← Voices of Search // A Search Engine Optimization (SEO) & Content Marketing Podcast
Voices of Search // A Search Engine Optimization (SEO) & Content Marketing Podcast
How To Win Customers In An AI Search-Driven World
00:27:48
Loading summary
Transcript23 lines
- [00:00]
Jordan Cooney
The Voices of Search Podcast is a.
- [00:02]
Chris Andrew
Proud member of the I Hear Everything Podcast network. Looking to launch or scale your podcast, I Hear Everything delivers podcast production, growth.
- [00:12]
Jordan Cooney
And monetization solutions that transform your words into profit. Ready to give your brand a voice?
- [00:19]
Chris Andrew
Then visit iheareverything.com welcome to the Voices of Search Podcast. A member of the I Hear Everything Podcast network, ready to expedite your company's organic growth efforts. Sit back, relax, and get ready for your daily dose of search engine optimization wisdom.
- [00:38]
Jordan Cooney
Here's today's host of the Voices of.
- [00:40]
Chris Andrew
Search podcast, Jordan Cooney.
- [00:43]
Jordan Cooney
AI is changing search forever. Brands must adapt quickly. Traditional SEO tactics are losing effectiveness. Visibility in AI results requires new strategies. How can businesses maintain relevance in this AI driven landscape? According to Search Engine Land's 2023 survey, 60% of consumers are now using AI search tools for product research. This highlights the urgent need for brands to optimize for AI powered search experiences. Here's the challenge. AI search doesn't rely on traditional ranking factors. It prioritizes different content elements. Brand visibility requires new optimization approaches. How can companies ensure their appearance in AI generated search results? This is the Voice of the Search Podcast. I'm Joran Cooney and joining us today is Chris Andrew, co founder and CEO at Scrunch AI, which helps businesses optimize their presence across generative AI platforms. Chris will share how to win customers in an increasingly AI search driven world. Chris, welcome to the Voice of the Search Podcast.
- [01:55]
Chris Andrew
Thanks for having me, Jordan. Great to be on here.
- [01:58]
Jordan Cooney
Absolutely. And I mean, no better time than now to talk about where AI is going, how discovery is changing for consumers. You know, I just want to just jump right into our first question about AI and how it's fundamentally changing the search landscape. You know, what are the differences that SEO professionals need to understand between what is traditional search and what search has been for the better part of 20 years and where AI powered search experiences are going.
- [02:26]
Chris Andrew
Yeah, great, great question. You know, to go back to the stat you listed in that opener, 60% of consumers using AI for search in 2023. That was really before these models had web search capability. So that was consumers accepting out of date information. Right. They were using ChatGPT to ask a question and the model would warn them, hey, you're searching for a product, but this is six to 12 months out of date. But still consumers prefer the experience, right? Still consumers move towards simplicity. And so when I think about what's changed, it's that I don't want 10 links from my search experience. That was Just the best we could do at the time, right? I want an answer. I'm searching because I want an answer to my question. So at Scrunch, there's a couple themes we talk about with our customers, which is, first and foremost, consumers move towards simplicity. Folks are habituating to AI search because I get an answer instead of 10 links. Now if I'm getting an answer, that means I'm not visiting as many websites. So one of the themes we're seeing unfold in the search space is fewer people are visiting websites early in the funnel. I'm moving my research and discovery process to ChatGPT, Perplexity, Meta, Claude, even Google AI overviews, which means I get an answer to my question rather than having to go visit three or four different websites. In practice, I'm outsourcing browsing. At its simplest, I'm outsourcing browsing to an AI crawler to make sense of Internet content for me to determine what I need to see to answer my question. And so I think that's the big shift that we're seeing, you know, in terms of how it's impacting search engine professionals. I think sometimes these shifts get a little overblown. I think search engine optimization is still the foundation for how these models work. Right. If you take the time to understand what's happening behind the scenes is these large language models are looking at the search index, they're looking at sites that they can consider to use in answering the question, but once they are on your site, they're downloading that information, synthesizing it with other sources and determining how to answer the question and who to cite. And so there's a number of significant changes in that space. But at the core, I think the customer journey is changing. I'm less likely to visit a website, I'm more likely to get an answer from my search experience, I'm more likely to refine that answer in a chat based experience with my mod of choice and maybe move along from there. And so that's a really, really big change in the customer journey that we're tracking with our clients.
- [04:56]
Jordan Cooney
No question. One of my favorite examples is the recipe experience.
