It’s time for another episode of The Search Session. I’m Gianluca Fiorelli. Joining me is Emina Demiri-Watson, a digital marketer based in Brighton, UK.
We talk about staying focused as AI search creates more possibilities, pressures, and distractions for marketers. Emina shares why new opportunities should be approached with curiosity but also tested against real business needs. She also explains why strong foundations, a broader understanding of marketing, and closer collaboration across teams still matter as search continues to change.
What you’ll learn in this session
Why SEO and AI create information overload: developments emerge faster than professionals can follow, official information is difficult to navigate, and regulation is only beginning to catch up and encourage greater transparency.
How to prioritize new AI search opportunities: assess their potential impact, start with small tests, and make sure you always implement or fix the fundamentals first.
What makes e-commerce ready for agentic search: aligning website HTML, schema markup, and product feeds, the “unholy trinity” of e-commerce search, before adding autonomous agents.
Why SEO and PPC should co-own Merchant Center: collaboration helps identify crawling issues, align landing pages with product feeds, and prevent disapprovals caused by mismatched data or outdated images.
How to design e-commerce experiences for users and LLMs: combine filters, product finders, application pages, and rich visuals to guide different audiences and surface the right products.
What generalist marketing brings to SEO: instead of hyper-specialization, marketing principles provide a broader view of audiences, business KPIs, and cross-channel search journeys that can make specialists better at their work.
Why brands shouldn’t fear direct competitor comparisons: clearly showing how your product differs and addressing the gaps users care about helps AI platforms understand when to recommend you.
How to understand audiences beyond traditional personas: combine internal and external data, sentiment analysis, and jobs to be done to uncover what users actually need.
Join us to explore the opportunities in AI search through the eyes of a marketer who is, in her own words, “too curious to specialize.”
Topics covered: AI search · agentic search · e-commerce SEO · structured data · product feeds · Merchant Center · faceted navigation · visual search · zero-click search · brand visibility · competitor comparisons · generalist marketing · audience research · cross-team collaboration
About the Guest
Emina Demiri-Watson
Head of Digital Marketing at Vixen Digital
Emina has more than 15 years of experience in digital marketing across SEO, paid media, social media, email, automation, analytics, and AI search. Her work connects technical, analytical, and creative marketing disciplines.
At Vixen Digital, she has been leading digital marketing strategies since 2021 across SEO, paid and organic social media, email, and HubSpot automation, with a strong focus on data-informed performance. Vixen Digital is a Brighton-based e-commerce marketing agency that combines SEO and paid media with analytics.
Emina writes about SEO and digital marketing and contributes to industry publications and events. She also co-hosted SEOs Getting Coffee with Sean Carroll, a podcast featuring practical conversations with search professionals.
Video Chapters
Transcript
Full conversation between Gianluca Fiorelli and Emina Demiri-Watson.
Gianluca Fiorelli: Hi, I'm Gianluca Fiorelli. Welcome back to The Search Session. Today we are going to have a guest with Eastern European roots but living in the UK. She defines herself on LinkedIn as “too curious to specialize,” and this is something that we are going to talk about. She is the Head of Digital Marketing at Vixen Digital. She touches practically all the things, from SEO to AI search, but also PPC, CRM, HubSpot, Xero, and so on. But she has a special love for search. Our guest today is, I hope I pronounce it correctly, Emina Demiri. Hey, Emina. How are you doing?
Emina Demiri-Watson: It's a great start. I forgot to unmute my mic. Hi, Gianluca. It's great to be here. Yes, I'm good. I'm good. How are you?
Gianluca Fiorelli: I'm fine. Fighting with the heat because it's still very hot here in Spain, but okay. We have to live with it.
Emina Demiri-Watson: Yes.
Gianluca Fiorelli: And ready to start this. I don't want to be like those guys who are already sharing on Instagram that Carey is defrosting, but it's only four months till Christmas. From September to the end of the year, these four months usually run very, very quickly. For instance, this episode is going to be broadcast live in early October. So surely the last quarter of the year is going to run for everybody, and everybody will start saying, “Where is 2026? It was just yesterday.”
But okay, I'm fine. And talking about you, I have a classic question I ask all my guests: How is SEO treating Emina Demiri-Watson lately?
Emina Demiri-Watson: Well, I think it's treating me well. I'm still in it, so it must be doing something right. I've just come back from holiday, so I've been out of it for a few weeks, and I kind of have this niggling expectation that I've missed way too much. In SEO, if you're away for three weeks, it's a very, very long time. A lot of things have changed. So I've been trying to catch up, and as soon as you catch up, you're already out of date.
