Why top of funnel content stopped working in SEO

Google has been handing us a research project every time we search for a couple decades now and that is changing. That's all about identifying the search intents that you want to target and making sure that you are satisfying those search intents with useful information that is relevant. You need to be focused on your customers not Google. We used to start our SEO campaigns with articles. Now we try to do video first and then use the videos to write the content that we need. Dale Betrron is the founder and CEO at Fire Sparks Fire and Spark and I would love you Dale to tell a few words about yourself, your expertise and actually a business. What do you do? Yes. So I am Dale Bertrand. Um I'm an SEO guy. I run an SEO agency. Um we focus on um SEO only at our agency. So we don't tend we don't claim to be world class at anything else. Um and Fire and Spark is the name of my agency. Great. Thank you. Thank you for the intro and I think uh

we can move forward with the slides. So the stage is yours please. All right. Great. So I'm here to talk about SEO which is changing quite rapidly. We all know the technology is changing. AI is changing the way that we work as marketers. It's also changing search platforms and the way that they help our customers to find the information that they need when they're making a buying decision. And it's also changing our customers search behavior. And that's what we're going to talk about today. Now, inbound marketing was popularized by HubSpot. And I want to say hi. Hi. Uh, inbound marketing was was popularized by HubSpot. So, that is basically the idea that you're generating loads and loads of content. That content is capturing traffic from people who are looking for information related to your business, your product, your services. And when they land on your website, they have the opportunity to uh you have the opportunity to convert them. Uh the inbound marketing playbook is starting to fail for a

number of reasons that I'm going to go into um in my presentation today, but we need new strategies for building content for SEO. And that is what I'm going to talk about. So just continuing on with with my background, my intro here. So I've got a technical background. Um, I run a marketing agency. I studied AI in grad school. That's how I got into SEO. In school, I I was um studying computer engineering, so quite technical. Um, I worked on a team that built a supercomputer for the NSA. That is the National Security Administration here in the US. And that is a story best told over the over a beer since it involves uh spying on people in the US and people overseas. And then also, I'm the founder of um my SEO agency, which I've already talked about. But we have a bunch of clients. Our clients have logos and now I am going to move forward. So search behavior is changing as I pointed out because of AI, because the technology and search platforms are changing. Now if you think back, if you're as old as I am, then you might remember using the

yellow pages. And in the yellow pages, the phone book, what you would do is if you're looking for a photographer, you find all the photographers, you make some phone calls, one or two of them calls you back, and you hire your photographer. The yellow pages of the phone book taught us to search by looking for the category and making some phone calls. And it worked, even though it wasn't nearly as efficient as Google. And Google taught us to search in keyword language, keyword queries. So, we searched for things like best ice cream near me or lawyer in Boston. Um, it worked. It worked even way way better than the um the phone book, but it's still this weird query language and we don't always get what we want. And when it gives us a bunch of websites and web pages in response to our query, it's really handing us a research project. So, Google's handing us um all of these pages and telling us to find the answer ourselves. But now there's AI tools like

chatgbt where we can talk to a large language model that's been trained on everything that's been written on the web. And talking to this large language model is like talking to an expert in the field or a friend. Imagine if I'm going to Seattle and I'm looking for a restaurant for a client meeting. I'm lactose intolerant and this isn't my best client, so I don't want to break the bank for this lunch. I could have that conversation with chat GBT and get an answer as if I'm talking to a friend that lives in Seattle that has expertise in lactose intolerance and um clients that you know you don't want to break the bank on. Um but basically what it's teaching us and it's teaching our customers that they can have conversations to get answers. They can develop ideas with in conversation with AI. They can figure out what type of product they need. Do they need a product at all? Can they do it themselves? They can figure that all out in conversation with AI. Our customers are just starting to learn to search

with AI. So, this is early days. We're in the first inning to use a baseball metaphor there of this game. So, it's important that we're paying attention to how they're using it and what we need to change on our end. But the most important thing to remember is that your customers are lazy. And what I mean by that is they're human like me, like you. We don't want to do more work than we need to do. So, one of the reasons that they're going to gravitate towards these new AI platforms is because they don't want to do any more work than they have to do. Um, Google has been handing us a research project every time we search for a couple decades now, and that is changing. So, let me show you a little bit more about how that's changing. the old school inbound marketing playbook um with traditional search. What we would do is we would create top offfunnel informational articles um that are relevant to what we do. Maybe I run a landscaping company and I I mow lawns and I build beautiful um landscapes uh