- [05:01]
Chris Andrew
Right.
- [05:02]
Jordan Cooney
For the better part of the last two decades, our recipe experience has been putting in one of our favorite foods to eat and just getting a bazillion pop up ads, a very long story about grandpa's way of making lasagna, and a very confusing and muddled way of getting to the ingredient list and steps that you need to complete this recipe. But today, if you go to Any AI model, put in your favorite recipe, you get a very simple, easy to digest, no pun intended, list of ingredients and steps into the process, which is really the simplicity, the ease of use that is AI models and AI search and the direction that consumers want. Right? They just want a faster, easier way of getting to the outcomes that they need. In order to achieve this, though, we have to be able to really focus on optimization for a whole new set of models, a whole new set of experiences. I'm curious, Andrew, how we do that? What does optimization, how do we make it in this new AI search? How do we become present there?
- [06:08]
Chris Andrew
Yeah, I mean, the recipe example is a visceral one, right? We've all experienced it. We go to Google and ask for, you know, I've got this in my fridge, I want to cook. And you get a link to a recipe and you see the life story. Three ads, an email pop up, 50 keywords, a backlink, and you can't find the recipe itself. You go to GPT, you ask the recipe you put your ingredient in, you get an answer back that simplicity is a delight. And I think ultimately consumers move towards simplicity. And I think in optimizing for this, you need to think more human language. And what I mean by that is I think we've had a period of time where folks have built content for Google. It's like Google's the primary consumer of my content. If I'm found by Google, I get listed by Google. Paid organic maps listings. Depending on the nature of your business nowadays, think about the fact that large language models have been built by humans, trained by humans, and they're looking for language, they're looking at the intent of the prompt. What is the question that this person is asking and what is the best closest match to answer this? What patterns am I seeing in the. In the data to kind of uncover what this person needs and what sites are going to be a closest match to the intent of the question. And so I was recently on a call thinking of, you know, you need to write for humans and design for bots, right? If the primary consumer of your content becomes the chatgpt crawler, the Perplexity crawler, the meta crawler, the AI crawlers accessing your site, you need to make sure that the content is available in a simple form, right? These models are looking for unstructured data, looking for language on your site that they can download, scrape and synthesize into an answer. And so at the same time, you need to write for a human because the model is thinking like a Human, it's looking for language. And so we're finding that the long tail of content is of increasing importance. Right? The models are going deeper to find the intent and the match to what the human is searching for. Other things that we're finding is these models are trying to be unbiased in their answers. They don't want to just go to the company website and answer the question. They want to get user generated content, be it from Yelp or Reddit or review site, they want unbiased reviews. From a listicle perspective, we're finding that our customers that have lists on their own website about them versus competitors really drives a significant impact, right? It's like, hey, here's my product compared to my five peers. We're cheaper, they're more expensive, we have better features, they have fewer features. That comparison is part of what the model is trying to do behind the scenes. And so if it can find a source that does part of that work for it, there's a greater likelihood you're going to be incorporated into the ultimate answer. And so again, I think writing for humans, but designing for AI crawlers, meaning AI crawlers don't necessarily need a deep visual experience. In fact, it's pretty well documented in the search engine community that JavaScript causes a lot of problems for crawlers. While the Google bot has gotten quite sophisticated in its retrieval and understanding of intent, these AI crawlers are more elementary. They're looking for content and exact match and struggling when it's buried on the page. And so those are some of the big themes that we're tracking in terms of optimizing for these models basically in real time as the shift begins to unfold.
- [09:24]
Jordan Cooney
And you know, Chris, you brought up a couple interesting points and I want to dive into this more because the differences between traditional SEO and where we're going today in terms of supporting these AI models, making these AI models successful in understanding what our products and services are, go beyond just the content elements. You just brought up a great example of crawl and crawl capabilities. But when you think about prioritization and what we should be prioritizing not just as SEOs, but as marketers, as business owners, as leaders and executives in companies, how do we prioritize the aspects that are going to be critical, the elements that are going to be formative, and the structure by which we serve content and information to these AI search models to become more present. So essentially what do we prioritize? How do we create that pegging order of things to do to be as effective as possible? With our time and resources to show up in these AI models, I would.
