Amazon Product Page Optimization in a Rufus AI World

Michael Erickson Facchin: What's going on, Badger Nation? Welcome to The PPC Den Podcast, the world's first and longest running show all about how to make your Amazon advertising life a little bit easier and a little bit more profitable. Today on the show, we have our dear friend Emma from Marketing by Emma. Come and give us our annual reminder, our annual check-in on how to improve our product pages, not only for discoverability, but also conversion. Some of this conversation was inspired by a little activity that I started doing recently, where I started giving my Amazon product images to AI, asking it to describe it back to me, because I know that Amazon is scanning my product images with their AI and then using that as part of our new search algorithm or Rufus discoverability. So I think this topic is incredibly interesting about how to optimize for the customer journey, how to optimize your product pages in the world of AI searchability. Let's jump in. Emma, great to have you back on the show. Thanks so much for being here again. Welcome back to the PPC Den podcast. Emma Schermer Tamir: Thank you for having me. I'm excited to be here with you today. Michael Erickson Facchin: So one thing that keeps coming up. For so many people, I actually think it's a change in the way that people interact with computers and the internet. It's sort of the change from searching for keywords, going to Amazon, typing in some keyword, and instead beginning to actually use natural language more and more with our computers. So being able to go to an Amazon search box and actually start a conversation with Amazon Rufus. That actually says, hey, give me some gift ideas for people who play guitars, for my husband who plays guitars, for my friend who plays guitars, who plays electric guitars, right? Like more and more of these searches will start with Actual natural language, which changes a lot of stuff. And it also makes the search results look very different as well. So I think this is a great topic you came here with today. Has this been coming up more and more for you in conversations with clients? Emma Schermer Tamir: Yeah, it's just been coming up in everything. I mean, even I spent the weekend with my parents for Mother's Day. And rather than going to Google Maps to find a restaurant or To figure out how to cook artichokes or all of these different details. My dad's first instinct was to just ask ChatGPT and he's not totally. of a tech novice, but I would also say that he is not somebody who's the most cutting edge. And so to see the way that this technology has just become a cornerstone of so many parts of his life and decision making and everything like that, I think also just really drove the point home that this isn't Something that is down the road. This isn't something that is only being adopted by those that are in the know. This has really permeated the average person's life. In a pretty significant way, I actually I think I just saw an article last week from TechCrunch, and it was about ChatGPT's new shopping update. And it was saying that OpenAI had reported over one billion queries the previous week with ChatGPT, which is mind boggling. Michael Erickson Facchin: Yeah, it's a top 10 visited website these days. And out of all the top 10, I think it was one of the only ones that gained traffic compared to like it took traffic from the other 9 in the top 10. It was the only one who gained. So like Instagram went down, Google went down, everything went down a couple points, like two or three points, and ChatGPT went up about like 10 points, which is unbelievable, right? I also just looked up and this is a quote way back from October 2024, Amazon Rufus What's handling 270 million daily queries? This is way back in October 2024, accounting for 13% of all Amazon searches. Emma Schermer Tamir: Well, and what's staggering about that is I honestly think that Rufus is sadly Kind of struggling in the back, trying to pick up speed. It just, from my experimenting with it, I don't think that it offers the same experience and functionality as any of the other top AI chat tools. Even just the way that it visually displays the information, it's not as functional. And then it's... Michael Erickson Facchin: By design probably, like the Amazon wants to keep it in a tiny little box. Emma Schermer Tamir: Yeah, and I think that's really it really hinders it because if it's meant to replace the search results page, but then you're needing to look at and compare five or so different products, But they're all in this tiny little box, one stacked on top of the other. That's just not as easy to do than if we could see all five together. I actually, I think that Perplexity does the best job of this because they are, they present the information in a variety of different ways. I think understanding that it's helpful to have things presented where you can learn a bit about each of the products, but then they also do a cool table where they'll have what it determines are Not just pros and cons, but maybe some of the key features that people might be considering, really helping to quickly and easily wrap your mind around