Episode 08: The Importance of Creativity and Expertise in AI-Dominated Marketing

In a world where AI is increasingly dominating marketing strategies, how can marketers stay relevant and continue to deliver value to their clients? In this insightful episode of Digital Clinic, Aaron Burnett engages in a captivating discussion with Ian Lurie, a seasoned digital marketing expert and Founder and former CEO of the acclaimed agency Portent. Together, they explore the critical importance of creativity and expertise when handling the challenges posed by AI, sharing practical advice on adapting to AI-driven search results, building owned audiences, and maintaining ethical marketing practices. This episode is packed with valuable insights that will help digital marketers thrive in an AI-dominated future. 

Ian Lurie’s Background and Portent

Aaron: Welcome to Digital Clinic, the podcast that goes deep on critical digital marketing trends, strategies, and tactics for the healthcare and medical device industries. Each episode brings you expert guests sharing the knowledge, insights, and advice that healthcare marketers need to be successful in this complex and rapidly evolving digital landscape. I’m Aaron Burnett, CEO of Wheelhouse Digital Marketing Group, and with me today is Ian Lurie, the Founder and former CEO of Portent – a story digital marketing agency based here in Seattle, Washington, that Ian established in the mid 90s, was acquired by Clearlink in 2017. Ian continues to be a highly sought after digital marketing expert, conference speaker, and consultant, and he also teaches digital marketing courses at the University of Washington. Ian, I’m excited to be able to talk to you. Thanks very much for being with me. I’m looking forward to the conversation.  

Ian: Thanks, Aaron. Yeah, I’m very excited too. 

Aaron: Here’s what I thought we might talk about. We might talk about your background a bit, and I thought that might lead us to Portent. Then if you wish you could talk about what’s happened recently with Portent.  

Ian: With Portent, I have some strong opinions about the people who led to this, so are you comfortable with me expressing them?  

Aaron: Oh, I would love for you to express them. I saw a link to the video. 

Ian: Okay, so you know. 

Aaron: Oh my gosh. 

Ian: When Clearlink did the deal – and anyone who’s interested can just go Google “James Clark Clearlink” and see the interview video and see the town hall video he did that is now so infamous – you get a sense of the business thinking on his part. It’s my understanding that some clients found out on social media, that they no longer had an agency. There was a lot of, “It could have been handled better,” I wrote about that.  

Aaron: That may be the understatement of our conversation. 

Ian: Yes, and there are still amazing people from Portent out there that are looking for work. I think you said you’re going to publish a link to a spreadsheet that lists them all. One of the best assets of Portent was always the team. Strong brand and a fantastic team. Clearly, they have crapped all over the brand at this point, so we can forget about that, but the team is absolutely magnificent. If you can get any of those people on your team, you’re going to benefit.  

Aaron: Well, as you know, I wholeheartedly agreed. We hired two of them, and they’re fantastic. Interestingly, from the outside, the brand still has value and the brand is respected. The behavior of the brand under Clearlink is not respected, but I think it’s understood. 

Ian: Yeah, I think so. In the end, what you have now is a bunch of clients who feel like they were abandoned, and that’s never good. You have a huge team, that have all said wonderful things about Portent after shuttered, and they were still left out on the street with very little time, and from what I understand, not very good severance packages at all. That impacts the brand, and even though I haven’t been there in any capacity for five years, I feel it very personally.  

Aaron: The lack of care is stunning.  

Ian: Exactly. It’s not just about being kind of “woo-woo, squishy, nice” to people, which I believe in strongly, but that is also how you build a healthy professional services business.  

Aaron: Well, I mean, this is just about fundamental decency. So, I’m curious about this. You present yourself, in some quarters, as sort of the SEO curmudgeon, and yet my sense of you and your actual ethos is quite different. You’re optimistic about people, you’re kind and caring to those people, while at the same time being somewhat curmudgeonly.  

Ian: I don’t think they’re mutually exclusive. I like to call myself an unlikely idealist. I am very curmudgeonly because I look straight at what a lot of people are doing in our industry, and I say, “You’ve got to be kidding me. Do you really think that’s going to work?” I said someday I’m going to write a memoir, and it’s going to be, “Am I missing something?” That’s what the title is going to be. I look around and I see people doing stuff, and sometimes I just think, “I’m not sure why you’re doing that.” What makes me curmudgeonly is that people are taking these assumptions, whatever they might be, and they are applying them to other people’s businesses. A client will come to you and say, “Hey, I need help. Please help me grow my business. I make this much money. I want to make this much money. We’re kind of struggling right now.” You, the marketers, say, “Okay, we can help with that. What we’re going to do is we’re going to hire a bunch of writers, and we’re going to bury your site and crappy content.” For two months, it works great. Then that business loses all their organic traffic and they suffer and have to lay people off. That’s the part where I get angry. I’m not mean, but I do get angry about some of the behavior I see sometimes. That’s what makes me a curmudgeon because there’s certain fundamental rules of marketing that have nothing to do with digital. It’s just the way it works. You’re selling to human beings. Everything out there, Google, everything else, is attempting to emulate those human beings. If you try to game the system too much, or if you forget those basic assumptions behind marketing – “Audience, you have a problem. Here’s how you fix the problem. Here’s how we’re going to help you fix the problem better than anybody else,” – if you forget those three basic questions and you fail to answer them, then nothing else you do is going to matter. Cranking out all this kind of “meh” content and doing all these other things that I’ve always been a little suspicious of, it’s a real disservice to clients.  

Aaron: It sounds like the through line is a duty of care.  

Ian: Yes, that’s a good way to put it. I do feel a lot of responsibility for team, for clients, and for the industry. I went to law school, and when I told my family, I was going into marketing, it was really the only profession even less respected than law. There’s a reason so many people approach marketing such suspicion. That is the “sell at any cost, grow sales at any cost” mentality. That’s also just not good business. That’s the part that I always try to tell people; it’s not just about a duty of care, which I do feel. I also ran and run profit seeking enterprises. I mean, I want to make money. You don’t do that for the long term by engaging in tricky, short-term, fast growth, fast burnout behavior.  

