Got an ecommerce challenge? Awin has you covered. With Awin’s affiliate platform, brands of all sizes can unlock endless marketing opportunities, reach consumers everywhere, and choose partners that fit their goals. Control costs, customize programs, and drive real results. Learn more at awin.com/emarketer.

Marcus Johnson (00:00):

Unlock more growth with Awin. Tackle acquisition, conversions, and retention by tapping into a network of over a million affiliate partners. That's right, a million. With everything from partner management to reporting and payments in one easy dashboard. The best kind of dashboards are easy. Awin helps brands drive real scalable results. Visit awin.com/emarketer to learn more or just for fun. It's up to you.

(00:36):

Hey, gang. It's August. Nope, it is August, but it's also Monday, the 25th of August. Grace, Drew and listeners, welcome to Behind the Numbers, an EMARKETER video podcast made possible by Awin. I'm Marcus and joining me today for today's conversation, we start with our tech and AI analyst, living in California is Grace Harmon.

Grace Harmon (00:56):

Hi, Marcus. Thanks for having me.

Marcus Johnson (00:57):

Of course, of course. And special guest, the CEO and founder of CMO Huddles and host of Renegade Marketers Unite podcast based in New York, it's Drew Neisser. Welcome to the show, Drew.

Drew Neisser (01:09):

Hey, Marcus. Thanks for having me.

Marcus Johnson (01:11):

Of course, of course. We start when we have a special guest on with a speed intro. So two questions for Drew, one for Grace. First one just for Drew is what do you do in a sentence?

Drew Neisser (01:23):

Easy. I help B2B CMOs via community coaching and content, and that community is CMO Huddles.

Marcus Johnson (01:31):

Very nice. And second question, we'll start with you, Drew, what was one of your favorite movies from childhood?

Drew Neisser (01:37):

So I was a little bit of an old movie geek and there's a movie from the '40s called Sullivan's Travels and it was about a film director who goes undercover and ends up in prison. And what he discovers is that at the end of the day, it's all about making people laugh and bringing a little joy to people's lives. And that stuck with me for a long, long time.

Marcus Johnson (01:59):

I like it. Did you have it on VHS or was this one you used to watch on...

Drew Neisser (02:04):

I watched it on television and then I owned a VHS copy of it. Yes, for those folks who remember that.

Marcus Johnson (02:11):

Very nice. I have my favorite movie of all time in VHS, Nightmare Before Christmas. It's just the best. Tim Burton and the Danny Elfman soundtrack. Grace, how about you?

Grace Harmon (02:22):

I liked Corpse Bride and Finding Nemo. So a little animated.

Marcus Johnson (02:27):

Good choices. Well played. Well, they're the folks we have for you today. Today's real topic, the internet's business model is breaking thanks to AI. Now what's one? All right, folks, so today we're talking about how AI is changing search and the internet's business model. Pew Research Center just did a study about AI summaries and as you can see from this chart on the screen, it found that when people see an AI overview, which are Google's results, they're half as likely to ever click a link from Google. You also found that after folks see an AI overview answer, they're more likely to end their browsing sessions. Drew, what do you make of this research? What was your biggest takeaway?

Drew Neisser (03:18):

Well, first I think it actually understates the truth rather than overstates the truth. I think that there's less clicking than they're even seeing, and you can see that in the amount of site traffic that is declined to so many different websites. So I don't think there's any doubt that gen AI and LLMs are changing organic site traffic. If that was your business model, you certainly need a new one.

Marcus Johnson (03:45):

Yeah, Grace, many AI searches, they're just fulfilling people's curiosity without the need to click into Drew's point. CNN one such example, traffic down 30% year on year, HuffPost down closer to 40%, both that number from a similar web. What's your take on this study? I think you wrote a piece about it, I believe

Grace Harmon (04:05):

I did, yeah, and I think it's exactly what you said, which is that these options are pushing people out of search really quickly and that changes everything for brands and for publishers. I mean in addition to AI overview, you have Perplexity's AI search, all of these different engines that are creating content, but ending discovery. So it absolutely changes search behavior.