what the different options are to be able to evaluate what's the best choice for you. And Rufus, not only does it not present that in a more helpful, utilitarian way, but sometimes it just also provides these really bizarre Follow-up prompts where I'm like, where are you going with this? Like I was, um, I was filming a video a month or so ago and I was looking for an example and I needed a sun hat, like a big wide brimmed sun hat for the hot Las Vegas summers. And I, so I was lurking in Rufus and one of the follow-up questions was, What's the best hat for different face types? And so I clicked it because I was curious. And then it just gave all the different face shapes and the different types of hats that suit it. So it had nothing to do with sun protection. It had nothing to do with my face shape. It didn't. It was like, not only did this not solve my problem, it actually overwhelmed me with more information than was helpful to the point of where it foiled The ability to make a sale. Michael Erickson Facchin: Yeah. And I wonder too, like, it seems like the Rufus has a really small context window, like it doesn't maintain a long history with what you're trying to get. And also it's like tiny, even on like on desktop, it's incredibly tiny. You know, it's really interesting in that regard. I mean, if you were to zoom out even beyond Amazon, like will people use like Gemini on their Android phone or Siri, once it becomes smarter, on their iPhone. So just begin to ask questions there. Like, hey, I'm looking for... Whatever, can you help me find it? And people will get used to those rich sort of cross-platform, like just scour the entire internet on Amazon and off Amazon for whatever it is that I'm trying to look at. So it is pretty interesting. I've seen people use Amazon in my day-to-day life, like look at a product, and there was a moment of frustration where they're looking at a product page and they're just like, does it do the thing? Searching for a 40-pound kettlebell, you click on something, and it takes them to a 30-pound kettlebell. Or they want one with a rubberized surrounding on it, and they're only looking at a galvanized metal one. And it's like, how come I couldn't find the thing? So the normal search box, I feel like people can get a little frustrated with. And then looking at the page, having a question about the product, and Looking up at a huge mountain to try to scan the page. I have seen people like normal day-to-day life Where's that thing that I can just ask about this page? Does it do the thing? And I've seen that actually be inaccurate on the page itself. You ask a question like, is this made in the USA? Or is this a certain material? Or how much does this weigh? What are the measurements? And I actually have it go wrong, which I think is interesting and it's hard to tell currently. I mean, I know best practices, but it's hard to tell right now if that's like a Rufus issue, not scanning the product properly, or if it's the, because I have seen this a lot of individual sellers, not filling out all the fields of information for a product. Emma Schermer Tamir: Or concurrently filling out the fields with keywords that are not necessarily accurate to the product. With the hope that they can scoop up some of those customers. So if they're selling a kettlebell that is a 40-pound kettlebell, but they feel like, well, maybe somebody that's searching for a 30-pound kettlebell will also want a 40-pound kettlebell. Or maybe they'll realize that, why would I get 30 when I can get 40? Because I'm stronger than I let myself on to believe. And so it's this sort of like, I think some of it is hopeful thinking, but I think also some of it is a little bit of smoke and mirrors. I hope that perhaps you'll be able to win over the unassuming customer, which is really not great for your long-term business success. It's actually why I'm really, people probably won't like me for saying this, but I like the new badges that Amazon is using with the, this is a higher frequency return. Versus there's also the opposite, which is people return this less often, because to me, this is a direct response from Amazon's part of how do we begin to crack down on this? And obviously, a lot of the onus is on the sellers to make better choices. And so if they can illustrate that, hey, Trying to manipulate people into buying something that is not actually accurate to what they're searching isn't good for anybody. And so let's clean this up a little bit. Michael Erickson Facchin: Yeah, so this concept of like optimizing for AI when it comes to SEO. I've seen it. I just saw an Instagram ad for myself. It said like a generative SEO optimization conference in Austin. Go attend like so like this is becoming more and more of a topic, right? How do you optimize whatever it is that you want people to go to for It's like optimizing for AI is pretty interesting. And you know, the first reaction is really like, what is the actual difference? Because like good product page optimization, has it changed very much? Or is it sort of reinforcing the best practices? Emma Schermer Tamir: Everything that I'm seeing, everyone that I'm talking to, all of the experts that I'm either