for my customers and I'm writing about different eco-friendly landscaping products or when to cut your lawn or when to plant your lawn. like it's just all highlevel top offunnel stuff and when people are searching for anything related to my service they would land on my website and get information um and then maybe they just disappear but now I might capture their email address or I might cookie them for retargeting and then eventually I have the opportunity to direct them to a sales page where they give us a call or I stop by their house if I'm running a landscaping company and give them an estimate. That's the way it used to work. Now, what's different now is we have generative AI. So, in generative AI, they're they're going to have questions about landscaping or whatever else is relevant to your service, but they're not going to land on your website when they get the information. They're going to get trained by Google, Google's AI mode, or they're going to learn from AI overviews. And this is very different or they're using chat GBT and they're

learning in com in conversation with Chat GBT or some other AI. That means they're getting the top off ofunnel information in the awareness stage of their journey without ever visiting your website. Super important. Instead, we would like to make sure that the our customers when our customers are searching, they're going to get their initial information from an AI platform, maybe Google's AI overviews, but they're not going to land on our website until they have already been educated. Um, so at this point, they're in the consideration phase. They're in the middle of their buying journey and they're aware of their problem. They're aware of one or two solutions probably if they're landing on your website. They've decided that they want a product or service uh like you sell to solve their problem. And we need to be publishing midfunnel content. Midfunnel content looks different from that top offunnel educational blog post that we've been we've been u publishing for a long time. And and I'm going to show you in a minute what that what the midfunnel content should look like. Um but then we

have the opportunity to drive them to a sales sales page. Um this is really what is changing. Um so we're going to talk more about how search engines are changing. Imagine if you were looking for a CRM. Uh you the way Google and other search engines used to work was they would do lexical search. Lexical search means that it was looking for the keywords you typed in on pages online and showing you pages that have those words on them. So, if I'm looking for sales CRM for contractors, it's going to find a page with the title sales CRM and contractors in the on the actual page. And that's how it works. Now, we have semantic search engines. Even Google, forget about just AI platforms, but just straight up Google is understanding the intent behind the search, not just looking for those words on pages. And in this case, using this example, it might find a page that's about lead trackers, and it has the word builders in it, not contractors, because it's relevant even

though it is not even though that page does not have the exact words on it that uh we are searching for. Let's um let's keep going. So, how your customers are searching, they're using ranking engines like traditional Google. They're also using training based LLMs which are AI chat GBT. But here's the important thing we need to wrap our heads around. Uh going forward, they will be using hybrid ranking in LLM systems. This means that they're using something like Google's new AI mode. Probably people haven't played around with it yet because it just came out a week ago, but that is a hybrid where uh Google's doing some ranking and then giving you an answer from its AI. Then it will it will annotate the answer that it gives you with the citations which are pages that showed up in the ranking system. So you can see how the AI and the ranking are working together to identify the sources that the AI should use to craft its

answer. So what this means for us as marketers is we really want to be optimizing for both. And I am going to um get into that vector embeddings. This is an important concept. You don't need to understand it, but um it's important for you to understand that this is one of the bedrock bedrock technologies that search engines are using. What they're doing is taking your pages on your website, converting them into a numeric vector, plotting that vector in a multi-dimensional space, and when somebody searches for something like maybe used cars in Montana, they're measuring the the distance between the query and all of the documents online in this numeric multi venture uh multi-dimensional space. That's a long way of saying that they're using um this math behind vector embeddings to figure out how semantically similar every document on the web and different sections of documents on the web are to the query. Um the reason why I went through this long uh winded technical

explanation is because that has nothing to do with finding words on pages like keywords. Um keywords are quite obsolete. Like what we really want to be doing nowadays as marketers is to understand what your customers are looking for, the information that they need to make a buying decision and creating content that is relevant to what they're looking for. That has nothing to do with keywords. That's all about identifying the search intents that you want to target and making sure that you are satisfying those search intents with useful information um that is relevant. And that's what these the this uh vector embeddings um technology that you do not need to understand um is doing at the end of the day. But it's not matching keywords. That's important. Now, the reason why one of the reasons I point that out is because we're all using SEO tools, but the SEO tools are not keeping up with the technology that I just explained. SEO tools are doing things like um like domain authority and um they're looking at traffic which