Marcus Johnson (04:24):

Yeah, there seems like there's just a mounting evidence that this is having a significant effect. Marissa Jones who writes for one of our newsletters was saying, "Ahrefs study found AI overviews, decreased clickthrough rates by over 34%. There was a C Interactive study noticing a 70% drop in organic clickthrough rates in search with AI overviews." She also points to authorities cited from The Guardian found sites previously ranked first in search results could lose nearly 80% of traffic for results that fall below AI overviews.

(04:59):

And then similar again, they measured about a hundred million web domains and estimate that worldwide search traffic by humans fell 15% in the year to June. Certain sites, health ones in particular closer to 30. Drew, in AI's defense though, I mean, Google does say they say clickthrough volume from the search engine has remained, as they say, relatively stable year over year. Though some types of websites are getting more clicks, others getting less. You could argue it's not all AI's fault. There are other reasons that folks are visiting sites, less, social media listening to more podcasts, things like that.

Drew Neisser (05:34):

Yeah, I think that's fair. First of all, this is going to take longer than you think. It's not going to be like suddenly there's going to be no organic side traffic and suddenly... I mean, I remember working on a brand a few years ago that was still selling typewriters. People actually bought a million typewriters every year, long after computers were developed. So it takes a long time for the... And people... For the first time, I think I saw the Google search usage went below 90% their share, but it's like 89. Right?

Marcus Johnson (06:10):

Yeah.

Drew Neisser (06:10):

I mean, they're still the dominant starting place. So that's part one of this is consumer behavior changes slowly over time. And so, yes, AI has accelerated, which was probably something. We're spending more time on our phones. We're spending more time on TikTok. And there's a reason why every single platform is trying to do vertical videos because people are on mobile devices. So I think, yes. And LLMs are killing organic site traffic.

Marcus Johnson (06:44):

So what's the advice or what's some of the conversations that you're having with the CMO community with regards to this? Because LLMs, large language models versus browsers, it is that LLMs are giving you one answer, browsers give you a list, and so there's fewer chances to get in front of customers. But like you said, things are happening slowly but surely. So what are CMOs talking about when it comes to this conversation?

Drew Neisser (07:11):

Oh, gosh. So there's a lot of angles to answer this, but number one is when your organic search traffic goes down, you start to think... And people are on LLMs, how the heck do you get your brand to show up on an LLM? There's a lot of conversation about whether you call it AEO or GEO, and a lot of different folks experimenting with this and we're seeing a lot in that area. And I can go list some of the things that they're trying to do, but here's the one thing that we know so far is that nobody has a clear answer on solving this problem because of the way LLMs work.

(07:48):

If you go in and ask it a question today and you ask it once and then you ask it again, it'll give you a different answer. You have to do a thousand queries to actually get to a percentage of where 70% of the answers look like the same. So it's so different than SEO tracking. It is really hard, so it's very difficult right now, but everybody wants the answer to this question, how do we make AEO work the way SEO did?

Marcus Johnson (08:16):

Yeah, I mean, Grace, it's difficult because AI overview is appearing I think in about 20% of results. So it's significant, but it's not everywhere. And Eli Goodman, chief executive and co-founder of Datos was saying that over 90% of all AI searches are informational or productivity-based rather than you looking for links type search. What do you make of this? What seems to be transition but slow transition towards LLMs?

Grace Harmon (08:49):

Yeah, I mean I guess going back a little bit, Google did contest that Pew Research study, but I don't think... It's undeniable that Google's or AI overviews and AI are changing how people engage with search. And with instructional content that Pew Research Study found that W questions who, what, where, when, why, how those trigger a lot more AI overviews.

(09:10):

I think that you also end up seeing less diversity of answers. So a lot of the time AI overview is generating links from Wikipedia, often from Reddit, but you're not seeing necessarily all of those smaller publishers or smaller brands, those smaller webpages. Yeah, I mean, SEO and GEO work very, very differently. For one thing, we don't necessarily understand everything about generative engines and how they serve as content.

