SEO Post-AI: Build Value, Measure Lift | Ian Lurie
Welcome back to The Search Session podcast! I’m Gianluca Fiorelli, and today I’ve got a brand new episode for you, in which I have a great conversation with Ian Lurie, SEO pioneer, guest instructor at the University of Washington, and a self-proclaimed nerd who’s never afraid to dive deep into the details.
Here’s a snapshot of the ideas you’ll take away from this episode:
The difference between today’s young professionals and those of the past, from Ian’s university teaching experience.
Why chunking optimization isn’t new—it’s how humans naturally process information.
How Ian Lurie uses AI as a collaborator and not a tool for automation.
Quick wins with big impact: image compression and caching for strong results right away
Unless you’re a publisher, business metrics are what truly matter
Why the industry needs retraining: lift-based measurement is valid, important, and essential for marketing performance
Let’s get started.
Video Chapters
Transcript
Gianluca Fiorelli: Hi, I’m Gianluca Fiorelli, and welcome back to The Search Session! Today’s guest is a true expert in Dungeons & Dragons and just about every kind of fantasy board game. Honestly, I wouldn’t be surprised if National Geographic hired him for his deep knowledge of capybaras, too.
He’s a big nerd with a deep love for Battlestar Galactica, The Lord of the Rings, and especially, he loves SEO. He was the founder and CEO of Portent, which he eventually sold. These days, he’s a freelance consultant and—if I’m not mistaken—also teaches. Let me double-check… yes! He teaches at the University of Washington, the State, not D.C. Today’s guest is Ian Lurie! Hey Ian, how are you doing?
Ian Lurie: Hey Gianluca, good! How are you?
Gianluca Fiorelli: I’m good—just back from a much-needed break and vacation. Though I think you’ll agree with me on this: as freelance consultants, our vacations tend to be short, and honestly, we should be a bit more protected in that sense.
Even living in a country like Spain, where technically I am paying for my own vacation, it’s still tough. I’m not an employee, so I don’t get the same benefits or privileges that come with that.
Anyway, we’re recording this at the beginning of September, and the episode will probably go live toward the end of the month. So let’s say—we’re kicking off a new season of Search Session, and I’m really excited to have you as a guest.
And one thing that’s not going to change is my classic starter question for every guest. So, Ian, how is SEO treating you lately?
Ian Lurie: Well, it's not boring. I'll say that. What's the best way to put it? I alternate between being really excited and really pissed off on an almost day-to-day basis, which folks who know me might say, “Oh, that's just Ian!”. But it's, it's much more than it used to be.
I'm sure everyone's going through this. AI, reduced click volume, those are all things that are getting people really wound up. And it's exciting on the one hand because I get to help clients navigate this. And on the other hand, it can get really frustrating because people don't necessarily want to hear the long-term approach. And I'm sure you've encountered this too.
Handling AI Overviews & Client Confusion
Gianluca Fiorelli: Completely. In fact, as you were talking, I was thinking back to when I started this video podcast back in February. I totally get it—especially after the summer, when AI Overviews started showing up by default in nearly every Google search result. The level of confusion has been massive.
We could make sense of it, sure—but even now, months later, the confusion is still very much there. And while it’s already big within the SEO community, it’s even worse among business owners.
So how are you dealing with that confusion—especially when it comes to clients, potential clients, and business owners?
Ian Lurie: I haven’t figured it out yet. I’m being brutally honest here—and if you want to end the podcast right now, I’d totally understand!
But really, I haven’t figured it out—more than in previous times in SEO. And I do not accept the whole “SEO is dead” narrative. I never have. I think that’s just poor marketing from folks who want to acquire clients and feel like they need to say something that sparks controversy. Which I understand—but I don’t agree with that narrative.
I think steady teaching—explaining things to clients one piece at a time—and also being pretty honest with them, saying, “You know what? We don’t have it all figured out either.”
So, what I’ve been recommending to people—and the one thing I’ve seen that works—is to stick to the things that have always worked in marketing, right? There are certain things, and if you execute them well, it doesn’t matter what platform you’re on, what channel you’re marketing on—it doesn’t matter what’s happening with Google or OpenAI or anyone else.
You’re still going to build a dedicated audience, which will help your business grow. And that’s how I’ve dealt with this with most clients.
Gianluca Fiorelli: Yes, I think I’m doing almost the same thing—not at your level, maybe—but I remember you’ve always been a strong advocate for staying rooted in marketing, without getting too caught up in the specific subsets like search, social, or whatever.
