x
READY TO BECOME A MEMBER?
Stay up to date on the digital shelf.
x
THANK YOU!
READY TO BECOME A MEMBER?
THANK YOU!
Transcript
Our transcripts are generated by AI. Please excuse any typos and if you have any specific questions please email info@digitalshelfinstitute.org.
Welcome to unpacking the digital shelf where we explore brand manufacturing in the digital age. Hey everyone. Peter Crosby here from the Digital Shelf Institute in the 2024 holiday shopping season use of AI chatbot shopping rose 6000% over the prior year according to Adobe. Those are searches driven by the intent of the shopper, and many of those queries never even make it to a product page. Winning your place in those shopping journeys will require mastering the new capabilities of generative engine optimization or GEO Max Sinclair, founder and CEO at eCom joined Lauren Livak Gilbert and me to provide a primer of what's changing and what's possible to start seizing the opportunity faster than your competitors early movers. Advantages at stake and the answer won't be more keywords. Max, welcome to the podcast. We are so excited to have you here today. Thanks so much for joining us.
Max Sinclair (01:07):
Thank you. It is a pleasure to be here.
Peter Crosby (01:09):
Well, just when I thought I knew what SEO was all about, all of a sudden there's a whole new thing, a whole new set of capabilities and knowledge and something that's feeding the consumer experience that our customers and listeners have to be able to wrap their arms around and that's generative engine optimization. How do you optimize for the agents out there that are the fuel of the mission-based shopper in some chatbot somewhere and you spend a lot of your days diving into how all that works and how the algorithms and where this new world will come to life, and so how it's going to disrupt shopping and then of course the lives of our listeners. So first of all, we'd love to talk just some of the shifts you are seeing out there in the world of search and shopping.
Max Sinclair (02:06):
Sure. So 60% of US adults who've used an AI chat bot in the last 30 days, so that is chat GPT, that is roofers or plexity or some other AI search engine chat. GPT is now getting a billion messages a day for reference, Google gets 8.5 billion searches a day. So of course Google is ahead, but I mean the catch up is staggering that they've closed that gap from zero in two, three years. And what is also very interesting to me is how people shop is changing. So 50% of perplexity users ask a follow-up question. We don't have data on chatt, PT or roofers, but you can imagine it would be similar. And additionally, the questions are getting longer. So the average perplexity query is between 10 and 11 words compared to Google, which is two or three on average. So these are definitely getting more traction and it's also changing how we all are interacting and searching and finding information.
Peter Crosby (03:17):
I love what that says. I mean, first of all, getting 60% of US adults to do anything, it's amazing. So when you think of the rapid nature of the growth of the use of these tools, whether in some cases the consumer knows that's what they're doing or not, is a really impressive number. And then I love the follow-up questioning that really goes to the center of what this is about, which is the ability of a consumer to go on a mission and get the answers they need to make a decision as quickly as possible. So within these stats are really kind of the storytelling of how the world is changing right now.
Max Sinclair (03:58):
Absolutely. I think everyone listening I imagine has probably done this themselves, so they can easily imagine how you would do some research on chat GPT asking a quick question. It's probably a daily usage I imagine for most of your listeners, and that extends beyond those who are plugged into technology and it's really widespread now across the population. So it's definitely a shift that sellers, retailers, brands, agencies need to be thinking about today because we've already, we're at 20, 25 now. We're at that point of massive adoption beyond when this all launched and was kind of the hype in 2022. I think we've arrived. And another interesting stat I'll just throw out there because I'm enjoying the stats. So do you guys know how long it took to get a million users on the internet
Peter Crosby (04:53):
Months?