(09:33):

So there isn't really a perfect strategy like there is with SEO. It isn't just relying on keywords. It's about contextual recall. And like Drew said, you can put in a thousand different queries or a thousand different times the same query and you can get a thousand different results. I don't think that there's a great answer for how to perfectly appear. I think what we've written a couple times is websites might want to add things like FAQ pages, really specific product specs. Adding summaries I think is the big thing for what those engines can take in, what they'll scrape and what they'll regurgitate later.

Drew Neisser (10:08):

There's more to that though. I do think there are probably seven different things that you can do right now. You don't know how much it's going to work, but I will tell you that a number of brands are already seeing the result. They're seeing inbound traffic as a result of an LLM referral. So that's happening too. Brands are getting discovered, but certainly nowhere near the way they would've from SEO.

(10:31):

There is the Q&A's and extensive Q&A's. There is having the basic information that an LLM wants. Do you have your competitive information? Do you have your pricing information? There's certain code that you can put on there that the... And it is funny because early days of SEO, it was about code and trying to trick it with code, but now there is literally pages that you could probably put up that will be readable by the LLM that no human will ever see.

Marcus Johnson (11:02):

Let's talk a bit about this shift in how the internet did work, the bargain that was the promise of the internet and how that's changing. AI is killing the web. Can anything save it? It was an Economist's article I was looking at for this episode and The Economist is writing the AI search can slow traffic to websites, which in turn removes the incentive for content to be created, which alters the balance of the open web as we've been discussing. As AI changes, how people browse is altering the economic bargain at the heart of the internet.

(11:37):

The piece was saying, "Human traffic has long been monetized using online ads. Now that traffic is drying up. Content producers are urgently trying to find new ways to make AI companies pay them for information. If they can't, the open web may evolve into something very different, but what might that look like?" The Economist had two examples, one, saying some folks like TollBit as a paywall for bots proposed a pay as you crawl system in which AI bots are charged for reading site's content.

(12:08):

And then the second one here, others, people like ProRata are working on systems that analyze chatbots answers to determine where the information came from so that sources can be compensated kind of a credit sharing model so to speak. Drew, what do you make of these?

Drew Neisser (12:22):

So I'm a big fan of The Economist and I generally really, really respect... I think I disagree with the premise fundamentally.

Marcus Johnson (12:32):

Tell me about it.

Drew Neisser (12:33):

Number one, part of it is the incentive to create high quality content is more important than ever because that's the only way you'll be able to cut through the slop. Think of The Economist. Are they going to stop creating content because suddenly there's less organic? No, they're not. And by the way, The Economist is a premium brand known for high quality content and it's a subscription based model.

(12:55):

The notion of free content is the problem here. But if you are creating content that is good enough that your subscribers will pay for it, in a world where there's so much distrust, I think the reputation of the content provider is never going to be more important and therefore because there's so much more fake and garbage out there, people are going to seek sources they trust and they'll pay a certain amount of money to get that content as it can.

Marcus Johnson (13:22):

Yeah, Grace, you actually wrote about this saying AI search has trust issues. You're saying just 9% of people fully trust AI overviews. You say 1% of users click on a source link in an AI summary, that's from the Pew Research study. However, over 70% said Google Search is either the same or better than before AI overviews launched according to exploding topics. So it does seem like there's a bit of a paradox here. Similar to how we see with a lot of things, people don't trust news on social media, but they use it all the time.

Grace Harmon (13:52):

Yeah. No, I think that's contradiction there in terms of how much people trust AI versus how much work they're willing to do. And to a degree it's kind of hard. AI overviews doesn't always surface the links to what exactly it's talking about, but I agree with Drew that brand is a really big deal. I actually did have a question for you Drew really quick, which is Google had rolled out preferred sources. You can start developing your Google search results to be more like an algorithmic feed where you're selecting exactly what you want to show up, which I think limits discovery to a big degree.

(14:25):

I was wondering if there's any brands or CMOs that you're working with that are having more interest in these indie publisher sites like Substack, like Medium, that have a really big focus on discovery?

Drew Neisser (14:36):

I think it's still early days when it comes... And again, our community is all B2B so you're going to see I think more experimentation in B2C initially than you will in B2B. Suddenly Reddit is on everybody's radar and they're all trying to figure out, "Oh gee, how do we play in that space?" And that's obviously really hard. So I don't know if I have a good answer for you on that one.