SEO has been, as it has for the last couple of years, a little bit noisy, so it can be a bit tricky to see the woods from the trees. But it is still good. The clients are doing well, which is always a good thing to see, and yes, happy so far. Excited to get back into it, really.
Gianluca Fiorelli: Cool, cool. That's wonderful news from you. So yes, it is very messy, maybe sometimes in terms of information overload.
Emina Demiri-Watson: Yes.
Gianluca Fiorelli: It's the biggest problem today, also because, thankfully, there are people like Mark Williams-Cook, Aleyda Solís, and others with wonderful newsletters who are doing great curation. But sometimes you want to understand things directly from the sources, and some sources are very hard to navigate.
For instance, we always blame Google. It would be great if OpenAI had the same clarity and the same developer domain, with all the information about OpenAI presented clearly and in an organized way.
Maybe the latest decision of the European Commission will oblige them to do something like this, having been recognized as a large search engine.
Emina Demiri-Watson: Yes, that's it. I think it's going to change, and it already is, but it's not... You know how it is with technology. It moves super fast. Regulation is a very slow beast, and it's this kind of cycle of technology that we're in where only now regulation is starting to catch up. That's going to change, I think, how some of the organizations operate, and for the better. I think for the better.
Gianluca Fiorelli: Yes, I think so. It wasn't so different in the early years of Google, for instance. We didn't have a webmaster tool. We needed to wait for it. All the developer documentation came later and is still getting better now, years after it was published.
How to Prioritize New AI Search Opportunities
Gianluca Fiorelli: Talking about information overload, there is also opportunity overload. I think the two things are connected because we don't have clear guidelines about the things that can be done, as we are in a very mobile-sand situation for everything related to AI search.
Probably there is also a correlation with what we prioritize and what kind of intervention we can prioritize. What is your opinion? Were you able to find a reasoning framework in order not to get lost in this ever-changing landscape?
Emina Demiri-Watson: Shiny things syndrome. It's hard, particularly for me personally. I'm extremely curious, so whenever I see something new come up, I immediately want to try it and immediately want to test it. There's nothing wrong with that. I think that's a really, really good attitude to have.
But there's a difference between doing that and then doing that for clients. When you're doing client work, you need to be a little bit more strategic and not just go down every rabbit hole you would like to go down.
So what we usually do is the usual stuff. You try to estimate what the impact might be of the changes to your roadmap. We do our roadmaps three months in advance, and it starts from the data. We analyze the data and the performance, so a lot of the ideas and the roadmap come from what has been working and what hasn't been working previously.
Then we have a very long list. We usually start with a very, very long list, and then it's about thinking about what to prioritize based on the impact. In GEO, AI search, or whatever you may call it, it's even more difficult to estimate things because it's quite new. For SEO, technical SEO has always been a little bit tricky to forecast.
But we do our best in these mushy areas, like technical SEO and AI search. It also depends on what the change is and how long you're going to be running it for. Sometimes you can run a very quick test to see what the initial impact of a change is going to be and then turn it into a bigger project.
One of the things we do is try to find an MVP that we can run and then maybe turn it into a full-on project. Sometimes it's about seeing the woods from the trees. We've all received emails that say, “What are you doing about AI search?” Or a C-suite member has read an article and seen a new standard coming out, and now they're very keen on pushing the entire roadmap toward AI readiness.
But actually, they should be dealing with some of the fundamentals that are still wrong. So there are a few things we do. We always look at the fundamentals first, not just the fundamentals of SEO, but the fundamentals of marketing. When that big list comes to my desk, if I know that the client is missing some fundamental things, those are the priorities.
There's a bit of education and upward management that you need to do with clients to explain why you are not implementing an ARD standard. Instead, you're doing some of the fundamentals, like schema markup, for example. If the website doesn't have that, if the website doesn't have a brand, or if you haven't done the fundamentals right, I will always push back on some of the new standards unless there is a very clear business reason why that would be the right choice for you.
Up to now, I personally haven't had clients where I thought, “Okay, I'm going to implement this new standard rather than do some of the foundational stuff.” Maybe it's the size of our clients. I went a little bit around, but does that answer your question? So it really is using your experience.
Gianluca Fiorelli: Yes, everything is clear. Substantially, you are saying, “Okay, try to calm down the enthusiasm of clients and say, ‘Let's first fix all the fundamental things because we can create the most wonderful semantic, entity-centered, search-focused content, but if it's not crawlable or parsable, it is time wasted.’”
In fact, I figure that this can be substantially summarized with a definition that I found you use on social media: tools are just tools.
Emina Demiri-Watson: Exactly.