And I think that’s the way—“This is the way!” as Mando would say—to approach things right now. In many ways, it feels like we’re back to being students, like we were 20 or 25 years ago, trying to understand everything from the ground up.
And we’re lucky now, because there are some very skilled—maybe even nerdier than us—people who really understand the technical side of AI and retrieval, and they’re digging deep into how AI search works. That’s something we didn’t have back in the day. We were all just experimenting, trying things out, and seeing what worked.
But regarding all the polemics around “SEO is dead,” GEO, and so on. It was just a few days ago when, suddenly, this job posting from OpenAI came out for a content strategist, and it revealed a few interesting things.
First, this content strategist must have strong SEO knowledge. Second, this person would work directly with OpenAI’s SEO team. Which means, yes—OpenAI has an SEO team. Not a GEO team, not an AIO team, not an LLMAO team… an actual SEO team.
So even if, more broadly, clients want to call it whatever they want—I’m not going to waste time and energy trying to correct them. No, it’s not GEO, it’s SEO. You can call it Mickey Mouse if you want. But at the end of the day, what I’m doing is SEO—applied to search and AI.
And I think—for our own mental sanity—that OpenAI’s announcement kind of closes the debate.
AI, Automation Hype & Critical Thinking
Gianluca Fiorelli: So, let’s move on. You were talking earlier about education. In your role as a professor at the university, what are you seeing in students today—maybe not just students, but younger professionals—that’s different from what you saw back when you were running Portent? Do you see some differences, or not really?
Ian Lurie: I’ve always liked teaching—even when I was at Portent. That was a big part of our culture. I wish I were a full professor; I’m just a guest instructor, by the way, which means I don’t get an office and I get the oldest classroom, with all the stuff that breaks all the time.
But what I see now with students is that they’re really grappling with the technology in ways that people coming on board at Portent weren’t. Because folks joining Portent already knew something about SEO. They already had some familiarity with digital. They came in with certain assumptions—they weren’t necessarily thinking about marketing when they first showed up.
Whereas my students—they take a course in marketing first. Then, the very next semester, they take my course on digital marketing. So in some ways, they actually have a better grounding in traditional marketing than a lot of people I’ve seen in the industry—not just at Portent. They have a better and more recent awareness of traditional marketing.
And most of them are entrepreneurs, right? They’re looking to start their own businesses fairly soon after graduating. So the students coming into my class are actually much more open-minded about some of the technology out there, and also some of the fundamentals. “How do you rank better in a search engine?” Well, we’re going to start with basic marketing concepts and then build up from there.
The other thing I see with students coming in now is that they’re very focused on trying to automate marketing—which wasn’t really a major topic, even as recently as two or three years ago. Back then, people weren’t seriously talking about using computers to completely automate, say, the content creation process in marketing, or fully automate how you process leads as they come into your business.
Yes, there was some automation—but not full automation. And now, a lot of students coming into my classes are working under the assumption that they can fully automate those things. So it’s kind of two sides of the coin. There’s a real strength, and also a real weakness.
The strength is that they often have a better grounding in traditional marketing. The weakness is that they’re unsure—they don’t have a good detector to understand when a technology can’t actually do what they’ve been told it can.
AI being the obvious one. All of my students come into class assuming that AI can generate all their content for them—whereas I think most people coming into an agency wouldn’t believe that. That was a really long answer—I hope it was helpful.
Gianluca Fiorelli: Oh no—that was a really good answer. And I think it touches on something that goes beyond just your students. It’s something we’re seeing more broadly in the so-called Gen Z generation.
I’ve even seen it with my own sons. Sincerely, I had to teach him, for example, how to properly use ChatGPT when doing research for school—and, more importantly, how not to blindly trust it.
And I remember, in one of the past episodes of The Search Session, the guest was Judith Lewis. She was talking about how, when it comes to highly technical topics, AI—or rather, machine learning and LLMs—can trick you. They can offer solutions that seem correct on the surface but could actually end up damaging a website instead of making it work better.
Gianluca Fiorelli: And it makes me think of something that’s becoming pretty common here in Spain. In schools, they’re starting to ban smartphones entirely—kids can’t use them during school hours.
Maybe we should try something similar—just for one day a week at work. Imagine if, for that one day, your access to Gemini, ChatGPT, or whatever tool you use was blocked by the agency’s server. You’d be forced to do everything manually, like we used to.
Because what happens if all the models go down—or there’s simply a blackout?
Ian Lurie: Yes, just think about how we navigate in our cars now—at least here in the U.S. You can’t get anywhere without GPS.
When I was younger, I managed to get around Seattle—places I’d never been before. I’d land at the airport, pick up a rental car, and find my way using this thing called a map. Now? I can barely get across town without a navigation app—Waze, or whatever else I’m using.