Grace Harmon (15:06):

And Reddit is so huge. I mean for the amount of community building and niche communities and just built up community trust there is in terms of you're getting a product recommendation. I know I'm talking more about B2C there, but if you're getting recommendations from your community there, I think it'd be a lot more trusted than what you're getting from Google.

Drew Neisser (15:29):

And by the way, I think that's a major problem for the LLMs if people don't trust them and if it drags Google down because I think there was... The interesting deal, the bargain you had originally with... And I really wonder what Google is thinking because if I were them I would've said, "There's still Google it exists and there's a new LLM product." But instead they're trying to do both and you can't ever do that.

(15:53):

So it's a business issue for them. Obviously they have to defend, but they had this cash cow that they've decided to use and it'll be really interesting to see how their ad revenue performs over the next five years.

Marcus Johnson (16:10):

Yeah. A lot of folks feeling differently about Google to a certain extent. Neil Vogel, head of Dotdash Meredith, and people who want food and wine and things like that were saying, "We had a very positive relationship with Google for a long time. They broke the deal." Grace, what do you make of those two models of how the internet might work, the pay-as-you-crawl system for AI bots and then the credit sharing model?

Grace Harmon (16:36):

I mean, I guess the base, I'd just say that I think media publishers need to be compensated for when their content is used for training or on answers, but I think it gets really complicated because there's only so much control you have. I mean, you're talking about a pay-as-you-crawl model. How enforceable is that? There's a lot of websites that have been trying to put code onto their sites, instructing bots on what they can and can't scrape, but there's only so much you can do. So I think it's great in theory.

Marcus Johnson (17:05):

As with a lot of things, we were talking a bit about AI search, having trust issues, and there was an Atlantic article from Emma Maris questioning whether AI slop might finally cure our internet addiction as chatbots make much of the web unreliable in turn nudging folks offline. She points out that YouTube long destination for real people who know how to repair toilets, make omelets or deliver engaging cultural criticism is getting less human by the day.

(17:31):

As the newsletter, Garbage Day reports that four out of May's top 10 YouTube channels were devoted to AI-generated content, and recently the fastest growing channel featured AI babies in dangerous situations for some reason she says. Reddit is currently overrun with AI generated posts. Many AI photos and videos have become indistinguishable from reality, making it much harder to trust anything online. Grace, how much of a chance is there that AI slop will nudge folks offline?

Grace Harmon (17:57):

I think a good chance, I mean it erodes user trust. It erodes platform utility and a huge amount of it. It's like SEO spam times a hundred. I think that it will nudge some people offline. I think it'll make people have less trust in what they're seeing. You're talking about what percentage of internet users think that AI has made the internet a better place. I don't think that stat is going to stay the same forever.

(18:22):

I think it also might spark more of a slow web revival, so niche communities online, newsletters, not to bring Substack up again, but Substack, Zines, things like that. So I think that you might see people nudged offline and also just away from what these legacy platforms that are getting more inundated with AI.

Marcus Johnson (18:43):

Yeah.

Drew Neisser (18:45):

I'm not sure, I agree with the very premise that people are going to change behavior and move offline because there's so much slop. I do think... And when you look at the YouTube baby things, I mean, that is a moment in time for YouTube babies. Let's face it. They are almost irresistible. You just can't not watch them. And in fact, they're addictive and that's why these things are kind of winning. But it's like saying it's really hard if it is an addiction of sorts to suddenly go cold turkey.

(19:23):

I respect that Atlantic article and how they personally felt it was driving them, but I suspect, again, I don't think you'll see a major change in what I hope will happen is that people will seek out quality that people will seek out if they'll start to seek out human made and that there will be actually this sort of reverse. With every trend, there's a counter trend and as Grace talked about, that while people are going to the AI slop, they more will seek this not AI, real human stuff. But that doesn't necessarily mean it'll happen offline.

Grace Harmon (19:58):

Yeah.

Drew Neisser (19:59):

Right.

Marcus Johnson (20:00):

Yeah. There was the point in the article saying we might see a bifurcation. The internet's new era may push AI skeptics to spend less time online while another group ramps up that AI mediated screen time, which could split the internet's culture or culture at large.