Gianluca Fiorelli: So, depending on what you need, you will decide whether to proceed or not proceed. But I know that you are a very curious person, so forget the Emina who is talking with a client and trying to calm herself down. Of all these things that are popping up like mushrooms after rain, what is making you feel more intrigued? Agentic search, meaning agents that are able to do things for humans who need something, or AI search, which is somehow in the middle? What is it?
The “Unholy Trinity” of E-commerce Search
Emina Demiri-Watson: We've been doing a lot in e-commerce as an agency. We've moved into the e-commerce space. So recently, a lot of my work, and what I'm really excited about, is this idea of the disambiguation of data, structured data in particular, but not just that. Schema markup, but looking at this interplay between your schema markup, your product feed, and your website.
I call it the unholy trinity of search in e-commerce because you have the really messy website HTML that is so difficult to standardize. That's not your structured data. Then you have your schema markup, which is your structured data. But again, some of the schema markup is actually being used by Google, and some of it is not. It's always changing and developing.
Then you have the feed, and the feed is, again, structured data. Then you look at how those three unholy trinity layers connect. Do they match? Do they mismatch? What happens when they don't match?
Recently, I've been digging a lot into e-commerce. I think that's the thing with agentic search. The part of me that is the excited marketer, the really curious part, always wants to dig into that first. But I always have to take a step back because the whole idea of agentic search also scares me a little bit.
The web is a messy place, so putting something autonomous into a messy space can get really messy. While I'm super excited about agentic search, and I think it's coming and you do need to prepare for it, it's not one of the areas where I'm actively, constantly working on it with our clients because I'm more interested in, as we were saying, the fundamentals.
For me, the fundamentals are this unholy trinity that you have to have. That layer is what you have to have. If you don't have that layer, then putting agentic on top of it is not really going to end well, if that makes sense.
Gianluca Fiorelli: Yes, that makes sense. But I was wondering if you are already planning to test a sort of pilot, for instance, for UCP in e-commerce.
Emina Demiri-Watson: No, not with any of the clients. My clients are just not ready for it. It might be the size of the clients. I've been talking to Jarno, actually. He's absolutely amazing, and he was saying a similar thing. A lot of it is on paper. It looks like it's developing, but that's the thing: it's still developing.
So it's very difficult, particularly if you're coming into an infrastructure, even for enterprise companies, where it's already set, and you're coming in with a new standard. It can be tricky. Unfortunately, none of my clients are at the stage where I can do that. I can only dream about that, Gianluca.
Gianluca Fiorelli: But the unholy trinity that you are talking about is an interesting take because, from the very beginning, although now it's a little bit better, I remember latency between the product structured data in the HTML and the feed containing the same data. It was very common to have mismatches.
Emina Demiri-Watson: Oh, yes.
Gianluca Fiorelli: And when there is a mismatch, the shopping experience, the merchant experience, is substantially crap.
Emina Demiri-Watson: Yes.
Gianluca Fiorelli: You are not going to achieve that dream excellent status, that excellent store status that is also giving you a boost in Merchant and organic Merchant search results.
Why SEO and PPC Should Co-Own Merchant Center
Emina Demiri-Watson: But what you're saying is interesting because it's something that brings you into this idea of cross-team co-ownership and collaboration as well. A lot of the SEOs I know don't really work with Merchant Center, which, from the get-go, is a big mistake.
You go into Google Search Console and use it for shopping because it's kind of a free listing, whereas Merchant Center is absolutely essential as a tool for any SEO. I was very surprised when I found out that there are SEOs who don't really log in. But I could understand it because, traditionally, it's always thought to be a tool that the paid team uses because they're running the paid ads.
Gianluca Fiorelli: Yes.
Emina Demiri-Watson: I can understand it. I don't agree with it, but that's the movement we need to see.
Gianluca Fiorelli: The paid team, the PPC team, owns Merchant Center.
Emina Demiri-Watson: Yes, exactly.
Gianluca Fiorelli: At least there should be co-ownership.
Emina Demiri-Watson: Yes, exactly. When you think about some of the disapprovals, this idea of your landing page and your feed not matching and your product being disapproved is not new. That has been around for a very, very long time.
You can see a flag. There's a difference between a paid person who looks at it and just says to a client, “Look, these products are disapproved because of the landing page,” but doesn't understand why or what's going on, and co-owning it, where that information is shared with the SEO team. They can tell you why those pages are not being crawled, for example.
Or you receive a client email saying, “Why are we running this ad with an image that is really, really old?” That's one of the questions you get all the time, and it's very difficult for a paid team. They're going to look within the platforms they use. They'll look in Google Ads, see what's coming in for the product card, and look at the website.