Aaron: Yeah, I completely agree. I think, in many instances, the right response to what a lot of agencies do is outrage. 

Ian. Yeah, or like I said, “Am I missing something? Is there something I’m missing here? There must be a trick that I don’t understand.” Every time I say that, it turns out no, there’s no trick.  

Aaron: You mentioned that you have a law degree, and I read that you practice for an hour. I’m really curious to know what happened in that hour.  

Ian: I didn’t even practice for an hour. I went to law school, and I quickly realized it was not for me, but the other thing I am is extremely stubborn. So, I finished law school. During law school, I worked for a judge on the Ninth Circuit, which was an honor, and it was really interesting. It was incredibly high pressure but further convinced me that I hated law. I worked for a city attorney’s office. I did a bunch of stuff like that. If you take all that and smoosh it together, I think it counts as one hour of actual legal practice. But yeah, it really did help. My last year of law school, I paid for my books by working in a bike shop, and then I graduated and moved on to other things. 

Aaron: An impressive credential.  

Ian: Honestly, the education is a little helpful when you’re running a business and not for the obvious reasons. Just because it is a professional degree, you learn how to work in an office and run a business and write a contract and do some of the things that we otherwise forget and screw up our businesses when we’re starting.  

Aaron: Right, and I would assume, learn a systematic way to think about things. 

Ian: Yes, there is a way of problem solving and coming at problems, plus communications, that is surprisingly helpful in marketing, particularly in digital marketing.  

Leaked Google Search API Documents

Aaron: We’ll talk about AI a bit, but you mentioned something else. You mentioned the behavior of lots of agencies and sort of the rote approach that they take to marketing, the unthinking approach, best practices or “checklist” digital marketing. That’s an interesting transition to the Google doc leak. What is your reaction to the leak? What did you see that was surprising? What did not surprise and what’s your feeling about it?  

Ian: I wish I was that surprised, and I don’t mean individual bits of information. I mean, Mike King, and Rand Fishkin did an incredible job of digging and what they found. I went back and looked at the actual raw docs afterwards. I was looking, and I was like, I have no idea how you guys made any sense of this.  

Aaron: I exactly the same reaction.  

Ian: I read it four times. I picked up a few things. The main thing I picked up is, Google is very, very cagey about what they tell us. I found it very funny that two days ago, they published a statement finally saying, “Be careful about taking these documents out of context,” when all they ever do is give us information about how search works, completely out of context, right. Like Navboost, the idea that that user behavior impacts rankings, they’ve denied that for years. But, if you read back over their denials, they’re always very careful about how they deny it.  

Aaron: It’s an intuitively correct signal.  

Ian: Well, it’s intuitively correct. I have to admit, when I looked at it, I thought, “I wonder how they’re doing this,” because it could be very easy to game, but the truth is at the scale that they’re working, it’s not that easy to game. You’d have to hire; I don’t even know how you would do it. How many people would you have to hire? 

Aaron: All the people on Fiverr.  

Ian: Yes, exactly. You’d have to hire everyone on Fiverr, and put them on a very specific schedule. Yeah, I wasn’t that surprised by that. I was really interested in how many Twiddlers there are, these little mini algorithms that fire after the main algorithm has decided how to rank some content. There are so many of them, and it makes sense, now that I think about it, that what they’re doing is they are generating the rankings and then adjusting them. If the adjustments, those Twiddlers, do what they want them to, then eventually those Twiddlers are incorporated into the algorithm. Which means that Panda, and I apologize if there’s people listening to this and suddenly I just started speaking another language. 

Aaron: It’s okay. We’ll include a link to the Panda.  

Ian: Okay, thank you. Panda probably started as a Twiddler. There were also some Twiddlers in there that make me wonder if they do, at least accidentally, sell preference. Maybe not intentionally, but at least accidentally, there’s some stuff happening. If you look at some of the YouTube stuff in there and some of the things about Navboost, they’re pulling it all from Chrome, or at least a lot of the data is being pulled from Chrome. Chrome is a Google browser, right? I would stop short of saying Google intentionally maintains a stranglehold over the rankings with their own technology just because that seems like a big accusation to make, but it seems like there’s some necessary inevitable feedback loops that set up. If you’re pulling data only from Chrome users, then you’re getting data from a very specific set of people, particularly since all Android devices come with Chrome by default. That’s bound to have an impact.  

Aaron: Interesting. My reaction was in two different directions. One, that the things that were disclosed, had been, to me, self-evident for a long time. If you had ignored best practices that Google recommends, and so you ignore checklists and digital marketing, and instead you test and you look at evidence and you lean into what actually works, rather than the list of 1,000 things that you’re supposed to do because their best practice, then you would have been doing things that just make sense. On the other hand, I did think what was frustrating and offensive was how vigorously they attacked some people who asserted that these things were true, and how personal the attacks were. 

Ian: I always had a lot of respect for Google engineers. I’m going to say that right now. I lost a lot of respect for them the day that Gary Illyes did an AMA (Ask Me Anything) on Reddit and just went after Rand Fishkin. 

Aaron: I saw that. 

Ian: It was deeply personal. It was not a professional statement. There was no apology, and there’s never been one. They went after Ross Jones pretty aggressively, Another Google engineer jumped on Twitter to rail at me about railing against JavaScript because I will continue to maintain that JavaScript is a huge issue in SEO. They were like, “It’s not about JavaScript, JavaScript is fine.” Then they give you 500 pages of documentation on how to compensate for the fact that they’re using JavaScript. That part really bothers me. It is disrespectful, coming back to my them, and it’s really unprofessional. It’s moments like those that made me trust them less and less, and now seeing these documents just kind of confirms that we shouldn’t trust what they’re telling us.  

Aaron: Exactly. Would you run through your professional history? Talk a bit more about Portent and talk about the consulting work that you do today. 