There are parts of our brains, parts of our day-to-day skills, that we’re losing because we’re using technology. That’s not always a bad thing. If I don’t have to remember directions to get somewhere—that’s a little bit of mental storage space freed up, right? I can use that for something else… like planning my next Dungeons & Dragons session.
And the same goes for ChatGPT. But yeah, it definitely impacts us. Urban studies are showing this now—it’s changing the way we think. Is it better? Is it worse? I don’t know. But we’re definitely losing a big chunk of our ability to process and solve problems. We’re relying more and more on this other technology. This is getting way outside of SEO—it’s turning into more of a philosophical conversation.
Gianluca Fiorelli: No, no—but it makes sense. It extends into everyday life, especially as AI becomes more and more integrated into our daily life.
I remember there was a movie—I don’t recall the title—but it was about someone who got locked inside a house he was trying to rob. Everything was automated—so when the tech broke, all the doors locked, and he was trapped. I wish I could remember the name.
Ian Lurie: I think it was called Escape Room or something like that… or no?
Gianluca Fiorelli: No, not Escape Room. The one with Jodie Foster—she was locked in the panic room with her daughter when the robbers came in.
Ian Lurie: Oh, right.
Gianluca Fiorelli: The one I’m thinking of was something else. I think it was with Willem Dafoe. Let’s leave this as homework for you, dear listeners.
Chunking, Passage Ranking & Writing for Humans
Gianluca Fiorelli: And speaking of which—you used the word chunk earlier. So, here’s a very big jump, a leap. What do you think of chunking optimization?
Ian Lurie: I think chunking optimization is something we should have always been doing, because it’s how humans process information.
I actually don’t think information retrieval systems process content any differently now—at least when it comes to handling chunks of information—compared to the pre-AI era. I don’t see chunking as a new strategy. Presenting information in digestible pieces of content—in chunks—has always been important.
David Ogilvy was talking about this way back in the Mad Men era of marketing—the ’60s, ’70s, ’80s. It was always about breaking up content into easily processed chunks of information. That’s nothing new.
So this idea that chunking is a brand-new optimization tactic? If calling it that gets people to finally do it, then fine—let’s call it chunking, let’s pretend it’s something new. But really, it’s not. It’s just a new way to talk about an existing concept.
It’s not a major change in how search—or any kind of discoverability—works.
Gianluca Fiorelli: Yes, I agree. And if we think just in terms of information retrieval and how Google works—well, passage ranking was announced a few years ago already.
That’s how, for example, Google is able to generate features like People Also Ask, featured snippets, and so on.
In fact, with featured snippets, when you click on one and it jumps directly to the highlighted section on the page—that's essentially the same behavior you see with AI Overviews when you click and get taken straight to the selected part of the content.
So yes, I think this whole “chunking” idea is useful—especially if it finally pushes some technical writers or copywriters to realize they don’t need to be James Joyce when they’re writing something. Because sometimes the real challenge lies in the tendency to get a little too creative.
So this shift back toward clarity, toward being synthetic, using things like the inverted pyramid structure—it may not always feel like the most elegant style, at least not for me—but it is effective for writing on the web. Short paragraphs, each one self-contained and conclusive, and so on… it works.
Ian Lurie: Right. And you know, when Google introduced passage indexing—leaving aside their profit motives, how they treat the industry, and how they handle information—what they were really trying to do was emulate the way people think, in a way that makes them more money.
People have always thought in terms of passage and content. They developed passage indexing to deliver a user experience that’s closer to how we naturally absorb information. It’s not like Google invented a new way of processing the information. They’re creating a model of how humans process information. So this whole idea of chunking and passage indexing—that is the way we’ve handled information for a long time.
And understand, I’m not skeptical of it. I think it’s great that people are doing more marketing and they’re exercising their creativity in ways that better conform to chunking. I think that’s awesome.
I just don’t want people to think that it’s a revolution. And the only reason I don’t want them to think that is because then, when the next thing comes along, they’re gonna stop doing it. And I don’t want people to stop doing chunking.
I would love to see more people doing it—and doing it more consistently—and not even thinking about it in terms of search and discoverability. But if that’s what it takes to get people to do it? Great. Chunking is a new thing. It’s cool. It’s awesome. Everybody should do it.
Gianluca Fiorelli: Yes. It’s like when there was that definition I really didn’t like at all - SEO copywriting - as if it were just a set of rules that copywriters or technical writers had to follow.