But sometimes it's an image that is still indexed because somebody didn't remove it from the server, and that's not something a paid team will check. We've had instances where the process of managing old, outdated product images had a knock-on effect on Merchant Center and product disapprovals because the internal team wasn't removing them from the server.
They would just remove the images from the product listing or product page, but they were still there. So you had all these products with really old images appearing in Google Ads, and we were actually putting budgets behind this. So yes, it's tricky.
Gianluca Fiorelli: Google Merchant has changed the policy for images now, so that kind of situation is going to be even worse. I'm talking about old images with embedded text inside. There is a difference. We are not talking about text that is on the package because that is part of the product.
Emina Demiri-Watson: Yes.
Gianluca Fiorelli: But the classic, very early 2000s-style image where you have an offer or “50% discount” all over the image.
Emina Demiri-Watson: Overlay.
Gianluca Fiorelli: You have to explain that it doesn't work like that anymore and that the product is going to be disapproved because of the Merchant Center guidelines.
Designing Faceted Navigation and Rich Media for Users and LLMs
Gianluca Fiorelli: Talking about e-commerce, I want to talk about UX navigation, faceted filters, and parametric finders.
Emina Demiri-Watson: Oh, yes.
Gianluca Fiorelli: They were already a problem in classic SEO: when to open a facet, when not to open one, how to deal with canonicalization, and the problems with Google accepting our canonicalization rules.
But you recently also talked about what this can mean, in a negative sense, for product recommendations by LLMs. If the fundamentals of faceted navigation are not well controlled and governed, we are probably going to have LLMs suggesting a product through a facet that is not relevant to the buyer.
You are visible on ChatGPT, AI Mode, or AI Overviews, but we already know that people don't click on these kinds of interfaces. If there is also this kind of misalignment, even the very small opportunity for a click is going to disappear completely.
What are your thoughts about this? Can you explain your idea about this kind of problem? It is not discussed very much.
Emina Demiri-Watson: You're talking about products that are surfaced in chats, obscure ones that you might not have a category page for. Tell me a little bit more, because when I think about faceted navigation and chats, I always think about chats on your website that help users.
Gianluca Fiorelli: Let's maybe reframe it. It's like designing faceted navigation and category pages for complex user bases, for instance, engineers versus trade contractors.
When you make this kind of distinction based on the personas or people you want to target, it can sometimes lead LLMs to hallucinate on high-intent e-commerce decision pages. These hallucinations may eat the click, especially on ChatGPT, where the clicks can sometimes be interesting. Obviously, this can hurt conversions because people are going to land on a page that means nothing to them.
Emina Demiri-Watson: Yes. I think it all begins at the beginning, which is being very clear about the user journey on the website. Who are the users and personas visiting your site, and how are they using your faceted navigation to help them do what they want to do?
As an example, we have an engineering client whose products are really complex. On the one hand, their audience includes people coming there to change a particular part. They sell flanges and low-loss headers. I'm not going to bore you, although it's really exciting to finally understand what the client does properly.
They have users who come in and just want to buy a part for their particular use case. They know exactly what part they're looking for, and their quickest entry is either the site search or perhaps the category page for that group and its filters.
That is a very different experience from somebody who comes in as your “man in a van” and knows that they're doing a specific job but isn't sure what they actually need. They might only know the category. They don't know the size or the pressure. They know absolutely nothing except the category and what their job is, like jobs to be done.
They might come to a category page with lots of filters. If you put all of those filters into that category page, you are going to alienate the users who came to simply change a part. If you're not very specific about the application and don't guide the user, you're going to alienate the man in a van.
So that one category page needs to serve two audiences, and all of that comes into wireframing and mapping the user journey. For this particular client, they had two very distinct audiences, and those category pages had two jobs.
One was to serve the main intent, which was, “Let's whittle it down to the product that I need.” But it was also to educate the man in a van and provide escape hatches on the category pages for those who might not know exactly what they need.
That can be done through a finder developed for that audience in particular. But what I also find really interesting is that sometimes it's about funneling them down to application pages first.
It's really simple. You put a module at the top of the category page that can say, “Not sure what you need? Are you doing this for a heating system? What system are you using?” They can escape from that product page into an application page that also has some of those products as a module.
That is useful for an LLM as well because you're not starting with just a category page and a bunch of filters. You're starting with rich information about an application of the products you're selling, and all of that is really good content for an LLM to learn from.
If somebody, such as a man in a van, comes to ChatGPT and says, “I'm changing a boiler for this and this. What kind of low-loss header do I need for this?” they can explain what they need in quite a lot of detail. We know that the queries are really long for LLMs.
You're going to surface if you have that information on your page. That doesn't necessarily come from your filters. It comes from the application pages and how well they are structured around users' needs. Does that answer your question?