Ian: I’m going to go all the way back. I grew up with two PhD scientists. My parents have and had PhDs in physics and engineering. In the 1970s, we had a TRS-80 model one sitting in our office, we had a Heathkit, all that stuff. So, I was very comfortable around computers. I say that because I majored in history. People always say, “You majored in history, how did you end up in digital marketing?” Because I’ve always been very comfortable around computers. It never occurred to me that computers are something you had to go to school to learn to use. That sounds disrespectful, and I don’t mean to be disrespectful. I’m not a programmer or an engineer, but I’m not intimidated either. I got a history degree, went to law school, did not enjoy law school, as I said. I did enjoy working on the law journal I worked on there. I worked on the environmental law journal at UCLA, and I enjoyed the editing and the writing. I worked in a bicycle shop. I enjoyed introducing people to products. I did not enjoy sales, but I did enjoy kind of showing and guiding people to things, and I love cycling, so it was just nerd heaven for me. Then when I moved up to Seattle after law school, I worked in an engineering firm that needed someone with some legal expertise and some technical expertise. I did that for two years and decided I wasn’t really going to become an engineer and told the owner that I was leaving to go start doing some marketing, copywriting, things like that, and he said, “We’ll be your first client.” That’s how Portent was born. Initially, it was called The Written Word. I’m a little bit of Shakespeare nerd as well. It started out as The Written Word, and I ran it until 1999. We were doing marketing, copywriting, some technical writing, some web design, and just starting to sort of touch on search. I briefly sold the company to another company. It’s a long, sad story, and it did not go well. By the end of 2000, we were out on our own again, and that’s when we started getting into search, expanding the paid search as well. Still doing a lot of web design, a little bit less copywriting because we were getting too expensive for people, more and more digital marketing consulting, and a lot of analytics. When Urchin first came out in the good ‘ol days, we did a lot of work with Urchin, and just kind of grew from there, you know. Sold in 2017. That’s the whole history.  

Now, I am a digital marketing consultant. I spend about half my time on SEO and the other half on helping clients with digital strategy and content. I work with some clients on team processes and getting teams to work together well within organizations. I do some training for clients as well. Actually, it’s sort of a sideline, I guess, I don’t know what I would call it, but I’m actually teaching digital marketing at University of Washington.  

AI as a Marketing Tool

Aaron: Well, let’s shift and talk about the thing in the zeitgeist, other than the Google doc leak, which is AI. Last year at about this time, you wrote that ChatGPT was not AI. It was a parrot that had spellcheck. Do you still have that opinion?  

Ian: I do. It’s changed a little, and I’ll tell you the parrot expression, I got that from an article that calls AI stochastic parents, I think. For GPT, stochastic parents is one of the seminal, scientific but also accessible pieces of writing about AI when it all started. To me it is still retrieval augmented generation of language, meaning you build this enormous model of language. You have all these words out there, you do a lot of math, you figure out which words are most often closer to which words, and then you decide, “Okay, someone just typed in a prompt. These are the logical next words in response to that prompt.” It is still parrot-like, but it is becoming much more sophisticated and harder and harder to detect. It is still only as smart as the content that’s been given to it, and it’s always a little bit less smart than the content that’s been given to it. Because it makes mistakes and doesn’t realize that it’s made sometimes hilarious mistakes.  

Aaron: Yeah, that’s certainly true. It’s also highly variable in its predictability and consistency. It gets smart and dumb in cycles, which is interesting, and problematic. 

Ian: And it still can’t do math. Seriously, I will take GPT more seriously as a marketing tool because there are things I use it for. For me, I’m at best a third-rate Python coder, but I know what I want and I know what’s possible. GPT helps me write Python that I never would have been able to figure out, in less than four or five days, and now it’s a half day job for me. It does some things incredibly well. It’s actually good at organizing data in spreadsheets, if you need to do that, if you need to check your work. Just don’t ask it to do a calculation. The day can do a calculation is the day I start taking it much more seriously. It’s kind of an overall generalized intelligence still. 

Aaron: Do you have the same opinion of other models?  

Ian: Yeah. ChatGPT for me is still pretty high up there. Perplexity is a better coding tool, in my experience. Gemini is just kind of interesting in the way it does things. I don’t think it’s better than ChatGPT. I do think it’s quite good. GPT is better at code, in my experience, and then there’s very specialized tools, Copdilot and things like that that are also really useful. I really recommend to people to get Kore, which is, I don’t know if you’ve seen it, but it’s an AI tool, but it also lets you run many models next to each other. I can generate a prompt, get a result using GPT4, and then I can also get the same result using Gemini, just by pushing a button. Then I can get the same result using Claude, and then I can compare all those results all in one place.  

Aaron: I haven’t used Kore; that sounds really valuable.  

Ian: It’s convenient because you can do all that anyway, but for some reason, I don’t do it until I have it all in one place. There’s some other stuff it’s really good for too. Wil Reynolds has a good short video he did on using it.  

Aaron: Well, so you mentioned Wil. It seems to me, it’s absolutely true, that there is a continuum of opinion regarding the implications of AI for digital marketing in particular. Wil Reynolds has, for the most part, been on the end of the continuum, which is, “We don’t know the timing, but it’s going to eat our lunch at some point, and we need to prepare for that.” There are others who are the other end which say, “It’s a parlor trick. It doesn’t really have a lot of value.” Where do you fall? 