I was always pushing them: “Please, learn the rules—but then forget them. Because if not, you’re going to end up creating the most awful content on Earth.”
Ian Lurie: Yes. ChatGPT—well, OpenAI hiring a content strategist should tell us that ChatGPT cannot generate content.
Gianluca Fiorelli: Totally.
Ian Lurie: And that’s because there is no fixed set of rules. ChatGPT is applying a set of rules. We have to create content based on what we know will appeal to our audience. Content that will answer their questions—whether those questions are explicit or implicit—when they come to see our content—read it, listen to it, watch it, whatever.
I’ve always hated the terms SEO copywriting and SEO content. The moment I hear it, I know exactly what I’m going to see—and it’s not going to be good.
Gianluca Fiorelli: Yes, and this also makes me think of something else. It was actually even before COVID when the news came out that Google had started hiring for more humanistic roles—like linguists, people with backgrounds in linguistics, sociology, anthropology, and philosophical languages, and all this kind of knowledge.
It reminded me that SEO has always been a very technical marketing combination. But I suspect, and without taking anything away from the importance of technical SEO, that there’s been this trend where technical SEO always focuses on the stage. Even Googlers, when they talk to SEOs, often focus mostly on the technical side. When it comes to content, the advice we get is usually just: “Write good content.” And that’s it. Nothing more.
So don’t you think—maybe this could be a good recommendation, not just for the younger generation, but also for people who might be struggling right now—that instead of going too crazy into trying to understand how LLMs work—which is totally fine, of course—we also make sure we don’t forget how our language works? Don’t you agree?
Ian Lurie: Technical SEO makes you discoverable and ensures you're visible. It does not ensure you're valuable, right?
That’s where creativity, content, and understanding your audience come in—knowing the answers you need to give them. Understanding what will make a piece of content—whatever modality you're using—appealing to that user. That’s what really makes the difference.
Technical SEO—look, half of what I do is technical SEO. It’s very important. But it’s just about ensuring discoverability. It doesn’t build value. Building value is a whole other game. And you have to understand both if you’re really going to be a good SEO—never mind a good marketer.
AI as Creative Partner, Not Automator
Gianluca Fiorelli: Totally agree. And earlier we were—well, not exactly joking—but proposing that idea of having one day a week without AI at work. So in your day-to-day, how do you use AI to support you?
Ian Lurie: I don’t use AI as much as other people might—I don’t know. But I use AI by collaborating with it. I don’t use it to automate anything. There was a great article in The Atlantic—and if anyone listening wants the link, I’ll happily send it to you. It was about using AI. I think the title was something like A Better Way to Think About AI.
And the authors—who are from Google and MIT—talked about how AI is not a good automator. It’s a good collaborator. So if you can’t make good decisions without AI, you’re not going to make good decisions with it.
I generally use AI to critique my own ideas, to follow up, and to brainstorm. I use it to help with some technical aspects of what I do. Sometimes I’ll use it to help me rewrite things. I mean, anyone who uses Grammarly is already using a kind of AI. What I don’t do is try to automate anything. I don’t automate workflows.
I use AI to consolidate data. I use it to help with visualizations, sometimes. But what I’m not going to do is have it automate something like analysis.
I’m not going to say to the AI, “Please give me 10 SEO recommendations for this page,” because the best it can do is the average. That’s how AI works. So I’m not going to use it to automate. And again—long version of a very short answer: I use it as a collaborative tool, not an automator.
Gianluca Fiorelli: You’re not alone. I also use it more as a companion, because working alone—okay, yes, we both know a lot of people in the industry—but we can’t always message our friends just to ask something.
So I use AI more like someone who helps me clarify my own ideas. I start chatting with it like:
“Okay, I have this in mind...” and then I present, for example, a checklist or a process I’m thinking through—always including the relevant documentation or context. Because without context, it can just invent things.
So I’ll ask something like, “Am I forgetting anything?”—you know, the classic little detail you overlook. And in that way, I find it’s a really useful companion.
Ian Lurie: Yes.
Gianluca Fiorelli: And then, obviously, for tasks I used to do almost manually—like merging very different types of structured data—AI is really helpful for those kinds of things. But you have to know what you're doing. Because many times I have to correct Gemini or ChatGPT: “You’re doing this wrong. This is not correct.”
Otherwise, you're just copying, pasting, and ingesting the output—then the developer says, “We did it,” and suddenly Google Search Console throws an error. So yes—use it for assessing workflows, not for creating them.
Ian Lurie: Yes, that’s a great way to put it. And also—to fill in parts of a workflow where the computer might actually be better than I am. Right? Because once I understand the workflow, or if I trust the information, it might be better at analyzing that information. Or it may just be different.
