the paid placements and also for the actual work that's being done. right? Ideally, you are working with an agency that does this with you and runs the outreach on your behalf. And there are a couple models for how that works. Some agencies will include the outreach and the placements in their retainer. And it's their job to figure out the economics.

Michael 32:14 – 32:39

Others go out and they acquire the mentions. You essentially pay the cost of the placement. As a pass-through, and then the agency charges a fee on top of that for getting it done, or you do it yourself. That does work too. It just takes a lot of time. It takes a lot of organization to set it up so that it runs every single month and you don't lose steam or momentum because you just did it one time and then you lost motivation.

Michael 32:39 – 33:04

So on priority within the list, I'd also try to prioritize working the sources that show up across several of your prompts first. And then work to the ones I would say where a competitor is named, and then you're not. And then the ones that just appear more than one in one AI system if you're tracking more than one platform. And then also just on expectation setting for all of these.

Michael 33:04 – 33:39

The directories are going to be your quickest wins. The roundups in those types of sites in that bucket is going take longer. And then when a placement does land, when we do get a mention, you might see it show up in the AI answers. Within a few weeks, a couple of days, could be a couple of months. That really just varies a lot by system and how often those prompts are being run. Okay, so taking a step back, let's say you have gone through that list, you've gotten onto what you can get onto, and you're rerunning that pull of citations

Michael 33:39 – 34:06

every single month, and you're working through the new sources as they are appearing. That is the most targeted version of this work. And I would always do that first. But let's go back to the number we talked about earlier. When we had said that companies with a couple thousand mentions across the web got named about 3% of the time. And then companies with 30,000 mentions got named about two-thirds of the time. The cited list is the shortest path because you already know that those specific sources are actively being used by AI systems.

Michael 34:06 – 34:32

But your overall mention count across the whole web is what drives whether you get named. And you know, no list of like 20 or 50 websites is going to get you from 2,000 mentions to 30,000 mentions. So once that list is up and running and you're doing that work, let's talk about the ways to raise your general mention rate across the rest of the internet.

Michael 34:32 – 34:57

Okay, let me start with something that is true of every route that I'm about to describe and talk through with you. So all of them are going to cost time. What differs is whether or not they also cost money. Okay. I'd split it and I am splitting it that way because that's how work gets resourced inside of a company. You either have hours or you have budget or both, and knowing which one you are spending is going to be a big part of the decision.

Michael 34:57 – 35:24

And if you don't have any hours, then somebody else's hours or what you're ultimately going to be buying. The other thing that's true across the board in these opportunities I'm going to walk through is that the test from what we talked about earlier around co-occurrence applies to every route. A mention is worth pursuing when it lands on a page specifically about your category or your industry.

Michael 35:24 – 35:53

And ideally next to companies in your industry. Which means that in a really easy mention on a random page or website that has nothing to do with what you do is going to produce very little outcome, no matter how easy it was to get. Okay. So here are the routes. I'll say these are roughly in order of what I would call durable value. The first route is publishing original data or research on your website that other people can cite.

Michael 35:53 – 36:20

If your company has data that nobody else has, and you publish it, every article that cites it is going to put your company's name on a page about your category industry. And it is going to keep doing that for years after you publish it. And you won't have to do anything else. So this is the only route that is in this list that I would say quote unquote keeps working on its own and you don't have to continue to work on it.

Michael 36:20 – 36:48

That's the first option. The second option is something that we've seen commonly in history, which is answering inside of SEO itself, is answering journalist requests. There's a like a reporter that is already writing a piece about your category or your industry and just needs a quote from somebody who knows it. There are actual literal platforms to help cultivate this. There's Qwoted, there is Featured, there's a Help a Reporter Out that shut down a couple of years ago, but it also did come back this year as a newsletter under Featured.

Michael 36:48 – 37:16

So this is like, I would say this is lower effort per placement than something like pitching a whole story, which is like PR. And the page is about your topic. And oftentimes we'll already have some companies, other companies or industry quoted alongside you, which is exactly the kind of page that we are looking for. This is also how you get into publications from the fourth bucket we mentioned earlier.

Michael 37:16 – 37:50

The editorial and news publications. These are publications, and ideally whose articles are already being cited. That's the second option. The third route in our list of five is going to be becoming a contributor at a publication that is in your industry or covers your category. Why? Publishing and being an author on those websites gets your name on every article that you publish there and also gives you a permanent author profile on their site that points back to your company, which means theoretically two mentions per piece.

Michael 37:50 – 38:18

And that profile can build up across everything that you write on that specific website. The fourth is awards. These would be best of lists and any roundups that accept actual nominations. These will land on the same kind of page as the like general roundups that we were talking about in our citation audit.

Michael 38:18 – 38:42

But the difference being, you might not have found these ones yet because they aren't being actually cited for your specific prompts today. You're gonna need to invest time and resources into things like award submissions. As these websites and companies typically only publish content about shortlisted companies or actual winners, not just people who submit for the awards. And then the fifth is exchanges.

Michael 38:42 – 39:06

This would be like a link or a mention that runs in both directions with a related company or a partner or a customer. Think about things like joint case studies or partner pages, your name next to a related company on a page about the work that you both do. Those are just five. There are honestly a lot of others.