having conversations with or I'm consuming their content, all are in agreement that really what this next stage of optimization best practices looks like is really doing what we should have been doing all along, which is writing very targeted, clear, Impactful content that accurately describes a product in a way that a customer is going to be able to understand. Because if a customer can clearly understand it, then the AI can also understand it as well. And so from that respect, it's actually wonderful because it should make the experience just overall from all angles, both as a business owner, interacting with customers, making sure that you're connecting with the right customers, but also when you have your customer hat on, it should all just be A better experience because we'll be able to more easily find, identify and connect with those products that are truly going to be a great fit. Because ultimately none of us want to be making sales to people who aren't going to like our products or who aren't going to, that that wasn't what they were searching for. Michael Erickson Facchin: I attended an SEO conference like many years ago and something the person said stuck with me for so long, which it's continued to this day, which is like, When someone searches something, do you have the best page on the internet to serve them? And it's like a lot of times I'll have conversations with people and like their answer is like, no. Like other people have like more engaging photos or they put more effort into like writing the descriptions or like they have more content. They've answered more questions on their page. And I think that's like a really interesting bar to set for oneself when trying to optimize Any page, whether it be their website or on Amazon, where it's just like, do I have the best page on the internet for this thing? And like, it's a sort of like, you know, strange question to ask, but it's like, do you? And I feel like that's a really interesting question, which like should make people like focus in the direction of like, what would the best page for my niche actually look like? So I'm seeing a lot of like answer Optimization, where it's like, does a product do this thing? Might be a concern that people have, like, does it do this thing? Like, is it this way? Like, what kinds of questions do people have about this, my product? And then actually preemptively answering those questions, having the answers on your page ahead of time. Happens so much. So like, that's one, that's always been a thing, right? You know, you want your pages to answer people's questions, of course. And I just feel like To me, that sticks out as something that got raised in importance as well, where that might not be a specific, super-duper keyword-rich answer, but it's going to answer the question of, is it compatible with whatever? It's durable for 75,000 hours or something like that. Those kinds of things, I think, are really interesting now. Answer optimization, because people will be asking a lot more questions about stuff. Emma Schermer Tamir: I thought that was interesting. Yeah, and I have been a conflict for such a long time when optimizing for anything on the internet is I understand that this is the information that I need to present and simultaneously we're operating in a much more binary system where if these certain phrases aren't in these specific places, then I'm never going to get in front of the customer to be able to answer their question in at all. And so that, in some ways created this hierarchy where the keywords were the top dog because they were the gateway to be able to speak to the customer in the first place. And so This is allowing for recalibration so that it's not so cut and dry and simplistic in understanding a product based on whether a very specific phrase in that exact order is included. And so then we can move beyond to what we would always be wanting to do anyway. You know, if somebody were at a store and came up to your kiosk You would begin by asking them questions to understand what they're looking for, what they need out of a product and all of those things, preempting the questions that they have, you know, like any good salesperson who understands how to anticipate even what those resistances might be so that you can provide the information ahead of time to minimize that friction and help reduce concern. And so I believe that we're entering like a very exciting time for marketing, not just because of all of the different ways that we're Able to reach customers and engage with them, but also just in the way of being given an opportunity to flex those creative muscles and really like up the ante with how you are putting your best foot forward because what was passable a few years ago is just not going to cut it. Michael Erickson Facchin: Here's an interesting question. You know, for a very, very long time in the world of like Getting found on the internet. A lot of times it was, don't try to appease the algorithm. Try to make it really good for humans and the algorithm will appreciate that. Would you say now that is less the case where you do need to appease the AI that is scanning the product page? Is there any difference