I used to have an agency, right? I had 50 people who would tell me if I was doing something wrong—or if they thought I was doing something wrong. AI isn’t terrible at taking a look at your assumptions and potentially challenging them—if you ask it to.
Gianluca Fiorelli: And you know, another way I find it interesting to collaborate with AI is when I have some crazy idea. I’m sure you also have these kinds of really lateral thoughts—where you’re connecting, I don’t know, elephants with lions—and you just want to see what comes out.
So I’ll propose the idea to AI and try to simulate it. It’s a way to understand: am I crazy, or not? And I think that might be one of the most creative ways to use AI.
Ian Lurie: Yes. We were talking about Dungeons & Dragons…
Gianluca Fiorelli: Yes.
Ian Lurie: ...and one of the first ways I used AI—and if anyone from my gaming group that I’m DM hears this, just… understand, this is what I use it for—sometimes I’ll describe what I’m planning for that night’s session. And then I’ll ask it a few questions, like: “Am I railroading my players?” Which, for non–Dungeons & Dragons players, means: “Am I taking away so many choices that they’re not going to enjoy the game?” I’m trying to send them toward a specific challenge, sure—but am I making it so artificial that they’re not going to like it because they’re being forced down this path?
So I’ll ask it, “Am I railroading my players?” or: “What would you do to keep things moving right here in this part of the game?” I’ll use it for things like that, but only because I’ve been running Dungeons & Dragons games for 45 years? I don’t recommend someone who just started last week to start doing that. But it is a great collaborator—for things like that—if you thoroughly understand the subject you’re collaborating on.
A D&D Party for Digital Marketing
Gianluca Fiorelli: Yes, totally. And speaking of Dungeons & Dragons… let’s see: if you could build the perfect team—the perfect digital marketing team—who would they be in a D&D party? What roles would they have? And who’s the boss they have to fight?
Ian Lurie: So… you mean like real people?
Gianluca Fiorelli: No, let’s say figures, marketing figures without naming anyone.
Ian Lurie: Phew. I was really scared there for a second. So, for roles, I’d want one highly technical person. And what would you call that? That would be the wizard, right?
And ideally, I don’t want them casting Fireball into the middle of the party—if we can possibly avoid that—which they’ll definitely want to do. Like when you find a full JavaScript front-end rendering tool, right? And then someone decides they want to move the entire website into that paradigm—and it just destroys everything. Kills everybody. That’s basically the wizard casting Fireball right into the middle of the party. Half our audience is probably getting completely lost at this point.
I’d also need a super creative person. Someone who’s great at crafting a really appealing message. That would be my bard.
And I probably need someone sneaky—someone who can quietly look at what competitors are doing, what’s working out there, what’s not working. Maybe pick up on subtle hints that other brands—or Google, or even ChatGPT—are dropping about how they’re boosting discoverability or strengthening their brand. So that would probably be my rogue, right? My thief.
I’d also need someone who’s good at leading the party—someone who maybe combines all of those skills into one. You know, someone who has knowledge of all the areas to some extent. I think Rand Fishkin used to talk about the T-shaped knowledge set. Someone who knows a bit about everything, but also goes deep into one area, whatever that is. That person would usually be the cleric.
Because in Dungeons & Dragons, the cleric can take action—they can fight, but they can also cast spells, heal people, and rally the group around a particular plan. So I’d say that would be another key member of my party.
And then, the only other party member I’d want is someone who can make sure that—when push comes to shove—everything moves forward. I need someone who’s really good at execution, at making things happen. And that would be my fighter.
Actually, I ran a game last week that featured what I think is the big bad evil boss. It was a devil who was creating cheap imitations of magic items. They worked almost as well as the real versions—but not quite. And he was just creating one copy after another, spamming the entire playable world in my Dungeons & Dragons campaign with all this ridiculous gear. And everyone started believing that this gear would just do the job for them.
And I’m serious—I did not make this up just for today’s podcast. This was a real thing. Actually, the abbreviation for the organization that he headed up was T-Mo. So yeah—that was the big enemy. And of course, he was powerful, scary, and hurt the party a lot. But in the end, it took the wizard, the rogue…actually, they didn’t have a wizard, and they almost paid for it. But ideally, that would’ve been a wizard, a rogue, a cleric, a fighter, and a bard. That's what I would've used ideally, to take care of that person.
Gianluca Fiorelli: Cool, cool—yes. And, hmm… I mean, that big boss spamming the world with “almost magic” items? That feels very relatable to what’s happening right now with all this AI-generated content.