- [10:32]
Chris Andrew
Say start from the customer perspective. And I think what's amazing about this shift is that's often where dialogue starts with a new customer of ours. We'll speak to a CMO who said, you know, I did a prompt and I thought my brand would show up, and it didn't. That's kind of surprising. The question I ask on Google vs. GPT vs. Gemini, powered by Google, yields three different responses. Why they're processing information and retrieving and returning information in a different way. So when I say start with the customer experience, there's an advantage in the fact that we are all shifting our behavior to AI search, right? Like, I don't need to come and sell you that this change is happening. You're starting to use GPT more day to day instead of Google, you're starting to use an AI overview instead of visiting five websites. And so one of the core things we do with our customers is we automate the creation of Personas for a given brand. Who do we believe you're selling to now? You can come into our platform and refine those Personas, but we think of Personas as the foundation to prompt generation and prompt monitoring. Because people are getting different results from AI models based on the lens from which they're sitting, right? So if you're a finance executive sitting in Manhattan asking questions of GPT, you're asking questions in a different way, you're getting answered in a different way, and different sources are being cited. That's fundamentally the case. If I'm a university student sitting in Boston asking questions about finance, I'm coming from a different lens and a different perspective. And so these models have memory built in. They're adjusting their answer based on the nature of the prompt, they're adjusting their cited sources based on the intent of the user. And so when I think about what organizations need to be doing, it's deeply understand the Persona of who you are selling to deeply understand their experience of what they are learning about your category, your brand, your competitors in their native search experience, whether that search experience is in GPT. Perplexity. Gemini AI overviews. Different populations use different models, right? So If I'm a B2B SaaS company heavily focused on APIs and developers, Claude Anthropic tends to have a bit more activity and volume compared to GPT, despite GPT being the leader and so understanding at scale, what is the nuance for how my brand is showing up? Am I showing up in top of the funnel, non branded queries where people are doing research but not yet asking about my company. That's why we spend so much time in the enterprise platform foundation of our product to say, look, these are your Personas, these are your competitors, these are the types of questions, these are the actual responses and citations. And if I'm a brand, really understanding that new customer journey is paramount. Right. I think the shift is that my consumer may visit my website. Way further down in the funnel, we're seeing that referral traffic from AI sources is the highest converting organic channel and in some cases it's converting higher than the paid channel. Why? Because they've done so much research in the model, by the time they click to my website, they're ready to act, they're ready to book a meeting, buy the product, talk to an advisor, whatever the action might be. They've made their decision in the native search experience. And I think that's the number one shift when it comes to what am I optimizing for. Start with the new customer journey, who are you selling to? What are they experiencing in the channels they prefer? And then start to think about the content gaps, the third parties that might help address those gaps and what you can do about it.
- [13:58]
Jordan Cooney
Overall, I'm loving this conversation because it's really a unique challenge that we're breaking ground on now, right? We only have about two plus years of understanding and experimenting and experiencing ourselves as consumers. How these AI models help us go through discovery, how we go through the process of seeking out information, finding new things to buy, understanding a new service, learning about the world, whatever it may be. These AI models help us consume information and become informed or capable of doing something. And as we learn more about that, it becomes increasingly apparent that the way we're measuring things is changing. And historically, as an SEO, it used to be the game of how do I get ranking data? How do I get the most keywords into a platform to tell me how I rank in these 10 blue links? And that even evolved over the decades to different SERP experiences and ways in which Google has evolved both paid and organic listings. But at the end of the day, measurement is a key component to success. It's the best way that we can understand whether what we are doing is working or not working. How should businesses track and measure their success in AI search environments? This has got to change and I would really like to hear your point of view on how it's changing and where we should go. Chris?