Max Sinclair (04:55):
20 years is the answer. I was going to say years. And do you know how long it took to get a million users onto chat? GPTA
Lauren Livak Gilbert (05:03):
Couple hours,
Max Sinclair (05:05):
Five days. So the adoption curve here is unimaginably faster and then the previous technology shifts we've seen and it's one to the people need to be thinking about
Lauren Livak Gilbert (05:19):
And there's no one dominant player right now, right? Like a chat GPT versus a Rufuss versus a perplexity versus Google has been air, the dominant search engine for years, but now it's dispersed across many different types of LLMs. Would you say that or do you think there's a clear
Max Sinclair (05:37):
Leader? I think this is great because I am generally pro competition and I think it brings innovation, but I think we're going to move to a place where we go to different engines because they're specialized for different things. So Amazon and Rufus obviously will be the best one to research products for. Some would argue that it already is. Some others would say, oh, roof is rubbish. I personally have bought multiple products from Rufus, so I'm a fan, but I mean it's clear that Amazon is going to win in this space, so there's no way they're going to give up that e-commerce one unless Shopify launched something drastic. Although interestingly, Shopify have done an integration with perplexity and that's the second one. And typically the perplexity user a kind of higher income tech inclined kind of folks. So I don't know if your users have come across it yet, but it's probably the best one for research and just understanding information.
(06:41):
And then of course Google is unlikely to really go anywhere. They already have AI views and we can imagine that increasing and then you can say, well what about copilot for work or being in, I would imagine that other players will enter with very early, maybe with Apple and in some regard when they get Apple intelligence working properly, maybe that is going to feature for more personal stuff. Who knows? So I can imagine us going to a place where we are using different AI search engines for different needs and it means that as brands and retailers and agencies need to be thinking about all of these different search engines now, and it's not as easy as just being like, we're going to optimize for Google. And then we're kind of done
Peter Crosby (07:32):
Though. I wonder, I dunno about you all. I struggle with the number of stupid streaming services. I have the thought of me also needing to remember which engine I go to, which thing I don't know, I feel like, but you pay for those
Lauren Livak Gilbert (07:49):
Services, right? Peter?
Peter Crosby (07:51):
Shut up. Yes. However I'm sure I was going besides my ridiculous, I
Max Sinclair (08:01):
Agree that the streaming is very annoying and especially for the UK listeners if you have some, we'll know the Premier League is about on six different channels including Amazon now. So it's impossible to legally watch the Premier League in full unless you're paying thousands of dollars to Sky and to Amazon. And so it is definitely annoying. But I think when it comes to search, I think it's a little bit different. Obviously I'm in the very much the early adopter mindset, but I'm a paid subscriber to chat GBT and Perplexity and Claude and I use them. I see the different engines are good for different things. So Claude is less good for searching but is very good for writing blogs and articles and it has a much more human sounding voice even though it's not connected to the internet. So not very helpful for search. So I already see this kind of in my own behavior having these different uses and I think it can be helpful because you know what they're good for and you're happy to go to different ones for them. Like shops, right? Apart from Amazon, right online in the real world, you're going to go to different shops for specialized things. You don't want to buy everything in one shop necessarily.
Lauren Livak Gilbert (09:27):
And Max, there's a general sense of anxiety around AI making choices for consumers, writing blogs, having the wrong tone, saying the wrong thing. What is the implication with GEO and with AI agents and brands as they think about using this and integrating more into search and how products are found and how consumers are interacting?
Max Sinclair (09:50):
Yeah. Well, I'm going to put my tin hat on and I'm going to say that AI has been making decisions for us for years, as we all know, and I think in a much more problematic way than the anxiety that generative AI brings. So generative AI is obviously when you're creating new content like product listings or blogs or whatever you want to do there, I think there's a lot less, and I think they're very good, certainly content's very good, but I'm sure others are good as well at controlling for brand voice. And the rest of it, I don't think it's particularly hard. Yes, if you are using something out of chat GPT, it's not going to be perfect because it's not built to sound like your brand, but it's quite easy to build some infrastructure and to train the AI on a bunch of brand voice and make stuff in your brand voice when your brand visuals.
(10:46):
That's not challenging. But I think the bigger detriment of AI has been in the historical deterministic stuff, and particularly with social media and the algorithms there and how they literally shift our thinking. And I think it's a big problem. TikTok and Twitter I think on different maybe political spectrums, you're doing different things, but they're definitely drip by drip influencing how we all think. And that to me is a much more present and dangerous implication of AI technology than generative ai, which I think is relatively benign. You're just creating a new product, listing a new blog, a new image of your product. And I think that's less scary really.