doesn't which isn't as important anymore and they're helping you do keyword research which is also not um the right thing to be doing until you've identified the search intents that you want to go after. So SEO is changing. I'm going to sum it up real quick here. Um no longer lexical search engines are semantic which means they're not looking for keywords on pages. They're measuring the um semantic relevance of your pages to the the real intent behind the search. And our consumers are switching from ranking platforms to these conversational hybrid platforms. And marketers have new tools that we can use. And in my deck here, you'll have some of the prompts that that I use. So, as I said, search behavior is changing. What you're going to see is information like this um showing that Google still has 83% of the searches, but there are these other search engines out there like chat GBT that you should be paying attention to or YouTube. Uh what I want to say about this is you you want to take it with a grain of salt because what actually matters at the end of the

day is the way your customers are searching. So you want to set up tracking and I borrowed this slide from Andy Crestadena. Thank you Andy. I wanted what you want to be doing is setting up tracking for AI platforms in your Google Analytics so that you can see which chat bots you're you're getting traffic from and what pages they're they're landing on. That's really important. But in addition to that, you need to talk to your customers. Remember I said search engines are no longer lexical. They're not looking for keywords on pages. That means if you can talk to your customers, not hiding behind analytics, then you'll understand the information they need, the questions they're asking. When you're looking at SEO tools and getting a list of keywords, it's not helping you understand the why behind the search. And it's the why behind the search that is going to help you build the right content. That's why you need to be talking to your customers. This is going to be in my slides when you guys get my slides. Um, but this is really just more about how SEO is changing. But the the

biggest point here is that you need to be focused on your customers, not Google. We've all been saying that for a while in in my industry, but it's even more true when it comes to um AI platforms. Now, when we're educating large language models in AI, um, remember the reason why we're doing it is because your customers are confused and your customers have questions. They used to go to your website to answer those questions. Now they're going to go to an AI search platform. So, educating LLMs, there's really three things we want to focus on. The first one is establishing presence with co- citations, which means to make which is I'm going to show an example of this, but making sure that your brand, your products, your services are mentioned online alongside whatever concepts and topics you want to come up in conversation for. The next one is PR. Like the more you're mentioned online, remember these AI large language models are training on everything that's been written on the web. So you want your brand and the right information about your brand and your products to show up in as many places as possible with PR

tactics and digital PR tactics. And then also reinforcing semantic relevance. So this is making sure the content that you're creating is relevant to the searches you want to show up for. Now, let me give you some examples. Um you might have this is a page online. It's a um product page on Amazon for the iPhone 16. There's a review on this page and I'm going to show you the review. The customer review says Apple intelligence is underwhelming compared to other assistants. Track TBT voice is impressive compared to this crap. So what's going on when AI finds this during its training is that it is associating Apple intelligence with underwhelming. So now it's going to believe and tell people that Apple intelligence is underwhelming and it's also associating chat GPT voice with impressive. So this is important. So you can imagine how you want your brand, your products to appear in the right context with in co-itations. And the co-itation here is capital intelligence and underwhelming. You want Dale's new

toothpaste that I'm I might decide to sell online associated with lifechanging if that's the concept I want associated with my new toothpaste product. Moving on here. When you're doing PR for AI search, remember the goal is you want information about your products, your brand, and the right messaging related to your products and your brand to appear online um as frequently as possible. And I know that my screen is disappearing briefly here, but I'm adding it back. You do that with thought leadership content, building influencer relationships so that you can get uh your information out there, doing pred traditional PR and then optimizing for featured snippets which is important. There's a lot of things we used to do for fe featured snippets that are still relevant and then managing your brand reputation which is really about your digital footprint uh for your brand online so that it appears as much as possible like frequency really matters when it comes to AI um in its training. We also want to make sure the content we're creating is optimized for clarity. Uh what we like to do is add it's not