Gianluca Fiorelli: Yes, but I'm wondering, isn't this the case I always use as an example? A plumber website, a plumber marketplace, or the classic situation where I go to a plumber and bring the things I have to buy with me because I don't even know their names.
For a client, couldn't you consider using visual search inside the website for these kinds of searches? Or optimizing the images on these application pages so they can be used in AI search?
Emina Demiri-Watson: With images, this is an additional thing you absolutely have to think about, particularly if you have products whose applications are different and quite complex.
What we've done so far is use a lot of schematics. Based on the application, you create images that are more like infographics than product images. And that’s part of it. You won't necessarily put them on your product detail page. That's not the image you're going to use in Google Ads or your feed. But it is something you can integrate into the application pages themselves. The more rich media content you can provide, the better.
Image search on LLMs is one of those things on my list to dig into. I know Myriam Jessier, for example, is absolutely amazing when it comes to image search. I have about five of her pieces that I still haven't read.
Gianluca Fiorelli: And my guide.
Emina Demiri-Watson: You wrote a guide, didn't you? There you go. You're going to the top of my list now, Gianluca.
I need to look at it because, in my mind, what we already know about optimizing images for SEO provides the foundations. But how can you make those images more LLM-friendly beyond the markup in the back end and things like that? I need to dig into it more.
The Generalist Renaissance: Why Broad Marketing Knowledge Matters
Gianluca Fiorelli: Yes. I want to change the topic a little. When I introduced you, and as I said earlier in the conversation, you are a very curious person.
Emina Demiri-Watson: Yes.
Gianluca Fiorelli: You really like jumping from one topic to another, and you have multiple interests. Somehow, this is reflected in one of your beliefs, and let’s call it the “generalist renaissance.”
For a long time, especially in recent years, there has been a strong suggestion for all SEOs that you have to know SEO in general but then specialize deeply in one specific field. It could be international SEO, for instance, but it can be JavaScript SEO or another area of SEO where you substantially hyper-specialize.
Now you say that the current situation suggests we should stop, look back, and return to a purer T-shaped marketing figure.
So, knowing marketing well, because many people who claim to “know” marketing don't really understand its fundamental principles. They may substantially be disguised as technical SEOs, and I say this with all my respect, but they don't know much about pure marketing or what drives a business to do one thing rather than another.
You suggest that people should return to this broader knowledge as a way to become even better specialists because they can better see the connections and correlations within their specialty.
Do you have a specific case or anecdote about a situation where having this broad knowledge, together with strong specialist knowledge, helped you find a solution for a client?
Emina Demiri-Watson: In general, a lot of my work has that at its core. One of the things I'm always on about, and which I think is really important and comes from the generalist marketing approach, is having a shared audience brief for anything you're doing, no matter the channel. Again, it goes back to what we were discussing about the fundamentals. So there’s that bit.
Then there are the KPIs. If you're working with KPIs related to business impact, such as revenue, you have no choice but to go wide. These days, the user journey includes so many touchpoints on the way to a purchase. I always say “these days” as if it happened today, but it didn't. It has been like that for a very, very long time.
It is no longer the case that somebody simply Googles you, visits your website, and buys something. They will scrutinize you across every single surface.
Talking about an anecdote, the other day, because we were moving house, I was buying a hanger for my hallway. It's just a hanger. We're talking about four hooks. I was sitting there reading reviews on Amazon, looking beyond Amazon at other hangers in the product grids, and searching for specific brands. I thought, “Maybe I should get it from John Lewis because John Lewis is a good brand.”
We're talking about a product that doesn't even have a very high price point. Obviously, maybe not everybody is like me, and they won't obsess so much about buying a hanger. But many people will no longer just choose the first thing they see and buy it. They will do their due diligence and look at several sources.
You can't account for that if you're very siloed into one channel. If you're not looking at how the brand and its products appear across all the channels, you're missing a very big part of the picture.
You're also missing a big part of the picture if you think search only means organic search on Google because that's not how people search. We're talking about “search everywhere” now, and that includes marketplaces, social media, and all these other surfaces.
You can't just open Google Search Console and say, “Bob's your uncle. I don't need to do anything else.” Look at what Google has done by integrating social media into Google Search Console.
Gianluca Fiorelli: Or the local search.
Emina Demiri-Watson: Yes, this is what I mean. It's not a coincidence that things are moving in that direction. I think it starts with the KPIs and a shared audience brief. Really understanding your audience is extremely important. You need to understand not only who they are, but also how they feel. How do they feel about the topics you cover? How do they feel about your brand? How consistent is your brand image across all these surfaces?