Peter Crosby (11:38):
I don't know whether I can rate the scare factor for each, but certainly we see forgotten who said this to me, but AI lies, but it lies with great ability or something like that. Just that the trust factor. Certainly when I think of our listeners, a lot of what they are trying to bring to market is the truth of their products and clear answers. And when you get that wrong, there's liability that comes with that. There's fines, there's consumer harm, there's brand. So when you're having non-human generate content, then there is the concern that that liability will somehow shift or so I think that's where the trust
Max Sinclair (12:31):
Issue. I think it's a fair question and I would counter that and I'll counter that with the example maybe of AI cars and Waymo. So people may be scared, and I spent some time in San Francisco to get in a Waymo conceptually no driver, but the reality is it's a lot safer in their marketing. They say we're the most experienced driver in the world, they are a lot safer. And actually if you think about getting into an Uber at midnight and it's a random person, you've no idea anything about them. And actually in many ways just because used to it, that is more scary. Are they a good driver? Certain I've had Uber drivers go the wrong way up one way roads and as
Lauren Livak Gilbert (13:21):
A female I agree, it's way better to be alone in the car.
Max Sinclair (13:25):
Yeah, there we go. And I think this translates directly into my world of product listing content, which is the, and again, I don't want to just talk about your content, but the AI we built is infinitely more intelligent than any one graphics designer or copywriter. It just is. It's got billions of data points, impressions, conversions, and it understands all that. And it's generating content. And yes, of course anything can make a mistake, but the error that maybe an individual kind of freelancer on Fiverr or whatever is going to make is going to be a lot worse. And the ai, especially when we're talking beyond just sticking something into chat team and generating it, of course it's got no context apart from what you put in the prompt. It's going to make some mistakes and it's not going to sound great. But if you are feeding it with all of the keywords that we're going to get you blocked on Amazon, for example, you're feeding it with many examples of the brand voice, it's going to do, in my opinion, a better job than an individual human and it's going to be more optimized and it's going to be a hell of a lot faster.
(14:31):
And I think it's a bit like getting in the Waymo compared to picking an Uber driver.
Peter Crosby (14:38):
I love that. I mean I think it really, one, it's inevitable. So we're going to have to figure this out somehow. And two, I appreciate the analogy. I think we sort of leapt into GEO without really defining it. And so maybe it's worthwhile, you just sort of what is it and how is it different the same than SEO, how related are they and how should our listeners think about that sort of practice, those set of capabilities that you need?
Max Sinclair (15:19):
Sure. So I mean as that GEO stands for generative engine optimization, I, and I dislike the term, I like the term because it's phonetically, it sounds like a progression from SEO. So I think it helps people kind of understand what it is, but I also dislike it because you are in that same way, you're kind of comparing it to SEO. And in my mind, at least with our customers, I encourage 'em to think of this as a new marketing channel. So you will have your paid ads, you'll have your social media, you will be doing SEO, and to me, this is a new marketing channel that rather than kind of giving this to your existing SEO and saying, Hey, here's, let's tactically tweak this and change that. To me this is a massive opportunity, especially now because not many people are thinking about it. And when I talk to a large brand and I say, what's your strategy on ranking on Chachi for T?
(16:20):
It's getting a billion messages a day and see most visited website of the world, what are you guys doing? They're not thinking about it. And obviously it's an entirely new marketing channel and I would encourage it to be the head of digital or the CMO or that kind of level in an organization to be thinking about it because it's a big opportunity. And another analogy that I give is we're kind of in this same era as the early two thousands when Google went from being a temp from Yahoo's volume to 10 times it, and we're in that same transformation as we discussed. I don't think it's all going to go to chat GPT, I think we're going to see fragmentation, but we are going to switch from everyone using traditional search to everyone using LLM based AI powered search in the coming months and year. And the analogy I give is kind of a bit like thinking about TripAdvisor. So TripAdvisor had an incredible SEO strategy in the early two thousands.