Drew Neisser (20:17):

And that's really concerning.

Marcus Johnson (20:19):

Yes.

Drew Neisser (20:19):

Because I mean, there's a really good chance that a lot of people, if you miss the AI wave, it's like if you might've had a grandparent, a great grandparent who said, "I'm never going to be on email or I'm not going to text and they can't communicate. If you ignore these tools in your personal lives and in your professional lives, you're putting yourself at risk.

Marcus Johnson (20:45):

A lot of it could be AI by proxy as well. Another point in the piece was that even if you never use something like ChatGPT, the rest of the internet is loaded with the stuff and people parroting their hallucinations it was saying. So you might get it indirectly anyway.

Drew Neisser (21:02):

Right. Really, the hardest part of all this is wherever you go, it's so difficult to know what's AI and what isn't.

Marcus Johnson (21:11):

Yes.

Grace Harmon (21:13):

There's a couple of examples in that Atlantic article that I liked. I think one that came to mind for me was this video that went viral on TikTok of a bunch of rabbits jumping on a trampoline and no one could realize that it was AI and people were saying they needed to go touch grass after they realized, or the example in that article of getting messages on a dating app and it feeling formatted by ChatGPT.

(21:33):

So I think it's unlikely that all of that means no one is going to spend time online, but it's just really going to change how they engage the internet. I also just don't think that that means that everyone is going to be taking full minutes to scrutinize everything they see.

Drew Neisser (21:50):

Yeah. So I come back to where... Are there going to be channels? If the New York Times had a video feed, I would say, "Oh, I trust that because a brand that I trust, or The Economist did. So if suddenly we had social places where the content was branded or under an umbrella brand that you trust, we're going to feel good. And there was a moment where social networks were going that way. They're not anymore. So now maybe there's a big opportunity for social networks where the content is actually a hundred percent screened.

Marcus Johnson (22:27):

Yeah. The content is really going to matter here.

Drew Neisser (22:30):

It really I.s

Marcus Johnson (22:30):

Newsletters you trust, paywalls where the content you're getting is high quality from a trusted brand. Drew, you speak to a ton of CMOs about a lot of things, but this I'm sure in particular. What would be one takeaway or one take from those conversations that you'd like to share with the listeners on this?

Drew Neisser (22:50):

Well, overall, I would say that we are dealing probably with the biggest transformation in marketing that certainly in my lifetime, and the CMOs are feeling that, and all of course because of gen AI, fundamental questions like, "What's my staff look like today and tomorrow? When am I going to have agents on my org chart? Do I need junior people anymore or do I not? Maybe I have a bell curve and or barbel and I only have people that are really senior or really junior." There's a lot of questions about people in that. There's a lot of questions about process. And then finally what we've already talked about how if SEO doesn't work anymore, then what does AEO... How do you get discovered? How is their brain going to get discovered in LLM? And that is a constant subject of conversation.

Marcus Johnson (23:41):

Grace, you've been writing a lot about this and I'm sure we'll continue to. What would be one take that you would share with brands who are listening?

Grace Harmon (23:51):

I think that for one thing, AI search isn't going anywhere. How it works might become less mystified, but I would agree with what Drew was saying earlier that GEO, really figuring out those strategies is going to be key because if you want brand discovery, if you want communication, if you want to establish a relationship, part of that means appearing in queries.

Marcus Johnson (24:10):

Yeah. Well, Drew is host of, as we said, Renegade Marketers Unite podcast, so check that out. That is all we have time for today's episode. Thank you so, so much to my guests. Thank you first to Grace.

Grace Harmon (24:22):

Thanks for having me.

Marcus Johnson (24:23):

And of course to Drew.

Drew Neisser (24:24):

Marcus, thanks. Great conversation.

Marcus Johnson (24:26):

Absolute pleasure. And thank you to the whole editing crew and to everyone for listening into Behind the Numbers, an EMARKETER video podcast made possible by Awin. Tune in Wednesday to hang out with Arielle as she guest hosts the Reimagined Retail show where [inaudible 00:24:39] presents its most interesting retailers of the month, which is August list.