Ian: I have said it’s a well-trained parrot, but I want to make sure people out there also understand that it is definitely not a parlor trick. It is coming for us, but so was Google, and so was digital. The Mad Men would have been horrified by the idea of digital marketing and the idea that they no longer stand between the products and the brands and the audience. The whole industry had to shift from “we’re the gatekeepers of access” to these massive media empires, and if you want to buy an ad in a magazine, you have to come to us anyway, so come to us for the creative too, and we’ll bill you for that. That has changed. Right when digital came along, suddenly it was, “Okay, now you’re paying us for our expertise and just our expertise. Sure, we’ll manage your paid media buys for you, but the truth is, if you really didn’t want to work with us, you could go do it on your own or give us your credit card and we can run the ads for you.” That changed the relationship, and AI is going to change it again. AI is going to change it because at some point, it will become smart enough to help make some decisions about marketing. It is nowhere close to that right now. When people talk about it eating our lunch, they’re mostly talking about content generation. They’re talking about how “Oh, I don’t need anyone to write my blog posts anymore.” I had a client who said, “From now on, we should be producing 100 highly technical articles a month. I mean, it should be easy,” and then he went and did it and came back and said, “It’s not as easy as I thought.”  

Aaron: You could be fast but not good.  

Ian: Yeah, exactly. The faster, better cheaper, right? It’s going to erode our lunch over time, and what you’re going to get is right now the sophisticated marketers aren’t fooled and sophisticated clients aren’t fooled. Because right now you have to be fooled to really think that it can really take over all of your marketing needs as far as creating content. Forget everything else because it can’t do anything else. It can’t do analysis. It can’t do forecasting. I’ve tried it, trust me. I hate doing forecasting, but it is chipping away at content a little bit at a time. We have to be ready and get better and better at helping clients use AI to generate good content and use AI to make good decisions. I don’t know if it’s going to reduce the money that we can make as marketers, but it’s definitely going to change our role. Also, by the way, our awareness of the shortcomings is part of the value that we’ll bring to it.  

Aaron: I agree. Going back to the discussion we were having earlier, I think this is a winnowing process, and it will eviscerate agencies or practitioners who are following checklists blindly or are doing work that can be well documented in terms of process, just repeat it. If it’s well documented and repeatable then that’s something that AI can now or in the future begin to do.  

Ian: It’s funny because when you were talking about checklists, I was going to say, checklists are important in digital marketing. There are certain things you want to list. The problem is that humans are bad at checklists. Humans tend to become obsessed with the checklists, and they stop thinking about other stuff. That’s why I always resist so much when clients say, “Isn’t there just an SEO checklist?” Yes, there is, but you really don’t want to start following it. You’re going to miss important things.  

Aaron: The value is in discernment.  

Ian: Yeah. You’re an engineering team, you’re asking me for an SEO checklist. The reason you’re there is because you can think about this way more intelligently than I can, so let me give you guidelines, but there are things that can be turned into a checklist, and that’s where AI can really help. You can already go into OpenAI and give it a set of instructions and say, “This is good, this is bad,” and then give it something else to say, “Is this good or bad?” It’ll give you a pretty good answer. You have to be very good at training the model, slightly less good than we had to be when it was straight up machine learning. 

Aaron: I agree.  

Ian: There are developers and natural language processing folks who are going to scream at me for comparing AI to machine learning, but it is a little bit easier now.  

Aaron: Yeah. We use it selectively. We actually don’t use it for content generation for clients, because it’s so obviously AI-generated content, but we use it for a number of specific tactical uses. We built a custom GPT that diagnoses issues with schema and then corrects them and generates schema. We have another that will structure and write content briefs against the content strategy. We don’t want it to write the content. Structure the brief.  

Ian: That’s an interesting way to put it. It’s good at introducing structure to things because that’s really what it’s kind of synthesizing out of that massive language model that it has. That’s why it’s so good code because code has a structure. In that language model, there’s a gazillion lines of code, and it’s using that to decide how your code should look. 

Aaron: The other thing that we did that has been valuable is we took Google’s content quality guidelines, and we trained a GPT on those content quality guidelines. We get a pretty robust assessment of any given site or page or setup pages and the score and diagnosis and suggestions. It works pretty well. You still have to continue to refine, but it works pretty well.  

Ian: Did you do that in in a custom GPT or a ChatGPT? 

Aaron: We did. We built a number of them that we use, but very selectively. We have found that if we actually want to generate copy, and again, we don’t generate copy that’s final copy, we’ll use it as a jumping off point to explore ideas or to structure ideas. We get a better result with Claude. ChatGPT goes straight to middle school purple prose really quickly.  

Ian: It uses words like transformational. 

AI’s Impact on the Future of Digital Marketing

Aaron: And delve. Yes, exactly. Let’s project forward then. What will be true in the future? What do you think a digital marketing agency looks like four years from now, five years from now?  

Ian: I don’t think we’re just going to be a room full of AI engineers. I know a lot of folks talking about prompt engineering and how that’s the future. I don’t agree with that. I think AI will become another tool. We still have to be creative marketers and creative thinkers. We still have to be critical thinkers. We still have to be always trying to unwind whatever algorithms are out there. AI is going to be just an assumption. We’re going to use AI to do an initial layout for a website. We’re going to use AI to come up with ten logos, or something like that. We’re going to have to have the judgment and the skill to turn those into real usable things, and that’s still going to be a significant undertaking. The real question for me is how the clients will change. That’s what I would like to see as the end result, and I think where we’re going. If the mass market, if everyone out there decides that “meh” is good enough, then we’ve got a problem. If every big company in the world decides that the crappy content that they generate, that doesn’t generate much of a response, is okay, then we got a problem. Our job as marketers will also be to continuously inform people as to why it is important that they not do it that way, which is something we do now. We’re telling people look, “Google’s great, but do not stick your whole business on Google. There are other things you should be doing too.” Creating a content farm may work for a little while, but unless you just have a garage full of Viagra you’re trying to move overnight, it’s not a good long-term business plan. Right. That’s the kind of education process we’ve always been in as digital marketers. I think we’re still going to be in that business.  

Aaron: I agree. I think the things that should matter today will become the only things that matter in the future, and that is truly deep expertise and creativity. 

Ian: And communications. I mean, the ability to just communicate and communicate effectively has always been what sets great marketers apart from everybody else, and that’s why you become a marketer, you’re a skilled communicator. No matter what your specialty, the ability to communicate well will still matter. I think we’re a very long way from AI being so good at communicating that it can replace us. Even if it gets to that point, it really does free us up to get even better at communicating in other ways. That’s always kind of been the way it has worked.  