(17:32):
They did two things. Firstly, they made many, many landing pages. So if you kind of want to find restaurants in London or Toronto or Vancouver, wherever you are, they would have a landing page for that specific need, which is obviously good for SEO. And the second thing they did was they would give back links. And obviously with the page rank algorithm, you needed these back links. So they would give little badges to the restaurants and the hotels and say, here's your number of reviews. And it links back to TripAdvisor so the customer can read all the reviews and suddenly they've got millions of back links from all these hotels and restaurants and they become a 14 billion market company at their peak and they get 200 million in organic traffic a month. So we have this opportunity now as brands to be the early adopters like TripAdvisor, make really think strategically about how we're going to approach it and have enormous returns. And there's many of TripAdvisors competitors who are now gone out of business, started at the same time, started in San Francisco, travel tourists being one that nobody's ever heard of and obviously many factors in that. But one of them is that they didn't spot as TripAdvisor did this ability to see that Google was going to be massive and you could hack in inverted commas the system of Google by doing these two strategies and make a massive business for themselves massive.
Peter Crosby (19:05):
So if GEO is the next version of that, and our listeners are thinking, yes, I'd like to win that space simply because that's where so many of their shopping missions of their consumers are going to go. And if you are not prepared to show up there, you lose. So with that in mind, I mean over time it's not tomorrow, but to your point, as we know with SEO to the early adopters did go a lot of the spoils and early advantage tends to be lasting for a while. So it sounds like that might be the case with GEO. And if so, maybe starting with Rufus as an example, if we were talking to our brands about how they need to think about preparing their data to show up in that channel, how do you think about it and describe it for your customers?
Max Sinclair (19:59):
The beautiful thing about Rufuss and actually all of Amazon and anything to do with Amazon is that they release science papers. So you can go on Amazon science, you can go and find science papers on basically everything, Cosmo, rufuss, all of them. And the science papers break down in a lot of detail how these new search algorithm are working. So my co-founders a PhD in AI and a postdoc, so obviously it helps to have an understanding of how to read these, but literally any layman can go in and with the beauty of Claude or Google, I forget what it's called, workbooks or whatever they call it or chat, you can just put a science paper in and you can ask questions to the AI bots about it. So you don't have to be a PhD in ai. But in terms of specifically rufuss and the science paper there, they talk about five facets that you need in your listing.
(21:05):
They call these subjective properties. And these are, and the meta takeaway from this, by the way, is that it's all about context and keywords matter much less. And I can talk about the technical reasons why, but the context that they say in this roofer science paper you need is the event relevant, the activity relevant. So what they're doing, the gold of the product, why would someone buy it and who the target audience is. And actually just having that both in the written content source and the visual content because by the way, all these AI models are multimodal so they can understand images. And SEO now is also visual in GEO. So showing the relevant target audience and how the products used in an image or an infographic just doing this is going to help you to rank on Rufus.
Lauren Livak Gilbert (21:58):
And what do you think about chat GPT in the same regard? So rufuss is an opportunity for brands to make sure that their products are being shown and they need to use the right context to be able to do that. What about from a chat GPT perspective because maybe someone's not going to chat GPT to buy something, they're definitely going to Amazon and rufuss is there, so they might have more of a purchase intent. What about from a chat GPT perspective?
Max Sinclair (22:22):
So chat GPT is reverse of Amazon and ironically named OpenAI, nothing published about Amazon publishes. I said, great details, 300 plus science papers on search. What's happening? Well, you don't know what's happening, but you can piece together what's happening. And also if you have data, which I assume most of your customers will, you can make a pretty good inference of what's happening. This is what the data's showing, this is what the science waves are showing. Chat GT is a different kettle of fish and we've got a separate strategy or product for that. But talking again for just listeners who maybe aren't building advanced AI systems, number one piece of advice would be start by just asking chat GPT about your brand. So whatever your brand is, let's say eCom, put it in, you can see immediately what CHATT PT says and also the references where it's getting that information.