Coming back to your question, throughout my career, I've always wondered whether I should specialize and focus on something specific. But every day, I've had instances where approaching the work through the audience, focusing on the KPIs that matter, and looking at the channels holistically rather than individually has produced good results. I don't know if that answers your question.
Rethinking the Search Session and Brand Recall Without the Click
Gianluca Fiorelli: Yes, totally. I think two questions substantially remain. The first is more of a consideration than a question. Maybe people should update their concept of a search session.
A search session is not just someone entering a query into Google and then clicking. This is also why I sometimes really like the concept of zero-click search, but at the same time, I don't really like it. It limits the situation to someone conducting a search and not immediately clicking a search result.
But if we consider the search session, even if we use the 30 minutes that Google defines by default as the duration of a session, people may end a query on Google, or move from query to query, visit YouTube and return to Google, or go to ChatGPT and then return to Google Search during those 30 minutes. That is the search session.
The click may still happen. It's not that there is no click. It may simply be postponed. Obviously, certain types of searches already worked like that with featured snippets or direct answers. If I ask what time it is in Bucharest because I don't remember the time-zone difference, Google gives me the answer. I don't need to visit a website.
But for other types of searches, the more common searches, people will probably click at some point during those 30 minutes of a search session.
Secondly, you were talking about silos. Even if people are now optimizing search visibility for Google and discovering Bing because Bing feeds ChatGPT, while Gemini feeds AI Mode and AI Overviews, don't you think that people still treat all these surfaces as isolated?
Emina Demiri-Watson: Yes, and I think...
Gianluca Fiorelli: They are siloing them without considering that, although they are different, they also have common denominators.
Emina Demiri-Watson: Yes, exactly. I agree completely. What you were saying about brand stickiness and brand impressions was interesting. It all comes back to that, and zero clicks have to do with it as well.
I'll give you an example. I always use this example. It is a little bit harder these days because the process of getting your brand out there is probably a little bit longer without the click than it was before.
Let's take HubSpot, for example. I knew about HubSpot long before I ever wanted a CRM. The first time I got a client onto HubSpot wasn't the first time I had heard about it. I knew about HubSpot because I had read quite a lot of their content.
What happens in a search environment where you are no longer clicking to see what HubSpot says about something? These days, if I were looking for those same articles, I would get an AI Overview or use AI Mode, for example. It might not say HubSpot.
So I'm not going to click, and I'm not going to see that HubSpot is the company behind all these resources. Does that affect me when I reach the point of deciding on a CRM? Because at that point, I might not know about HubSpot to the extent that I did when I was clicking through to the website. Does that make sense?
Gianluca Fiorelli: That makes sense.
Emina Demiri-Watson: I think that if you are a well-known brand and you're still producing content, not churning out AI slop but actually providing value, you can remain visible. HubSpot had valuable content not because they were regurgitating what everybody else said, but because they were producing really high-quality content.
Whether or not you now receive traffic from it, if you're producing quality content, your content is being used by LLMs. If you're very clear about who your brand is, who it's for, and all the other lovely things we discussed previously, then you will eventually enter that window when I'm actually buying the CRM. You will be mentioned.
If you've done all those things, even without the click, you will be mentioned. You'll be mentioned particularly if you work on owned media, as well as external citations and media you don't own.
Gianluca Fiorelli: Yes, and I think this is something people probably still don't understand. It's fine to use how visible your brand is in AI search as a metric, but the company your brand appears alongside is even more important.
Emina Demiri-Watson: Yes, exactly.
Gianluca Fiorelli: If I'm paired with another brand that is better known than mine, people will probably remember the better-known brand. Let's say I'm an obscure CRM startup that beats HubSpot on every feature, but we're not known.
If we appear alongside HubSpot, the reverberation of the HubSpot brand is going to obscure the smaller reverberation of this lesser-known CRM startup.
I don't know if you had the opportunity to see the results of Aleyda Solís's survey about how people are using AI and where they are conducting AI searches. I was surprised. I also commented that, from the very beginning, we have discussed how important it is to participate in Reddit and other user-generated content spaces in the right way. Obviously, I'm not advocating spamming Reddit.
You need to be present and visible where your audience is active so you can become important to that audience. This can also mean discovering the specific niche forums they use, identifying the real influencers for your audience, interacting with them, and creating co-marketing opportunities.
If we want to go even further, we can discover the Discord communities where people are discussing topics relevant to us and from which we can gain knowledge.
I think people have misunderstood all of this. The purpose is not simply to get a citation or co-occurrence on Reddit. It's about enhancing brand recognition so that when people see a new brand, they remember it because it was active in their forum or Discord community.