just like a straight up wall of text blog post. Um but we're building guides that have tables of contents, call out boxes, Q&A at the end, some key facts. We're really creating bite-sized chunks of semantically dense information that is um that the AI can find. Now, I've got a couple prompts. Um actually about a whole bunch of clearly I have like 30 prompts. I'm going to give you some, but I'm also going to give you a PDF. You guys can grab um my slides through this QR code. Um and what I'm going to talk about next is a bunch of prompts that I use to create relevant content that's relevant to the search intents that my customers are searching for. And then also to increase the semantic density of the content that we're creating. Now what semantic density is about is you can imagine a lot of using a lot a lot of words to convey a concept or you can imagine conveying exactly the same concept and information with less words. And that's what we mean by like semantically dense. So it makes it easy

for the AI to find it and it makes it easy for the AI to match it up to the search intents you want your uh brand to show up for. And then also it makes it easy for the AI to determine whether your content is relevant to the searches that you want to be relevant for. And um I realize that we might not have seen my QR code because I'm I'm losing my my screen share here every now and then for a weird reason, but here's the QR code. And what I'll say is the first prompt that I'm going to show you is really about relevant to engineering. Um so grab my slides using this prompt. Um, but the prompt that I'm about to show you, what it'll allow you to do is to um, well, basically I'll read it. Um, you're an SEO specialist focused on user intent. This is what we're telling the AI. Uh, your task is to fetch the URL below and determine the underlying action that the user wants to take or the information they want to acquire when they read this page and then translate that into a concise four to six words that begins with a verb. This is a search intent. So, what it's doing is it's taking a page on your website

and telling you what is the search intent that AI is likely to send to this page. Um, that is important. We what we like to do is we like to run this prompt against all of the pages on our website so we can get a map of the search intents that we're already targeting. And then here's an here's an answer just for this um content marketing software page where the search intended gave us back is compare content marketing software options. So basically the the content on this page is written to target people who are going online and searching to compare content marketing software options. That's different from another intent that might be they're looking to buy or sign up. It's just quite different. So you would create subtly different content and that's really important. Um you can also measure uniqueness with AI. I love this prompt where I say rate the novelty of each of these article top topics from 1 to 10 where 10 is the most novel. Remember that AI is trained on everything that's written on the web. AI is not great at being like creative and

creating new ideas, but it does know what's average because it was trained on everything. So, that helps. And you can see an example here. It basically gave a novelty rating of five out of 10 for one of the ideas I gave it. Um, if you're hiring writers, um, I love this. We used to hire writers like Ally. Alli's not her real name, but she works on my team, and we hired her because she's awesome at pitching ideas, storytelling, writing, editing, proofreading. These are the types of writers we used to hire. But with the changes I'm talking about in crafting the type of content that I'm about to show you, you really want what I'm calling more of a content updater, like somebody who's understands AI optimization. They know how to do data analysis. They're good at community engagement. They know how to do PR, uh, like I talked about for AI, and they're good at crafting messaging for calls to actions that are aligned with the search intents that we're targeting. That's different from your typical writer toolkit and skill set. So think about that. Also, my content updator, you can tell by the name of the role, she is updating existing content, which is

usually um the lowhanging fruit, much better use of your time than creating new content. Um that's why we call her a content updator. But what I want to show you guys is uh the type of content that we are building for AI and AI search engines. So we're doing a lot of datadriven comparisons. These are databasebacked content. Remember, if your content is either interactive or maybe it's a tool, maybe it's software or it's one of these datadriven comparisons, this is content that Google's AI and no AI actually can summarize. So, when we're talking about losing traffic because your customers are getting the information directly from Google instead of coming to your website, if you were to build a tool, and I'll show some examples, or one of these databasebacked type of data driven pieces of content, it can't be summarized because the data is in the interactive tool or the information's in the interactive tool or the data is in the database and your your pages are really um just presenting

the data with visuals. This is content that AI cannot summarize. That's important. Now, also original research doesn't have to be hard. I always tell people that you can find government data. You can find uh other sources of data online to to talk about in your original in your original research or do a survey. But original research is is great because it's novel and also it allows you to do build some of this databasebacked content for comparisons and then also for visualizations and that works really well. So now I'm moving on to video. So if you're um if you're doing any type of video, it works well. Like we we used to start our SEO campaigns with articles. Uh like I said, now we try to do video first and then use the videos to write the content that we need. Um there's so many video platforms uh where you can get visibility if you have video content. So that is one of the reasons why it works. Um, video also works well because your