(23:22):
And that's really helpful because in my case for example, when I did this, I saw that they are pulling from articles which is like two, three years old and pulling information, which is up to date as I'd like my business to be represented. So by just simply sticking in your brand and working backwards from that is a good starting point to see what's there. Now number two I think is similar to the rufuss and the Cosmo stuff, which is it's all about intent-based copy. It's all about, and visuals, understanding the pain points, understanding the goal that your specific customer has when they buy your product and the specific target audience and generating content for that. Thirdly, you can do OpenAI has got a bunch of partnerships with major news organizations which they use to train their dataset. So we're talking the American journalism project, the Associated Press, the Atlantic buzzfeed, a bunch more, so slightly longer term than maybe just like some of the other ones.
(24:32):
But if you can get yourself featured there, you know that eventually chatt PT is using that in the trading data and it's going to help you over the long term. The final quick hack one, which I would recommend people doing would be to make a custom GPT. So whatever your brand is, and again, I'll give myself an example, I can't think of any, but we made a custom GPT about how to help sellers can put in their ASIN and see how they rank for Cosmo. Simple custom GPT is now got over a thousand reviews on the GPT store and it helps you are literally giving in the GPT, it puts the orphan it links to our website. So you are giving chat GPT context about what your company does. I don't know if you're selling a dog food, maybe you make a GPT about how healthy your dog is and different types of food, different types of dogs may need or whatever it is. But you can show how you can tell chat GPT that your brand exists and what you do and how you help customers by literally putting it in their system. And of course that's going to again help chat GPT understand a bit more about you
Peter Crosby (25:51):
And doesn't chat GPT make Bing matter again?
Max Sinclair (25:57):
Yes. Yeah. So they are, as you say, I missed that one. They are built on Bing with their Microsoft partnership. So it is actually very easy. You can just kind of copy your Google place of business over to Bing takes like three seconds. And then you are also showing up in Bing and that's also going to help you in chat GT as well.
Lauren Livak Gilbert (26:22):
And Max, do you have any examples of working with brands on either rufuss or chat GPT that you can share about how they really made a change and really kind of influenced how they were showing up on these lms?
Max Sinclair (26:35):
Absolutely. So I mean how much I can probably talk about it. Our largest customer is a tire and we will have a case study with them shortly and we've been helping them specifically on AI overviews and in Google ranking. But on the roofer side we have tons of case studies with let's say smaller Amazon sellers. These are just available on our website. But case studies showing how we've helped sellers to go from outside the top 50 in the top 25 in the bestseller rank. And we've got a number of these on the website and basically on the roofer side is implementing what I said before, generating visual and written content that addresses the, as they call it, subjective product needs of the customer. So generating a lifestyle image of the target audience. So for one customer who is selling cigar cases, you have a lifestyle image of someone smoking cigar cases, smoking cigars, and then infographics talking about, and you are testing me here, but if I'm guessing the moisture levels and whatever else you need in the infographics for cigar cases. So you generate the visual and written content that really explains to the AI as well as a customer who the product is for and what it does. And you can see this improvement in bestselling
Peter Crosby (28:13):
And that improvement comes from putting that content where content on the PD DP on the product list or that's it. Okay, I just wasn't sure if there are other places
Max Sinclair (28:22):
You wanted, you're making you printing a plus content, you're making infographics, you are updating the copy, the keyword, the keyword stuffing here. I think basically every series seller has realized this now is over Amazon have come out and said they're going to penalize you for repeating words in your title. So I was literally talking to a seller a few minutes ago and she was selling a vendor selling cooking things with, again, I'm not a chef, but why is the kind of tie together meat. So it can be done for lamb, it can be done for Turkey, it can be done for chicken, it is like a cooking wire package thing. And she was saying part of the reason she reached out to me, she's seen this drop and it's quite obvious to see why because you've got Turkey wire, steak wire, all these same thing over and over again across the listing. The AI is smarter than that and it's not rewarding that it's rather you talk about preparing a delicious meal and having a photograph of people enjoying some steak or whatever. And it's kind of more intelligent now than just trying to find some long tail keywords that you think there's a gap and sticking all of those in a listing.