Which is feeding more AI—so I call it the mad cow syndrome of AI. AI eating itself. Generated content created by itself, and so on and on… until, eventually, it just explodes. Let’s see what happens.
Ian Lurie: Yes.
Site Performance Quick Wins
Gianluca Fiorelli: But, talking about something else now, because one of the topics you’ve always loved is performance. Website performance. It’s always been part of the classic JavaScript debate—and now, maybe we finally have a really strong argument: LLMs don’t render JavaScript. And when I say that to a client? They instantly go: “Okay. Let’s call the developers.”
Ian Lurie: All of a sudden, they’re paying attention.
Gianluca Fiorelli: Yes. But what do you think is the key area? The problem with performance is that it usually takes a huge amount of time and effort. And in a sprint-based philosophy, we’re always moving from one sprint to the next—push, push, push.
So ideally, if you had to suggest a checklist of priorities, what would be at the top? Something that’s relatively quick to implement but has a strong impact right from the beginning. Where would you start?
Ian Lurie: When I talk to clients about this, I try to make two things clear. First, that performance touches everything. It’s not just about search. They need to care about it because it impacts conversions, and it affects all the things that deliver value to their audience. And yes, it does help somewhat with search. It’s not a major ranking factor, but it is a factor.
So where do I start? What’s my top priority? Even today, I’d start with images. I’d start with image compression—because no one does it well. And, you know, people always roll their eyes when I bring it up. I tell them, “I’ll stop bugging you about image compression when you start doing image compression well.”
Then the next place I go is caching—how the site and the server are storing and delivering content. After that, I look at things like JavaScript complexity, how you’re delivering fonts, and other assets. But it starts with images and caching.
Gianluca Fiorelli: I totally agree. Images alone usually solve—I wouldn’t say 80%, maybe a bit less—but they make a big difference. At least enough to move you from red to orange. And this is somehow the average volume for websites and their performance. It’s not great, but it’s acceptable.
Ian Lurie: And I also know that, yes, JavaScript may be the bigger problem with a site’s performance - but if we can’t handle images, that client is already in trouble for performance. I'm not gonna be able to get them to do anything with JavaScript.
If a client can’t take a batch of five-megabyte images, compress them, and put them back on the site—or make sure Cloudflare is doing what it’s supposed to, or install Imagify - I know the chances of getting their engineering team to reassess and refactor the way they handle JavaScript are pretty much zero.
I wouldn’t give up on it, but it’s a long shot. So I usually start with images, just because it’s a good place to start. It's an easy win—and it can be a big win. I know we have to get them to the point where they can handle that and establish processes around it before we can even think about moving on to JavaScript. Maybe that's the teacher in me, and maybe I shouldn't think that way, but long experience tells me that's how it usually goes.
Metrics After Zero-Click
Gianluca Fiorelli: Yes, yes. And another thing I wanted to ask you—it's actually a common topic here on The Search Session—is about metrics. It's something we often talk about in the context of Google. But sometimes, I feel like we always put the magnifying glass on Google and forget that these issues are just as bad, if not worse, on ChatGPT and the other LLMs. At least with Google, even if it’s not the best tool anymore—Search Console—and Bing Webmaster Tools still provide the data.
But GPT? No. Perplexity? No. Claude? Never. The problem is, until last summer, there was at least some level of certainty. Like, okay—if we appear here, even with the rise of the zero-click phenomenon, we should still expect some kind of CTR and some kind of traffic.
Now, with AI Overviews, the links—tucked away in that side tab, in the source column—have a ridiculous 1% CTR. So traffic is shrinking, but impressions are rising, because suddenly, thousands of websites are showing up in position one, but with no clicks.
So there's this growing trend around "visibility." In that sense, impressions are becoming more important than clicks. And that might be valid in certain ways, but it’s really complicated to sell it to clients.
I know this might not be directly related, but you’ve talked about return on time invested, which I think could actually be a great metric. Something we could include in our overall metric package. But typically, how do you measure SEO success in a way that's understandable for your clients?
Ian Lurie: So, return on time invested is really about engagement. Clients don’t always understand engagement as a performance metric—which I think is fair. Saying, “This page got this many visitors, but they also stayed longer and converted more often,” can feel very abstract to a lot of clients.
What I try to get them thinking about and have always tried to because it doesn’t change—whether we end up with the alligator graph, zero-click results, increased impressions but decreased clicks, or something else—is the business metrics. Are you making more money right now or not? When the alligator graph started showing up—impressions going up, clicks going down—did your revenue go down? And when you made some changes to your site that we know increased share of voice but didn’t increase clicks—did revenue go up?