- [15:34]
Chris Andrew
Yeah, it's a great question. I think monitoring is the foundation I honestly believe monitoring is something every organization is going to need to do. The amount of companies I've met that have assigned a person to be manually running these searches across the large language models, putting it into an Excel document, turning that into a graphic, and not realizing that the results they end up getting back end up becoming very biased because the model starts to know, oh, this person's asking a question. This person wants to hear me say their company, right? If I keep asking the model these questions, it thinks that it knows who I am, it has memory, it starts to answer questions in a way geared towards me. You might even get it to a point where it's like in parentheses, your company is a great option in this field because it knows you. And so you really need to think in a more sophisticated way about your monitoring practice. And it's why we've built out such an enterprise platform focused on who are your Personas, who are you competing with, what are your themes and kind of traditional keywords? How do those translate to prompts in the new language of search? Because we're no longer in a search world where I put in the word, you know, Walmart, Nobody goes to ChatGPT and just types in Walmart, right? That's what we do in Google. We go to Google, we type in Walmart and then Walmart has to pay to show up first so that I click on the Walmart site. That directional search is quite the monetization engine. People in GPT tend to ask very intent driven questions, whether that's in the research phase or in the comparison evaluation phase. It's another reason we organize all of our prompts by stages of the customer journey. Is this a top of the funnel prompt where someone's doing a highly considered research stage that has nothing to do with your brand, but your brand should be present versus hey, I'm comparing brand A against brand B. Of course you're going to be present. You know, those components become really, really important to pay attention to. So on the monitoring side, I think monitoring needs to be looked at comprehensively. What types of things do I need to be showing up for in the top of the funnel? How do I need to be represented more towards the bottom of the funnel when my product is being compared against my peers and competitors? And from a comprehensive nature, these models are non deterministic. And what that simply means is that you're going to get different results for the same question at different points in time. If you ask GPT the same question twice, it will give you a slight variation on the answer, potentially different sources, potentially different answers. It's certainly giving different answers based on memory of people in different locations. But even the same exact question from the same user comes back with a different answer. So what does that mean? It means we need to track this at scale over time. You can't be like, oh, I did a search. We show up number one, because that same search might put you at number three. And a week from now, maybe you're not present. And so everything we do is at an enterprise grade, which is comprehensive monitoring multiple times throughout the week, capturing that information. Some enterprises that's multiple times a day, others is a handful of times a week. But ratcheting up that monitoring at scale so you have real time series data to say, okay, you know, let's pretend I'm a clothing brand. And making sure that my presence is showing up for the back to school specials is important to me. Well, we might need to run hundreds or thousands of prompts about that mother Persona prepping to, you know, clothe two or three kids going back to school. Is my brand showing up, yes or no, for this variation? For that variation? What partner sites are being linked? Is it going to my shopping page or a blog page? Tracking all of that at scale is the foundation to understand the changes you need to make. And I think from a monitoring perspective, people get a little confused in the space right now. I would encourage companies to track the things you can actually track with certainty. I think there's four things there. You can track the frequency with which these crawlers are visiting your site. Right. We can work with you to do that. Right. So if you want to understand how frequently is chatgpt scraping my website to look at my content, we can track that. The second thing we can track is referral traffic. Am I getting visitors from GPT, Perplexity, Gemini, Claude to my site? You can track that. Once they're on your site, you can track conversion. Hey, this was a referral visitor from a ChatGPT search. Are they converting at a higher level? At a lower level? What pages are they landing on? How much time are they spending? And the fourth thing you can track with confidence is citations. Am I getting my website listed in these AI search results or am I not present? Now, there's a million small things in there. Your presence score, your position score, your sentiment score, your ranking compared to competitors. But again, the four core things that looking to track, if I were an enterprise or any business thinking about the shift to AI search would be frequency with which these crawlers are looking at My content, my referral traffic of how many visitors am I getting from these sources? The conversion traffic, conversion rate, hey, compared to Google are LLM sources that are referring traffic to me, converting at a higher level and my citations, which sites that I own are getting referenced and linked to in AI search results and how does it impact my overall performance? Those are the four things I'd be tracking from a monitoring perspective.
- [20:35]
Jordan Cooney
When we think about performance, right? You know, most businesses are going to be thinking about conversions, they're going to be thinking about leads, they're going to be thinking about some kind of an event, right? And historically SEOs, the traditional SEO path was leveraging a collection of vanity metrics. What may it be a rank position, may it be click through rate, may it be visibility or some sort of an index of these KPIs to then understand how frequently you're showing up and at what degree you're showing up based on your competition or the market. Competition is very different when we think about an AI model and the results that come out of it. And you mentioned citations a couple times here. As we think about driving to performance, how should we be thinking about what metrics lead into that when it comes to AI models and monitoring and tracking?