Peter Crosby (29:51):
So I'm thinking again, of putting myself in the place of our listener, this sounds really hard and expensive. So tell me, and I, I'm being facetious, but our listeners are so pretty whelmed or overwhelmed with what's on their plate right now. How does this scale for somebody when you think of our listeners who have, some of them have thousands of products and sell on all sorts of different insights, what is the hope as this scales over the next few years to make this an achievable thing? It is pretty critical to get right,
Max Sinclair (30:33):
I think. And talking from personal, purely personal experience, number one thing is this is a topic that the senior leadership or at least senior marketing or senior digital leadership in a company need to be thinking about. It's a big change and a big opportunity. And I think without that kind of buy-in, I agree you're not going to get very far trying to fiddle around the edges with the existing graphic designers or the existing SEO teams because frankly they don't have the budget or even the motivation, if you're talking to people who aren't motivated by increasing revenue or somehow tied in the growth is part of their personal compensation, it's going to be very hard to shift behavior in the way that it needs to shift. So I think I would say is the number one thing, and I don't know exactly the profile of your listener, whether those are the decision makers in these various companies or if these people need buy-in, but in my experience it's say when you explain it to a senior leader in the company, they get it because they've used chat gpt, right?
(31:55):
Who hasn't used chat GP two who's in a digital e-commerce world? Nobody. So they understand it and it's not a huge leap to be like, okay, there's an opportunity here, we need to start thinking about stuff. And obviously there's a bunch of AI tools, I'm very biased about them, but there's a bunch of AI tools that you can use to help you to achieve this. And I think in many ways that's better than trying to get a human because if you find an AI tool which is understand these algorithms, they're going to be able to do a much better job as I talked about with the Waymo than trying to get people to do it.
Peter Crosby (32:37):
So when you talk to the senior leadership to get their attention and to make them understand the difference between current practice and where growth comes and future practice and where growth comes, right now I'm summing up on a, keywords are dead, but I dunno what's the right way to really bring to life to senior leadership? What a paradigm shift and slash early adopter opportunity this is.
Max Sinclair (33:05):
It cut to me and in my experience, it comes back to saying, this is a new channel, this is a brand new channel. We've absolutely massive opportunity depending on who your customers are, it could be a bigger or small opportunity so you can get data, and I use this when I talk to some people of the age ranges who use chatt pt. Obviously if you're selling products to 20 to 50 year olds, it's a massive opportunity. If you're targeting a slightly older generation, it's probably not relevant as much to you, but you can kind of break it down. And even actually by race, so surprisingly Hispanic use of Chachi Paters twice that than white Americans. So there's even more opportunity and I guess individuals can just quickly use perplexity would be my main recommendation to just say, this is my customer, what's their adoption rate?
(34:11):
And there's lots of studies on this, so the numbers are pretty clear and the headline number is a billion messages a day. That's pretty massive. So I think it's quite, people might not necessarily be thinking about it, but as soon as you mention it, yeah, they will start to think about it. And my view is in the future this will be a new organization, like a new department in major brands and retailers that they'll have focusing on GEO and it won't necessarily sit with SEO, I mean it might have similar people working on it. If SEO is still important, probably will still be important for a section of the population, but to me this is a big new opportunity and it requires a unique focus and maybe new departments in future.
Peter Crosby (35:11):
Well, as everyone can tell, this is a fast moving and fascinating moment. And so one of the ways you can keep on top of it, max does a terrific blog at his site and it's eCom, so ECOM, like eCom tent.ai. And so go to their site, lots of great articles. He about one, which is how search is changing because of these generative engines. There's a great blog post on insights from the Amazon science piece. So if you go to that log, you'll see a lot particularly on the top of the list that you can go to and start to dig into, get these concepts because moving fast and it's great to have somebody like you on Max to share the latest and give a sense of what's ahead. So thank you so much.
Max Sinclair (36:10):
Thank you Peter and Lauren, been a pleasure.
Peter Crosby (36:12):
Thanks to Max for sharing his GEO wisdom with us. Speaking of wisdom, you can find lots more from experts and peers at the 2025 Digital Shelf Summit in New Orleans and April. More info at digitalshelfsummit.com. Thanks for being part of our community.