Unless you're a publisher—and publishers are in a world of hurt right now—I understand that, and I wish I had a great answer for them besides, Google is horrible, and they're taking your business away from you.
When it comes to other businesses, it's all about the metrics that directly affect the business, and for most businesses, that's money. So, is what you're doing bringing in more money? It’s a pretty simple question.
If you can compare changes you've made in your various marketing channels over time to changes in business growth and behavior, you can get a pretty good idea of what's working and what's not.
I think people got pretty spoiled over the last 20 years with the idea that they could track behavior directly to revenue. That’s the exception. That wasn’t true until around 2002, 2003. And now, we're starting to return to the reality that you can't always track things directly—but you can always track business performance.
You can’t always tell exactly which thing is making it work, but you can track when you made a change and when business performance changed.
Gianluca Fiorelli: Yes. In fact, I think it's this loss of attribution that's making it feel more painful for many people, but also, not really for the business owner, but maybe more for the CMO of the company.
Ian Lurie: Yes.
Gianluca Fiorelli: Because they’re lacking the data to say, “We were doing things right,” to confirm a strategy. But at the end of the day, for them, it should be the same. I mean, if the business metrics are good, that means the overall strategy is working. It's more problematic, for instance, for an SEO agency to say that—because how can you attribute at least a tiny part of the contribution of SEO to the success of a business metric? This is becoming really complex.
Ian Lurie: Yes, and, you know, it always was complex. And I think what everyone has to research and relearn—because it's been around for a long time—is the concept of lift-based measurement. That is how, up until the year 2000, we measured the success and performance of marketing.
And we have to train ourselves, and we have to train our clients back to that. And, you know, we did this to ourselves, and it's going to be difficult and painful. And businesses may suffer or agencies may suffer, and CMOs may suffer, but we have to retrain the industry around the idea that, yes, lift-based measurement is valid and important, and it's part of how you measure the performance of marketing.
Gianluca Fiorelli: One last question about marketing and SEO. It's a very simple question. There is something that happened in the past month that really surprised you in a good sense, happening in the search industry? Because I don't want to talk about the many things that may have surprised you in the negative sense. What did surprise you in the positive sense?
Ian Lurie: I talk a lot about the negative, but there are a lot of things I'm very positive about. The rise of a really strong teaching culture in the industry. That to me is really exciting. And, you know, you see people like Mike King and Lily Ray who are out there, and they're not just sort of holding onto their secrets and their information.
They're really out there explaining things and sharing knowledge. That’s tremendously important for our industry and for its longevity—to ensure that it keeps moving forward. That was a very pleasant surprise for me.
Another surprise is that clients are much more sophisticated about AI’s place in marketing than I expected. They're much more critical thinkers, and I'm not insulting the intelligence of my clients. I'm just saying, there's a deeper understanding of the implications of AI than I expected, because clients are dealing with so many other issues too, right? They're running businesses, managing everything else, but they seem to be more thoughtful skeptics about it.
So, I’d say there are two things: first, a really strong teaching culture in our industry—that’s absolutely huge. The quality of information and teaching out there now is even better than when we were starting in the industry.
And second, the way clients are approaching AI, ChatGPT, and all the changes in search. Sure, there’s some frustration, but I see many of them being more sophisticated about it than I would’ve thought.
Gianluca Fiorelli: Yeah. For instance, in my personal cases, I don’t see clients asking the same things they used to. Back in the day, it was always, “How can I rank? How can I be number one?”—that was the classic question. Now, very rarely do they ask that. Instead, they ask, “How can we be more visible?” And they try to explain to me why they want that visibility.
So they’re more detailed, and in a sense, even more helpful in what I need to do for them. We also see that many are experimenting with AI to figure things out for themselves, instead of treating it like some kind of black magic recipe from Google for ranking.
And about the high quality of the community—this is something we’ve always been very vocal about, and promoting all the great people who create wonderful content and share it with the community. You mentioned Mike, you mentioned Lily, I could also mention Aleyda and Dan Petrovic. Practically all the guests of The Search Session are that kind of person.
And I’m thinking of you too. When I remember you, I always feel you should write more — it’s a little critique I have for you. Not just because you’re a great writer in my opinion, but because you have a way of presenting ideas that are very digestible — and at the same time, really important. So, as a fan, I invite you to write more.
Ian Lurie: I appreciate.
The Fireside Questionnaire
Gianluca Fiorelli: Now, let’s talk about you. Let’s stop talking about SEO, AI, and all the rest. Let’s make a transition from AI to your own interests, to something more personal. You’re a lover of sci-fi, like me.