Aaron: Yeah. I have also wondered about the potential slide to mediocrity, as AI becomes more pervasive.  

Ian: Yeah, and that’s what I was kind of touching on. The idea that if everyone decides “meh” is okay, then that’s what happens.  

Aaron: I think about that societally as well. If in fact, conversational AI becomes pervasive, and now instead of doing actual research and due diligence, we accept least common denominator synthesized response, and all we have to do is speak a question and we get a spoken response, and the response is met, I worry about that.  

Ian: Well, that becomes a deeper political and philosophical discussion that I could have for a week. The question is whether having tools that handle the basic 90% of problem solving and communication frees us up to get even better at the 10%. This is where I’m the cynical idealist. I think tools like this should free us up to do more and more amazing things. In marketing, in communications, in whatever, we should be able to improve the state of the art every single time we get a new technology. Digital marketing, as much as writing publishers like The Verge love to talk about how digital marketers are destroying marketing and destroying the industry, the truth is digital marketing, the practice of it, has really elevated marketing, if you compare what we do now for communications and effectiveness of communications, to the 1950s, 60s, 70s, that was some brilliant stuff, but it was not as effective. I think people who are good at this are going to get even better at it. The question is whether we can get better enough to show enough people that we’re worth it, that we can keep doing what we do. I would be shocked if that didn’t happen. I’d be shocked if AI actually ended marketing or somehow fundamentally changed marketing in a way that it looks nothing like marketing does now. I think what it’s going to do is the state of the art is going to increase considerably.  

AI Overviews and Search

Aaron: Yeah, I agree with that. Let’s shift to AI in search results, AI overviews, Google, what Microsoft is doing.  

Ian: Yeah, that’s a shitshow. That’s a shitshow for many, many reasons. I just changed us to PG-13. 

Aaron: I think the listeners can take it. 

Ian: It’s bad because first of all, it’s stealing. I mean, even more than Google’s practices with search snippets and all that feels like plagiarism because it kind of is. AI-generated results are plagiarism. They are generating content based on what other people have written, and they are not providing credit. Those three sad little links or whatever that appear in the content are not enough. When I think about what used to make my blood boil and still does is when someone would steal my blog post, put it on their site, and at the bottom say, “This is from Ian Lurie’s blog,” or “This is from the Portent blog,” or whatever and have one sad little follow link back to my site. They know exactly what they’re doing. They’re just trying to weasel out of being a thief of content. Google’s doing the same thing, and so is Bing. Their lack of interest in finding a better solution is, I wouldn’t say it’s shocking to me, it’s amazing. It’s going to come around and bite them. I mean, Google’s already been in front of courts in the US and the EU over other practices. This is only going to make it worse. So, it’s a shitshow for that. It’s really bad for brands because it is further reducing the real estate in which you can present yourself. Those AI Overviews show up, and they immediately shove everything else down the page, and they’re terrible. It’s interesting because Google, yesterday, published a statement saying, “We know there’s been issues, but it’s only like a few hundred out of billions of queries.” If it’s a few hundred out of billions of queries, then I must be very unlucky because I’m seeing lousy results all the time. When it tells me I can smoke two to three cigarettes a day when I’m pregnant, I mean, that’s not a small mistake, and don’t tell me that that was me asking a satirical question. That’s just lame. How could you possibly give me that answer? It’s got a lot of problems. It is being pushed out there too quickly because these are companies that make money off of stock, and they know that if they don’t present something with the letters AI in it, their stock is going to fumble.  

Aaron: Where they seem to have clarity is their advertising strategy in AI Overviews. That seems to be far more carefully thought out than the organic result. It seems to me that AI Overviews could potentially represent an existential risk for Google. Their business relies on fair use doctrine, and it’s not clear to me that an AI Overview respects fair use.  

Ian: When I talk about plagiarism, that’s really what we’re talking about is it’s not fair use at this point because you may have grabbed words from ten different articles to generate this AI Overview, and then you have two or three links to two pieces of content that are a tiny little part of what was written there. There’s also been examples found where those three cited pieces are not actually part of the content that was used to generate that. People have found that websites that are no longer live have content that’s being published in the AI Overview, which makes sense because the model is always lagging a little bit right. Also, Google depends on advertising. They must have had this discussion internally, somewhere. If you generate an AI Overview, no one’s going to click the ads. If you look at the patterns, I think what they’re doing is they generate AI Overviews for highly informational intent queries, where probably they’re not getting that many ad clicks anyway. Then they’re going to start using AI Overviews to generate actual product pages in their own results. There have been some examples already where if you search for certain products, Google actually generates a product category. A lot of people have been talking about this, but I think that’s probably why they’re not worried about it reducing ad revenue. But again, is that fair use? Google went after comparison shopping engines because they said comparison shopping engines were pulling products from many sites and building them into these big engines that brought no additional value. Now, what are they doing? They’re pulling products from many sites, and they’re using it to generate a product page. Are they adding value? I don’t know. 

Aaron: The new ethos is don’t be evil, except when it’s super convenient or profitable.  

Ian: Don’t be too evil. Don’t be so evil that people can point at you and say, “You are unequivocally evil.”  

Aaron: Operate in the gray. Well, okay, so the implications then for SEO. If we’re dealing with a land of AI Overviews that displace conventional organic results, how should SEOs respond?  

Ian: One thing that I’m recommending to a lot of clients right now is start figuring out how to blend informational and transactional content. This is partly because of those product results, whatever they are, that Google is generating now using AI and partly because of the way they’re constructing the AI Overviews. If you can tie informational content, which is often what is cited in an AI Overview, to some kind of transactional content, then you may get a link in the AI Overview that points to a page that has both your information and transactional contents. It’s not pure conjecture, but it’s a lot of conjecture, because no one has tested this yet. I have some vague concept of how these models work and what Google is trying to accomplish with them. This seems like that makes the most sense. You can see how much I’m kind of trying to keep wiggle room here.  