They recognize the brand in AI search, and maybe they will then search for HubSpot and buy from it. Some clicks and leads may come because they already know the brand. I think this is the real importance of digital PR, outreach, mention building, and all these kinds of activities.
Why Brands Shouldn’t Fear Head-to-Head Comparisons
Emina Demiri-Watson: I have always found it interesting that companies are reluctant to go head-to-head with their direct competitors. I can understand that reluctance.
You've worked in SEO for a long time, so you know those articles where you compare yourself with a competitor. If HubSpot goes head-to-head with Salesforce, there's no problem. But smaller firms might be more reluctant to go head-to-head with a competitor, even when they have a really good product.
There is this concern about proprietary data. SaaS is riddled with the question, “What the hell does the product even do?” You have SaaS websites where you don't even get to see the product because people are concerned about proprietary information. You're just shooting yourself in the foot with that.
With LLMs, this has become even more important because this content absolutely needs to be created. You can no longer have a piece of content that compares you with obscure competitors that are related but not entirely relevant simply because you don't want to go head-to-head with the big players or let somebody see your proprietary data.
Now you have to do that because many AI searches involve comparing tools and are very specific. For example, “These are the specific features I'm looking for in my CRM. Give me a list of five CRMs that have them.”
If that information is buried in your help subdomain, or worse, if it is nowhere to be found, and you don't have a head-to-head comparison with other competitors, you're simply not going to appear.
Gianluca Fiorelli: You're totally right. How can you communicate your USP as a SaaS, for instance, if you are not comparing it with, let's say, Hubspot, or Waikay vs. some bigger, hyper-funded new companies?
Emina Demiri-Watson: Yes. HubSpot didn't become HubSpot simply because it had great marketing and a large budget. It became HubSpot because it had a really good product and then marketed it properly. You have to have both. You can't have a shitty product. Can I say shitty? Sorry if I can't.
There's no amount of marketing that can sell a shitty product. You can put a lot of money into it, but the churn will be terrible. Maybe with legacy technology and things like that, but we're talking about current SaaS and a product that is really, really bad.
The caveat is that if you're targeting enterprise clients, you can have a very bad product, but as long as it integrates with the existing infrastructure stack, it may be easier to buy than a really good product. But then we're talking about features. Integrations are features you have to develop.
Coming back to what we were saying, this is why you can't just do SEO. Somebody working in SaaS SEO needs to understand where the market is going, what competitors have, and the features of both the product they're doing SEO for and the competing products. They need to discuss that on the website and work collaboratively with the product team to say, “Hang on a minute. We don't have this integration, but all of our competitors do.”
Somebody is going to come to ChatGPT and ask for this integration. We're not going to show up because we simply don't have it. What do you do then as an SEO? You have to go to the product team and say, “Can we please develop this? There's no point in optimizing if we don't even have the feature.” I don't know if that makes sense.
Audience Research Beyond Traditional Buyer Personas
Gianluca Fiorelli: Yes, totally. Taking inspiration from HubSpot, it was famous for popularizing the concept of buyer personas among digital marketers.
This is one of the many topics you like. But the classic buyer persona profile, the original one created by HubSpot and then copied at scale by many other tools and marketers, is based on demographic and sometimes psychographic data.
You say that this kind of buyer persona is not the right way to conduct audience research. What is the correct way to conduct good audience research that gives you the insights you need to build your strategy?
Emina Demiri-Watson: I always say that you have to look at external and internal data. Your internal data is the quantitative data you can find in GA4, for example. How are people already using your website?
Then you layer that with tools such as Hotjar or Clarity, or whatever you use for qualitative data, videos, and the rest of it. A really good way to do this is to bring your GA4 data into Hotjar. It's not even called Hotjar now. It's Contentsquare, but it's still Hotjar in my head. You bring in your events so you can segment your recordings. Otherwise, you're looking at thousands of recordings and can't segment the data to see the users who matter. So there's the data on the website.
Then there's the data you own that is hidden within your teams. This means talking to your sales and customer service teams and others across the company. It also means looking at recordings from demos. Particularly in SaaS, sales teams should always record their demos. There's so much information you can analyze, and this is where LLMs are really useful because they're very good at processing data at scale.
Then there is your external data, which is your digital footprint outside the website. This involves understanding how people search for you on Google. AlsoAsked is brilliant. It gives you People Also Ask questions and query trees so you can see the journey of a particular user. What is the next step? What is the next question? What's interesting is that it doesn't just show you what people are searching for. It also gives you an indication of their sentiment. We conducted sentiment analysis on queries for some of our clients to understand not just what they were asking, but how they felt about a particular topic. That is useful for content and website copy. So there's AlsoAsked and, obviously, social listening.