Michael 39:06 – 39:30

You can pitch journalists directly when you've got something to say. You can go on podcasts in your space. You can speak at events. Those all work to the same degree. But the five that I mentioned are probably the ones that I would start with. Those are the five, time intensive options. Let's now talk about the paid routes.

Michael 39:30 – 39:59

People get reluctant to talk about these. So let me just say it plainly. You can pay for placement. You can sponsor content. And as I said, most directories and also most review platforms do sell a tier that changes how prominently you do appear. But that also raises a question I get asked a lot. Which is whether a sponsored mention carries the same weight as a non-sponsored or an earned one.

Michael 39:59 – 40:26

My position is that I think it does and it should. And here's my reasoning: the only thing that has ever marked a placement as sponsored was a tag on the link. And we just established that a mention with no link at all counts about the same as one with a link. So most of the time, when we're talking about unlinked mentions, there's no marker left for a system to read.

Michael 40:26 – 40:51

And I would imagine that one of these AI systems, when they read a page, it's probably spending its processing power interpreting what the page says, not working out the commercial arrangement behind why the page actually says it. Now, nobody has as far as I'm aware, measured this directly yet. So to be clear, this is just my read of how I think these systems work, and I would be fascinated to be proven wrong on it.

Michael 40:51 – 41:23

But this is a this is my opinion on where I think it's going and why sponsored content is just as valuable as earned content. And then there's also a research route for finding more targets beyond you know what that original prompt audit we conducted turns up. So if you're looking for more opportunities, there's a lot of interesting things you can do. For example, you can take a competitor that is beating you in AI answers and you can pull the list of websites that are linking to them out of a tool like Ahrefs or SEMrush.

Michael 41:23 – 41:53

Both tools let you do that for any website or domain. You can then go through and you can look through that list looking for sites that pass our test we talked about earlier about which are sites about your category that also list companies like yours. You're gonna go through probably a lot of irrelevant sites in that list to find them, but they are likely going to be in there. Now the blind spot to this approach is that these tools really only show you websites that link to your competitors.

Michael 41:53 – 42:19

A page that just simply mentions your competitor without linking to them is not gonna be in that export. But for those, there are tools now. If you did not see, Ahrefs launched one in June. It's called Firehose, and that sends you every newly published page that matches a keyword that you set, which can be like your brand name or your competitor's.

Michael 42:19 – 42:46

And it's whether there is a link on that page or not. And it also has got a free tier if you want to check it out. Ahref's Content Explorer also does the historical version of that and it has a filter specifically for pages that mention a brand without linking to it. And then SEMrush also has a brand monitoring app that flags mentions with or without a link. So the unlinked point being is the unlinked mentions are findable now, both for you and for your competitors.

Michael 42:46 – 43:09

But taking a step back, the last thing that I would say about all of these available routes to us is that pretty much every one of them runs through an actual person. Like even when the content on another website was written by AI, a person owns that site. And they have a commercial reason for running that website. A lot of these people expect to be paid.

Michael 43:09 – 43:43

And a company with a strong and I would say, a reasonable budget put towards this work will likely not run out of websites that are willing to take its money in exchange for a mention. So every mention that you earn is eventually going to send somebody to your website. And that raises a question that I also haven't answered yet, which is what your bottom of funnel pages on your actual website are actually for if the brand mention and the influence is actually being won somewhere else

Michael 43:43 – 44:11

in an AI system. So the answer is they have two jobs validating and converting. Validating and converting. Think about who is arriving. This is a person who has already done the comparison inside of whatever AI answer tool that they used. They've got a short list and your company is going to be on it. And they're landing on that specific page that has gotten recommended, usually not your home page.

Michael 44:11 – 44:37

It's fewer visitors also than organic search sends you, but they are much later on in their process. We talked a couple of weeks ago in an episode about why visitors coming from AI tend to convert significantly better. And it's this. The evaluation already happened inside of the AI answer. So we must treat these as destination pages.

Michael 44:37 – 45:02

The person is there to confirm and the person is there to act. So for the validation part, I would just work backwards from our list of prompts for that specific entity that we are tracking associated with that specific page. What did this person learn about your company before they asked the question that produced your name? Whatever AI said about you is what they are going to walk in believing.

Michael 45:02 – 45:27

So the page has to confirm, I would say four things. That's what AI said you do is what you actually do, that you can be trusted to do it, that you've done it before for other people or businesses like them. And that there's also a very clear and obvious way to get started with you. These pages also still have to work in traditional search.

Michael 45:27 – 45:53

So there has to be enough on the page for a person to understand what your company actually does. It can't just be a headline and then a form to submit. But at the same time, we have to resist putting everything on the page. Because as we've been talking about, the primary job for our down-funnel pages is going to be about validation and conversion. So pulling it all together and also pulling together the series so far, right?

Michael 45:53 – 46:21

The infrastructure that we talked about in part one told these AI systems who your company is. And then the informational content in part two taught them what your company knows, and that you're an authority. And then the bottom of the funnel, is other people, other websites confirming both of those things on pages about what you do, often enough that your name comes back when the question moves from informational to who I should hire.

Michael 46:21 – 46:40

That wraps up this episode of The Search Signal. Thank you for listening this week. If this was useful, please go ahead and subscribe so the rest of the series finds you. And part four, which will be coming next, is the final one. And it is all about how to measure all of this. I will see you then.