Aaron: Well, it’s very early. As you said, nothing’s absolutely knowable.  

Ian: This is part of where I was talking about state of the art. Even before AI, one thing that I was seeing really great marketing doing is starting to tie together informational and transactional. In the example, I always use REI. For people who don’t know or who aren’t in the US, it’s a big recreational equipment company. Everything from bicycles to climbing equipment to kayaks to tents. Their website is a great example. They have “how to” posts on things like, “Here’s how you fix a flat tire on a bicycle,” and then in that post, they will have many, many links in very helpful places to products that they offer that will help you do it. “Use this patch kit. Use this tire lever.” At the end of the post, they’ll have a link that’s something like, “Click here, and you can add all these products to your cart,” and you’re good to go. That’s marketing, that’s sales, but it’s also helpful. I don’t have to buy any of their products and that’s already helpful to me. Then if you go to a product page on their site, they will often have links to those kinds of informational posts. There are some other sporting goods sites that do this very well, where they’ll just have sort of comments and reviews by product experts right on product pages saying, “If you’re a downhill mountain biker, then this bike has lots of travel in the front and rear shocks. You probably want this. If you’re a cross-country racer, you probably want to look at something more like this one over here.” That’s the blending of informational and transactional, and I think we’ve been going that way anyway.  

Aaron: Yeah, that makes a lot of sense. I also have seen some conjecture, and I’ve wondered myself if an effective approach might be kind of a continuation of some of the strategies that we’ve used to do things like achieve Featured Snippets for you understand the structure that Google is seeking. An AI Overview seeks to answer an informational query. Structure your content in a way that answers a query and make the job easy for Google. I also have suspected, and I think this has got to be true, that this is another area where being diligent about using schema will be really valuable. 

Ian: UX of content. It’s funny because I skipped over a whole bunch of stuff like that. Schema is going to be more important because resources are going to become even more constrained because these AI models consume an incredible amount of electricity and processing power and everything else, so the more you can structure your data, chances are the better you’re going to do because Google doesn’t have all the energy and all the resources left to do the kind of unstructured ranking analysis that you do now. Again, conjecture, but UX of content is also going to become even more important. The way your content appears and is laid out on the page, I think, will at some point, there will be a whole separate AI that just interprets and parses the way the content looks on the page. This is again, where the scientists are just going to yell at me. A component of the AI, if you are something like Google, will be dedicated to examining the page and how the content is treated on the page in a very, very careful manner, in a really detailed, manner to decide what the most important concepts are on that page.  

Aaron: That makes a lot of sense to me. But again, going back to what has been state of the art, we should have been behaving in that way anyway. If you had assumed that, for example, user engagement signals or ranking signals and that what you want is to serve user intent and to aid people in getting the information and achieving the outcome that they had, you would have been designing for that sort of thing anyway.  

Ian: Exactly. I don’t want to say nothing’s changed, right? That would be extremely unwise, but best practices, and again, we’ll get into the checklist thing, but certain good ways to do things provide the good user experience. Those are all things that still matter a lot. Google has always been caging. We all roll our eyes when they say, “Produce great content, providing good user experience,” but the truth is, there is something to that. We’re trying to market to humans. We need to provide something that’s good for humans, and that’s going to continue to be true. The only question with AI, and the thing that worries me, is whether we’re going to get credit for it, or whether the search engines are just going to grab it all and take credit for it for themselves.  Aaron: Yeah, I feel like the one area in which Google has been consistent from the beginning, was to say, “Produce good content and serve users and everything else will work itself out.” That was naive blandishment early on, but I do think that in a pretty clear way, what Google has done through the years is to become more sophisticated in actually delivering on that promise. Sometimes through subterfuge, but they still are looking to deliver on that promise.  

Ian: I feel like they would have been better served and would still be better served by simply telling us the truth. They don’t have to tell us the weight of all these factors, but if they had said, “Clicks impact rankings,” no decent marketer is going to think that they can use clicks to mess with the rankings. What we are going to say is, “Okay, we need to provide content in a manner that provides a really good experience in the SERP.” If dwell time matters, then we need to provide content that gets people to dwell longer on our sites. That’s what always bugs me about it. It’s like if you guys are just honest, you’re going to get what you want anyway. Yeah, there will always be people who try to game the system. They’re doing it anyway, and that’s not going to change. I think a lot of what will help you perform in an AI world is the same things that will help you perform in a more standard SEO model. The problem is whether we are going to actually get credit for it or whether Google is going to keep that traffic for themselves.  

Aaron: In the current environment, we’re not entirely sure what’s driving AI Overviews, where organic links have been displaced by AI Overviews in many instances. You mentioned that you’re advising clients to blend informational and product content in the same context on the same page. Are there other tactics that you think digital marketers should be considering to end around search results, to market in other contexts, to think more expansively about search? This seems a channel where, at least in the near term, maybe we have lost control, lost influence. So, I start to think about other places where maybe we can be more effective.  

Ian: Yeah, got another podcast episode we can do on this? This is what I’m recommending. So, clients are still coming to me always asking about SEO, SEO, SEO. What I’m telling all my clients, and have for years, is do not just depend on SEO. Owned media is actually your best asset. What you want to do is you want to capture people’s attention, and then you want to give yourself the opportunity to talk directly to them. That means some pretty typical stuff, like building an email list. I know people say, “Oh, no one uses email.” Lots of people use email. I have tons of clients that make a lot of money from a small, niche but very high-quality house list. You need to find ways to retain the attention of those users when they come and find you. I don’t mean popups when they try to leave the website. I mean, you do need to build that house list. You do need to do more traditional marketing and PR, in that you’re going out and trying to find places where you can go talk about what you do, for example. You do need to look at social media, but I have to tell you, I think social media is also earned media. I think the transition of Twitter to X to dumpster fire is a good example of that. We do not own the audiences that we build there, but certainly you want to look at that as an alternative. You want to build a house list of some kind, you want to have that list of people who you can reach out to and who genuinely wants you to reach out to them. If it’s ten people or 100 people or 1,000 people, it doesn’t have to be million people no matter how big your company is, it has to be built that you know they’re going to respond every time you contact.  