competitors are probably not good at video. Um, and you might tell me you're not good at video either, but get good at video because it's it's just going to be a huge competitive advantage for you and your brand if you're creating some video content. The last thing I'm talking about here in terms of types of content is interactive tools for decision support. Remember I said that it's important for you to understand how your customers make decisions, the information they need to make decisions, purchase decisions for your products, for your services. So you can create these interactive tools that Google's AI cannot summarize so that people do need to come. Um and then also they're sharable. Um you can do some PR around them. So it might be a pricing calculator, a needs assessment, a configuration tool, delivery estimator, product comparison, whatever it is that you need. But this is the new SEO content. And then um last slide I'll show here. When you're creating these interactive tools, you can create them with AI. And I have a prompt here. You'll see this when you get my slides, but it says create an HTML page that contains an interactive tool with an intuitive layout and a light blue color scheme. Here's the inputs. And then the name of it is how much should SEO cost?

And remember, I'm an SEO guy. I run an SEO agency. So you could imagine putting a tool like this on my website. Um but the reason why I'm showing this prompt is because you do not need a developer to create these interactive tools. these interactive tools you can create them in AI cha in a canvas or you could do it in in cloud or with gemini like it's it's you do not need to be a developer this is a noode solution you are only limited by your creativity as a marketer when it comes to creating interactive tools now this is my last slide I am done um I think I finished with enough time for some questions and um but I do want to give you guys my AI recipes the prompts that I use to do SEO strategy content and strategy. Um, maybe 30 of them and I gave you a couple already in the slides and that is my time for today. Thank you. Thank you very much. That was really insightful, really actionable. Thanks a lot for sharing additional materials. I think that our guests will love to try them out and to see how uh

it works for them. And uh yeah, we definitely have time for to cover questions. So, let me start with the first one. Hello Dale. In your experience, did companies see drops in engagement or trust due to overly gener generic AI content? And how are they adjusting strategies now to recover and build audience trust? So there's two parts to that question like one is about are they building generic meto low value content with AI and getting dinged, right? And then the other one is building trust with an audience. I think those are two different questions. So, we definitely want to avoid um creating meto content with AI. That's why I love the prompt that I showed you where AI can rate the novelty of your ideas before you build the content. Um the trick that I want to tell you there is it's not just article topics, right? Like you might use the prompt that I gave you and give it a list of article topics and say, "All right, well, these are the ones that are the most novel, so I'm going to write about those." I also

give it a list of content format so that and then the formats might be video or like I talked about interactive tools or datadriven comparisons or a million other things that you could think of you use AI to help you to to brainstorm around the different types of content formats but it's also content formats like what is the best way for for your customers given their information needs when they are making a purchase decision to get the content that they need. It might be a video, it might be a tool, it might be an article, it might be a guy. There's so many of them out there. So, I would do it I would do it that way, too. And then the second part of the question was really about trust. The reason why I pulled that out is because we used to think like there's a mindset shift that we have to make where we used to serve Google with content that contain the keywords that we wanted to drive traffic to our pages. Now, we need to serve our customers and think more of a like a publisher that has is basically narrowed in on an audience and serving that audience. And and I'll get back to trust. What you're doing when you do

that is you're building trust with your audience. And and what this should feel like is when they are looking for information um related to the type of information that that you publish, they navigate to your website directly without finding your content through Google because maybe they found it through Google or some other means the first time they got there, but now they're in your audience and when they're looking for information on the topic, they go directly to your website or your blog or or your YouTube channel, whatever that is. But that's building an audience for your content, which is different from building content that will show up top of funnel for informationational searches in Google. That's a very different mindset. Thanks a lot for covering it in so much detail. I hope we've satisfied the interest of our guest. And then we're moving forward to the next question. Do you think people will get tired of AI content? What's the next step for inbound marketing if that happens? Well, really depends on what we mean by AI content.