There's also data you can get through Reddit. Reddit has a great API. Connect to the API and start small by pulling 100 posts from a subreddit discussing the topic your company works in. Then you might conduct an entity analysis. That's what we did. We extracted entities from the most popular posts in a subreddit and layered sentiment analysis on top. This helped us understand which topics appeared repeatedly and what the audience was discussing.
Trustpilot is also really good, particularly if you're working in e-commerce and have product reviews. You can group the reviews, identify which entities appear consistently, and analyze their sentiment. You can discover a lot about your audience that way.
From Audience Insights to Jobs to Be Done
Emina Demiri-Watson: Coming back to your point about why understanding the wider context is important, we conducted a Trustpilot analysis for one of our clients. Many of the resulting actions were related to SEO in terms of content, but much of what we discovered concerned the wider business.
We could clearly see areas of the business that we couldn't change as SEOs because they required a business decision. For example, people were extremely unhappy with a particular delivery service for a certain group of products. That's an operational insight you take to the client. It has nothing to do with SEO, and I can't fix it.
So I think that's important. Before you create your persona, do your research and try to understand the audience. Obviously, there's the SEO side of things, such as Google Search Console, Semrush, and all that jazz. But then sit down and see what the data is telling you about who these users are.
It's not about Karen, who works as an HR manager. By all means, call her Karen and give her a job title. However, it's about understanding her online journey in relation to your brand and understanding not just her pain points, but what keeps her up at night.
A lot of the time, people list the usual pain points, but all of those pain points are actually attempts to sell the product. That's not what really keeps her up at night. It usually doesn't correspond with the data.
Gianluca Fiorelli: It's the solution people are searching for rather than the problem.
Emina Demiri-Watson: Yes, exactly. That's why I always find the idea of the jobs-to-be-done framework very interesting for SaaS. It's a very useful framework for anybody who works in SaaS because it immediately stops you from being so self-centered or brand-centered and only thinking about your brand.
It moves you into a space where you ask, “What are they trying to do? What is the audience trying to do?” It's a very useful framework. Hopefully that answers your question.
Explore SEO through a broader marketing lens with two other episodes of The Search Session:
Amanda King explains why SEO must reconnect with marketing fundamentals, business goals, and impact-driven priorities.
Lee Odden extends the conversation to audience intelligence, buyer journeys, cross-channel content, and long-term brand building.
Gianluca Fiorelli: Yes, totally. One hour already, so I think it's time to close our episode today. Emina, it was a really great conversation. Thank you for being my guest and a guest of Advanced Web Ranking for this episode of The Search Session.
Where can people meet you in the future, in terms of conferences, not necessarily as a speaker but also as an attendee? Where are you going?
Emina Demiri-Watson: I haven't been traveling much. My next event is a speaking engagement at the Search ‘n Stuff Antalya conference. I'll be talking a lot about e-commerce there. Are you coming, Gianluca? Are you going to join us?
Gianluca Fiorelli: No, unfortunately not, because October is quite a busy month for me. I will speak at SERP Conf. Rome and then SEO On Stage in Athens.
Emina Demiri-Watson: Maybe next year. So that's my next event. I'm probably going to be at BrightonSEO. I'm not speaking this time around, but I will be walking around. If you catch me in the hallway, please come and say hi.
You can always find me on LinkedIn as well. I live on LinkedIn, so just send me a direct message.
Gianluca Fiorelli: Thank you again. I hope to see you soon in real life.
Emina Demiri-Watson: Yes, we need to go dancing. Gianluca, you're my dancing partner.
Gianluca Fiorelli: No, no. Maybe not. Let's finish with this kind of situation.
Emina Demiri-Watson: It was brilliant. It was lovely talking to you.
Gianluca Fiorelli: Thanks to all of you for being our guests as well. Remember to subscribe to our channel, like this video, and ring the bell so you'll be notified when a new episode comes to YouTube or Spotify. Thank you, and bye-bye.
Podcast Host
Gianluca Fiorelli
With almost 20 years of experience in web marketing, Gianluca Fiorelli is a Strategic and International SEO Consultant who helps businesses improve their visibility and performance on organic search. Gianluca collaborated with clients from various industries and regions, such as Glassdoor, Idealista, Rastreator.com, Outsystems, Chess.com, SIXT Ride, Vegetables by Bayer, Visit California, Gamepix, James Edition and many others.
A very active member of the SEO community, Gianluca daily shares his insights and best practices on SEO, content, Search marketing strategy and the evolution of Search on social media channels such as X, Bluesky and LinkedIn and through the blog on his website: IloveSEO.net.