Aaron: I think that becomes even more key as we enter a realm in which we’re deprecating third-party cookies, and you now need to rely on first-party data for targeting and efficacy of marketing overall. We work with a lot of healthcare and medical device clients, and we’ve needed to rely on first-party data for a long time. 

Ian: Right, because of HIPAA, and things like that. 

Aaron: Well, and now, my thesis is, what has happened with healthcare, with HIPAA-covered entities, is going to happen for everyone over the next two or three years.  

Ian: Did you hear, in the EU I think, someone has taken action against Apple because some feature of Safari that’s coming soon, they’re introducing something that will remove ads from sites wholesale, which I know there’s some tools that already do that, but it’s going to be built into Safari and other Apple tools. There’s actually legal action being filed against the company that’s trying to make it harder to place cookies on computers.  

Advertising and Google’s Privacy Sandbox

Aaron: I wasn’t aware of that. I know there’s been quite a lot of action against Google related to their Privacy Sandbox, because it’s such naked self-dealing. 

Ian: Yeah. It’s interesting. They’re saying no third-party cookies, but of course, they own so much of the advertising world.  

Aaron: Oh, they get to know everything. You get to know nothing, only what they would give you. Have you looked at the Privacy Sandbox much? No.  

Aaron: Oh, it’s incredible. I’ll give you a couple of examples. Attribution modeling or even conversion reporting, instead of giving you accurate attribution modeling, either single touch or multi-touch, they will instead tell you that transactions occurred, and they will tell you that a number of transactions occurred within a specific financial range. The range might be $0 to $500. There were ten transactions within this range that occurred within this period. You don’t get to know specifically when or specifically how much, and then on top of that, they will salt the data with deliberately inaccurate information, so that there is absolutely no chance of specificity or tracking back to any individual transaction. In a multi-touch context, it’s even worse. 

Ian: It’s even worse. I was just going to say, so now, the thing that you could do Google Analytics that got clients to actually believe in things like owned media because you could show it was organic search, social media, direct, email, that’s going to gone. Yeah, great. Thank you for making me feel that much worse about the state of marketing.  

Aaron: It’s incredible. There was an analysis that was done by, oh, I’m going to forget. It’s an acronym organization. I’ll add it in the show notes. They did a study, a technical assessment, of the Privacy Sandbox, and they did it specifically with regard to advertising and typical use cases today They tested 45 use cases, and of the 45 use cases, there were three that remained viable, and all of the rest were not possible, either in whole or in part, unless you’re within the walled garden of the Google ecosystem. Even then, you don’t get the data that you would typically get. 

Ian: They say, “Trust us.”  

Aaron: It’s like PMax. Just let us do it all for you, and write us a check. It’ll be fine.  

Ian: Exactly. Yeah. It doesn’t surprise me. Again, we’re going to have to be more aggressive about it on the media. When I started, back in the day, there were still companies that made all of their money doing direct mail, right? That was what they did. They had these enormous databases, and they’re very good at merging data and doing things like that. We’re going to have to get very, very good at doing that without pissing people off. We can’t do it the old direct mail way of we get a little bit of information here, here, here, put it together, and now we start spamming you. We have to get very good at building opt-in and double opt-in lists. It’s not great. For the next few years, for those of us who have not become very good at it, you got a few years, but you better get going.  

Aaron: It really, as with many things that we’ve talked about today, it’s a return to the fundamentals of marketing. Actually get permission to communicate with someone who wants your information. 

Ian: And do it with really great, compelling creative that can’t be generated by AI. Tell a really fantastic story that teaches and sells at the same time. If you do that, you will start to build that home list, and you will at least get an earned media following across many channels. Yeah, social and search and all those are earned, and we can lose them at any time, but if you’re on lots of them, they’re probably not all going to turn into dumpster fires at the same time.  

Aaron: Elon Musk can’t buy them all at once.  

Ian: That’s right. Elon Musk can’t buy everything. Google can. Well, they could buy everything. Diversification is important anyway. Really, I cannot say this enough – in fact, I just wrapped up a semester of teaching, and the last thing I said to all my students was – just remember everything you do, whenever you’re doing any marketing, you need to think about how you can turn some of these users, this audience, into an owned audience. Always be thinking about that. I’m not just talking about, “Subscribe to our newsletter,” you’re going to have to get more creative than that.  

Creativity, Communications, and Ethical Marketing

Aaron: Yeah, I think that’s excellent advice. What haven’t we talked about that we should cover?  

Ian: There is a whole thing about the level of discourse, and I have a very strong belief that marketing is such a fundamental part of discourse in a capitalist society. We are so responsible for that level of discourse. I look at how much the level of discourse has decayed in the last 4, 8 years, or, I guess, 12 actually. I think we all have to take some responsibility for that, as marketers. I don’t mean, we have to own what happened. I mean, we have to own how we’re going to contribute to a better level of discourse, and that is very high-level, and I know it’s very idealistic. That’s one thing that I don’t talk about that often at this point. Marketing is communications, and communications can save the world. I have always said that.  

Aaron: That resonates with me. Alright, let’s dig into that a little bit more. What are the practical implications of that sort of an ethos?  

Ian: Well as marketers, the implication is you need to do things that don’t make you feel like you’re going to need to take a shower every two hours. The way you communicate about products, the way you communicate about issues and items, it needs to be clear and it needs to be done ethically. Ethics can be different from person to person and society to society, but it basically means you can’t be catering to worst possible impulses all the time. Manufacture scarcity is not a great marketing technique when it comes to how you improve the level of discourse. If you say, I have this thing, and I’m selling it, and it’s a digital thing, and it’s made of bits, but I’m only selling ten because there’s a limited number, that makes people cynical. That’s probably the best way to put it.  