Ian Lurie: Yes
Gianluca Fiorelli: So, let’s imagine our AI future as if it were a serious sci-fi movie. What kind of plot would it have?
Ian Lurie: Wow. Well, I don’t think — I hope — it would not be like Battlestar Galactica.
Gianluca Fiorelli: Well, in Galactica, we were using analog phones.
Ian Lurie: Yeah, but in that case, most of human civilization was wiped out because of AI. But yes, the survivors only made it because they used very basic computer programs. Like I said, I hope it doesn’t go that far. A good example is the Blade Runner franchise. I don’t know if this is the answer you’re looking for, but if you put the two Blade Runner movies together, I think that’s a better way to see it.
There are bumps, difficult times, and struggles, but ultimately, at the end, there’s potential for AI and people to collaborate and do something useful. So, I don’t know if that’s helpful or not, but that’s kind of how I hope it goes. I’ll admit, though, I think the Battlestar Galactica conclusion is more likely.
Gianluca Fiorelli: Yeah, but somehow it ended with a false promise, if I remember well, in Battlestar Galactica. I’m paraphrasing, it’s not the exact phrase: everything that has passed will come again, or something like this. You know, in the last frame, there’s Baltar with Number Six, playing these silly little robots in the streets.
Ian Lurie: Trying to deliver Uber Eats.
Gianluca Fiorelli: Yeah.
Ian Lurie: “All this has happened before and will happen again.” I think there’s something to that. Maybe that’s part of why I find Battlestar Galactica so appealing — there’s a lot behind the premise. I don’t want to get too philosophical, but people tend to plunge ahead with great new technology, and it often brings tragedy along the way. And then, in the end, they’re left with Uber Eats robots bumbling around in the streets trying to deliver food.
Gianluca Fiorelli: And let’s talk about your other great hobby, which is cycling. Why cycling in Seattle? Because I know Seattle — it can be harsh.
Ian Lurie: Weather-wise, and with drivers, yeah.
Gianluca Fiorelli: I remember that you bruised yourself.
Ian Lurie: I can show the scar. The last few years, cycling in Seattle has been a little sobering for me. Drivers have become very careless, and my luck hasn’t been great. I’ve crashed twice. The streets are a mess. It is improving, though. If you ride just outside the city, it’s absolutely gorgeous, so I highly recommend it. Cycling in downtown Seattle — if you’re into video games — definitely do it.
Gianluca Fiorelli: Like Frogger.
Ian Lurie: Like Frogger? Yes, like Frogger. Or, if you’re our age, the video game Tron when you’re driving the light cycles. There’s a certain amount of that. I’d say those are both good examples of what it can be like to ride a bicycle in downtown Seattle.
Gianluca Fiorelli: And what would you say to your younger self?
Ian Lurie: Professionally, I’d say “Spend a lot more time becoming a good leader, and less time trying to learn everything about search.” In the end, that was the make-or-break moment for Portent, and while it worked out okay, it was more of a struggle than it had to be. Personally, I’d tell myself, “Don’t bring quite as much of it home.” Those would probably be the two big things. Both are pretty cliché, but definitely the advice I’d give.
Gianluca Fiorelli: And just one last question: the longest D&D game you ever guided — how long did it last?
Ian Lurie: The one I’m running right now. If I’m remembering correctly, we started at the beginning of COVID, and it’s still going on. So that’s five years now, somewhat consistently every two weeks. For campaigns I’ve run, that’s probably the longest one.
Gianluca Fiorelli: Wow, that’s a long time.
Ian Lurie: It is.
Gianluca Fiorelli: Have you ever thought about writing down all the stories you’re inventing?
Ian Lurie: I have about 60,000 words of it written down in a huge Obsidian database. It’ll probably never be suitable to share, but it’s great lore for other games I run.
Gianluca Fiorelli: Cool. I like it. Wow, that was one hour — a wonderful chat with you. I’m sure our listeners and viewers enjoyed it. Let’s make a promise: in six months or more, let’s see where we are. Whether we’re still in the middle of something new or if we’ve finally landed somewhere new.
Ian Lurie: I think there will always be something new, but yes, let’s definitely do that. Gianluca, thank you so much. This was such a pleasure. I really enjoyed chatting.
Gianluca Fiorelli: And thank you — and thanks to all our listeners. Now, this is the classic YouTube semi-professional influencer thing: remember to ring the bell so you’ll be notified about new episodes, and of course, subscribe and give it a like. If you enjoyed this episode — which I’m sure you did — thank you, and see you in the next episode of The Search Session.
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.
stay in the loop