Um, I'm already tired of blog articles that are generated by ChatCBT and you could tell when you read them and hopefully Google's doing a good good job of not showing too many of them, but they're out there. We've all stumbled upon them. So, yes, we're already tired of AI content. when you're talking about blog articles written by ChatgBT, but there's so many ways to use AI to improve your content in a way that serves your customers and builds trust and builds an audience around your content and ultimately sells your products or services. And that's really thinking about how it can help you do the upfront research for content, how it can help you brainstorm formats and topics. Um, when after you're done writing a piece of content, you can give it to AI and ask it how would my persona respond to this? Is there anything missing? Is there information they need that's not in here? Um, did I make come to any logical conclusions that weren't supported um by what I wrote in this piece of content? There's so many many

ways to use AI to improve the quality of content, but asking AI to write an article start to finish on a topic is not one of them. Awesome. Awesome. Thanks a lot. Now moving to the next question on how can brands stand out from the AI's noise. Is it all about voice expertise and community things AI can't fake? Yeah, I mean I I think it's about I like the I got many ideas on this, right? But ultimately you need to be relevant to the searches that your customers are making, right? And and I've already talked about that. But then also new formats like these interactive tools that anybody can create with zero technical skill uh using AI. That that's important. That's like a new SEO format. And and I can't stress enough why it's important and valuable. Uh it's important and valuable because not only are they easy to create, but Google's AI can't summarize them. So you will get traffic and they'll be like sharable sharable pieces of content for what

you're doing. Also, we want to make sure that we're building an audience, right? Like really what it comes down to is there are there's there's a there's a channel I go to when I want to know the weather. There's a website and an app I go to when I want to know the weather. Um there in in similar to that, if I'm interested in a topic like SEO or AI or or tech startups, there's there are specific sources that I will go to because I trust them. I've built a relationship with them. And that's what you want to do with your brand. We spent so much time trying to build a relationship with Google that I still hear people asking the question um how do we get this to show up in Google? How do we get traffic from Google? What about serving your customers so that they come back time and time again until they're ready to buy? And that's how you're building um an audience and a pipeline going forward. Great and very comprehensive. Next question. Uh did AI flood the market with lowquality content hurting content marketing? And are we now seeing a shift back to strategy and quality? Yes. Yes. And yes, and yes. So,

um, goo AI has absolutely flooded the web. You, but we we don't notice as much because the search engines are good eliminating and filtering out most of the crappy AI content, but we still see it. It's just we most of it we don't see because they're pretty good at it. And like I said, that does mean that we're going to have to get back to building relationships with real humans and figuring out how to satisfy their information needs in a way that also uh serves the the goals that we have as business people and marketers in an organization that needs to build pipeline and do customer acquisition. Okay. Maybe you have like any uh specific advice in terms of how to uh build strategy to make sure you're out of this uh AI flat of uh of the market and still stand out with your unique voice. Yeah, I would say it's all about unique voice is really the operative word there. Figuring out what is that niche that you're going to serve with the content that you're creating. And

not only that, but figuring out like what is your opinion like your your brand needs to be opinionated about something and that's very important. So if you are consistently targeting the the niche that you have identified and and described and you're talking to your niche because a lot of us are digital marketers. We're hiding behind analytics and and and keyword research tools. So we're not actually even talking to our customers. But if you're doing that and building the right content for them with an opinion and and don't be shy about your opinion. Not everybody's going to agree with your way of doing things or the way you design your products or the way you decide to deliver your service. But whatever that is, own it because that is what's going to make you unique and build an audience around your brand. Great. Thank you for covering this question. And probably the last one is so if generic blog post BS are now easily produced by AI, have you seen a shift towards strategies like AIdriven personalization, interactive tools or

multimedia content as the new way to attract people? Yeah, absolutely. I mean, we're just getting started with that. So, you could imagine that AI personalization is going to be absolutely huge. Right now, what we're seeing is a number of different tools and methodologies for personalizing email content. And that seems where where it's starting and I would play around with that. And those tools are useful and effective to the extent that they have information about your customers. So what I would do when it comes to AI personalization is I would think about well what what information do you have? Do you know what pages they visited online? Do you know what their motivations are? Do you know do you are these longtime customers that you have you have a lot of data about them maybe in your uh customer database? and and I don't know what that is for you, but think about what information you have that you could use to personalize email. And then from there, we're going to surely be doing the same thing um on our websites and with social media and and with these interactive tools as you mentioned. But but start with email so

you can start really wrap your head around how to do personalization in your market. Great answer and great information to wrap up your presentation. Thanks a lot, Dale. We're very happy to have you here today sharing your expertise. Thanks. Well, thanks for having me. It was fun.