Aaron: You seem to be calling into question the whole NFT market.  

Ian: Imagine, and look at how well it’s going. That’s the best way to put it. We have to think of marketing not as a cynical exercise, but as a really good communications exercise because when you market cynically, people become cynical. When it comes cynical, they don’t trust the information that is put in front of them, which means the level of discourse descends because you tend to gravitate towards the things that you agree with, and you tend to ignore everything else and just say, “Well that’s bullshit.” In a very practical sense, if you have a product if you have a glass that you’re selling, you could say, “This is a fantastic glass, and all the other glasses have lead, and they’re going to kill you. Our glass won’t break and won’t do terrible things to you.” That’s kind of cynical because anyone who hears that knows that’s not true, right? But, they’ll be like, “The price is good, so I’m going to buy it,” and you encourage a brand, and it will do more and more of that. Plus, whether you believe it or not, you’ve absorbed some of that communication. If on the other hand, I take this glass, I say, “This is a really nice glass. It’s got a great feel to it. It’s better than our competitors because it’s clear and it’s more solid and it keeps things a little bit colder a little longer because it’s not really thin,” that’s really good marketing. The truth is, it’s going to appeal the same way to the same audience. It’s not going to terrify them into buying a glass. They’re still going to look at the price, and it doesn’t make me cynical because you’re not saying anything that feels like you’re pushing the bounds of reality, right? It feels pretty straight up. It’s not yelling at me, “Buy this glass. It’s the last glass you’ll ever need,” or anything like that. That’s a very simple description of it, but hopefully that helps.  

Aaron: I think that makes sense. I think, again, as with many things we’ve talked about, it’s the difference between a short-term mercenary strategy and a longer-term strategy where you’re earning a relationship with a customer.  

Ian: Yeah. Earning and owning a relationship. If I come back to cynicism and not fomenting cynicism, when people walk out of the store, or they’re finished checking out, you don’t want them to be like, “Thank God that’s over. Give me my product.” If you can do marketing that doesn’t leave people feeling that way, you’re going to have a better brand, you’re going to sell more stuff in the long run, and you’re going to improve the level of discourse. You can have your cake and eat it too. I mean, if you do the things that I’m talking about, that will improve the level of discourse in marketing and communications in general, you’re also going to do better as a business. No one can point to a company that lasted a long time, that is very strong, that doesn’t do that. Don’t tell me Elon Musk because right now we’re watching Tesla circle the drain. A good product, I think, that is now circling the drain simply because their marketing. They make marketing claims and say things that they have no business saying. How far did they set back self-driving cars? That’s a great example of just gouging out the eyes at the level of discourse. 

Aaron: Well, Twitter’s not doing so hot either. 

Ian: Well, X, he just let the lowest common denominator take over. I’m not sure how I quantify the rise and fall of Twitter. I’m not sure where that fits into this.  

Aaron: I do think what is true, an implication of the ethos that you’re talking about, is that if you seek, first and foremost, to be fundamentally helpful and generous with the information, the expertise that you have, that has the promise of changing the nature of the discourse. If you build a brand, if you build a relationship, based on positive interactions, rather than antagonism, that’s much healthier and much longer lasting.  

Ian: No one’s perfect, by the way. There are lots of brands to which I’m very loyal that have still pissed me off many times. I’m a big cyclist. I ride bikes, for the most part, made by a company called Specialized, and I love their products. They’ve always done a very good job of pointing me to the right product. Even when they’re trying to get me to buy a $10,000 bicycle, because I’m old and you can do that when your kids are out of college, they’re guiding me towards that bicycle, but they still do a very good job of saying, “You want this one, not that one, because this is the way you ride.” That feels helpful, and it’s very enjoyable when I take my bike to a shop, a specialized dealer, they’re super helpful. Somehow I always walk out having spent $300 on something, but I would never say I regret buying their products because they are super helpful and very positive. They sell the crap out of their products. 

Aaron: I think one of the things that I have admired about you – I’ve been aware of you for a long time because I’m in digital marketing – is that you have consistently been helpful to people and generous with your knowledge, and you’ve done it in a way that is not always typical in our industry. It hasn’t been self-aggrandizing. I haven’t gotten the sense that your chest thumping and trying to build a personal brand. I’ve had the sense that you have strong opinions and knowledge, and you want to share them. 

Ian: I suck at building a personal brand. That might be another way to put it. It’s interesting. 

Aaron: But you have one. I don’t think you suck at it.  

Ian: Well, I have one, yes. Question is whether it’s the right personal brand. It’s interesting because I have occasionally met people at conferences who are talking to me and they’re like, “I was so scared to walk up to you and say ‘Hi,’” and I was like, “Oh my God, I feel so bad.” That is not what I’m trying to do. I’m a cynical, New Jersey Jew, it’s just how I communicate with the world. I am so sorry. I want to help, so I appreciate that. I get a big ego boost and thrill out of teaching people stuff. That is part of it, too. There’s a lot of self-interest in it for me, but it’s just kind of my style.  

Aaron: Yeah, I think that’s great. It’s a good style.  

Ian: Thanks.  

Aaron: I’ve enjoyed the conversation.  

Ian: Same. That was really fun. 

Related Links  

Former Portent Employees (Free Agents)

Google Panda algorithm

Leaked Google Search API documents  

IAB Tech Lab Privacy Sandbox Report  

Navigating the Impacts of Google’s Privacy Sandbox on Digital Advertising

Questions or Comments

Please let us know if you have questions or comments about this episode by emailing Grace Johnson at grace@wheelhousedmg.com. Want to be a guest on a future episode? Fill out the Be a Guest form at the top of the Digital Clinic page to submit your inquiry.

Make sure to subscribe to the Digital Clinic on Spotify, Apple Podcasts, or Amazon Music to stay up to date on our weekly episodes.