- [21:36]
Chris Andrew
You know, one of the things we identified early is that your brand is no longer what you say it is. That's a scary thing for marketers, for leaders, for the board. For a long time, my brand was what I said it was. I could communicate it on a TV ad, I could get somebody to click on my website and read my mission statement or my vision or what my product is. Nowadays when I ask a question about a brand, a product or service in GPT. GPT is representing the brand, product or service, it's telling me what it thinks of it. That's a completely different orientation. Now the question is how is it arriving at that answer? And I observed early in the AI shift that many organizations were organizing their internal data. Makes sense. Organize your internal data for rag, clean it up, label it, get the right place sources in the right places and start running workflows and outcomes on your internal data sets. Protect it. Very few companies were thinking about how all of the external data in the world is being used to represent your brand. And so when I think about leading indicators and what to measure, I want to understand what are the sources that are answering for my brand? What are the sources that the models are using when they're trying to answer questions about me? Right, let's, let's pretend I'm walmart Right. If I'm Walmart, you better bet the AI models are accessing the Walmart website to use that as a source. They're also accessing the rest of the Internet, news articles, blog posts, third party sites, competitive sites that pit themselves against Walmart. Right. Walmart's a very large box retailer example here with an E commerce presence. But the message to be delivered is I want to understand how my brand, product and service is being represented. What are the sources, what are the things fueling this answer and how do I understand how I can impact that? Because let's say, for instance, the sentiment towards my brand is negative on a search engine, on a large language model, search engine via GPT, Perplexity or others, it's getting that content from somewhere. I was on a demo recently with a very large Fortune 500 provider who was really upset about how the model was representing their brand's products and services. When they used our platform to dig into why they were being represented that way, what they found out was that there was a lot of content on the websites they owned on their community page, on their help forum, consumers submitted feedback very negatively representing their brand and they were not going in and addressing those complaints. And so when the model was like, well, what's the reputation for Product xyz, they're out looking at the Internet, looking at this company's website, and they found a bunch of negative content on the company's website about their brand, products and services. And so when the model was being asked to compare that product against another, they were linking to the core domain and the subdomains of that business, saying, hey, well, here's what the company says about itself. The community site says it's a terrible product and there's a problem with it. No wonder the brand is being misrepresented. So when you think about how all of the external content in the world is being used to represent your brand, I want to monitor towards what are the sources that are driving the output that represent me. That's kind of my starting point if I'm a small, medium or large business.
- [24:46]
Jordan Cooney
Yeah, I love that. And I think that's such a critical transition from where we've been so inward focused on our own website and now looking beyond that to all these sources to better understand how we can influence the main metrics of our business as we close out this deep dive. What are some specific steps that SEOs and content marketing teams can take over the next 30 days to begin optimizing their digital presence in AI search platforms?
- [25:13]
Chris Andrew
Yeah, there's a self serving answer to this, which is come chat with us at Scrunchai. There is a manual process that you can implement yourselves which I would deeply encourage you to go do, which is go start paying attention to it. The things I listed out of the four things you can measure and monitor, some of those you can go start yourself. How frequently are these models hitting my website? Go look at your web logs. Go work with your technical team to understand that. Go set up your analytics platform to track referral traffic. Obviously we help you with both of those things and do that in our platform there's a number of things you can do on your own to start paying attention to it. Deeply understand what content that you've created is valued by AI is being cited by the models as the answer. Go look at the third party presences that these AI platforms are prioritizing to answer questions and think about your relation to them. Are they pay to play? Do they have the misrepresentation of your pricing or product description? Go get those updated. We are oftentimes seeing very basic changes resulting in a significant lift in presence or a significant lift in referral traffic for fixing misunderstandings from the model. I think one of the reasons this one of my beliefs is that all of search will be AI Search. I think some of the new acronyms are silly. Aeo, geo, aio. There's a million new acronyms. Nobody knows what to call it. I believe we're going to call it Search. When I search I want an answer. When I search I want an outcome. Minimize the work I need to put in. What's happening when these people search in the new way? Are you showing up? Go dig in and start to track it and start to test new content. Start to think about the fact that these models really prioritize rich content sources, rich text based content sources, be it a knowledge graph, a knowledge base, FAQs, glossaries, these types of resources are really rich assets that the models are looking to to try to answer questions at scale. And so those are some of the big things I'd be thinking about, you know, as I move forward.
- [27:11]
Jordan Cooney
That's excellent. Thank you so much Chris. We're going to transition now to our lightning round that wraps up this episode of the Voiceless Search Podcast. A huge thank you to Chris Andrew from Scrunch AI for joining us. If you'd like get in contact with Chris, you can find a link to his LinkedIn profile in our show notes or visit voicesofsearch.com for more information about Scrunch AI. Visit scrunchai.com if you haven't subscribed yet and want a daily stream of SEO and content marketing knowledge in your feed, hit the subscribe button in your podcast app or on YouTube and we'll be back in your feed soon. Okay, that's all for today, but until next time, remember, the answers are always in the data.