in appeasing as opposed to like making it really suitable for humans. Emma Schermer Tamir: No, I would say that that's actually one in the same in many ways where naturally writing about a product in a very descriptive way rather than having a list of keywords is going to check all of your boxes. Michael Erickson Facchin: Here's an interesting, so here's where my mind went with that question. I've been doing this more and more, but like, okay, so now You know, AI can read, like look at your pictures of your product and understand what's inside of it. So like it knows what sort of what color it is. It knows what it looks like. And it can probably make some guesses about how it's used. Like it can see the picture maybe and describe the picture. One activity that I've been playing around with when doing image optimization and people that I've talked to have been doing it as well, is like you actually take your images and like you give it to ChatGPT and ask it to describe what this is. And then like, see if it can really understand what it is that you're looking at. I thought that was interesting. And like, I think that's good to do. Like, I think it's interesting to see what does AI think of my image? Like, can it accurately describe what it is? I thought that was really interesting. And I guess in a perfect world, it's the same, right? Like the AI describes your images really nicely. And it's incredibly clear to people as well. But I almost wonder like, are we going to, you know, I'm just thinking of like people like slipping in hidden something into their images that can't be seen by humans, but is seen by AI, like red words on red text or something. And then I'm sorry, red, yeah, right. It's just like, okay, you see this, this all looks good, right? AI, like imagine I fed an Amazon image with like a hidden pixel of like, so much information. I don't know, like, That is interesting to me, but like giving a product image to ChatGPT, asking it to describe it and being sure that the description is what you intended it to be, I think that's been one new thing that has been really interesting to do. Some outcome from that was like, oh, like ChatGPT says, like, I'd noticed this, this, then that, and that triggers someone to be like, oh, I should like add this other thing to my image. Like it actually comes in more colors and it's actually different sizes. And that wasn't clear in my image. Like, let me accentuate that, which I suppose is good for humans and AI. I don't know what you think of that. Emma Schermer Tamir: That's ideally what we should be doing with anybody that is going to be interfacing with something that we're creating in addition to the non-human things like the AI algorithms that are digesting our data. And I mean, I use, it's so good for blind spots and also for not even just blind spots, but You know, you're coming in with a clear perspective, a clear understanding, and it's sometimes difficult to get the distance necessary to know how it's going to be perceived by someone else. And so I've gotten in the habit of using ChatGPT as like an editor, but like not an editor of like, make this better, but like, how do you perceive what I've written? Like, what are the, you know, what's the tone that you're reading? Like, who is the intended audience for this? Like all of these things. And I don't always, I don't give it any prompting to lead it. To what I'm hoping it will be, I want to know from an outsider's perspective, how is this going to come across? And that's such a helpful set of information to be able to get that was not very easily accessible. Not that long ago and so like I do that with everything from complicated emails to LinkedIn posts to I mean like everything it's just so Helpful because we're stuck in our own experiences when we are creating a thing or developing a strategy. And even if we are very highly skilled at what we do, it's still so easy to forget something or to not realize that that sentence could be interpreted in another way because you even have the tone That is sort of stuck in your head that informs that you know, we've all had that text message that we've sent that like, we thought was pure and good intention and it was a spark that started a fire. Michael Erickson Facchin: That's right. That's right. So, in terms of optimizing for AI, I think one interesting activity I've been doing is dropping my product images into ChatGPT and asking it to describe what it sees. I thought that's been giving me really good perspective. Does anything else come up for you? Are there any activities that you're doing more of, you're telling people to do more of now? Emma Schermer Tamir: Asking who it perceives the intended audience to be based on what I've given it because oftentimes I think it's easy to think that we're Crafting something for one set of people when really it's not necessary, you know, because our slang or cultural references or all these different details can just so easily slip in without our cognition that they're there. And I believe on top of that, what is so important as we enter this very different world of search And discovery and the shopping experience is that having very targeted ideas of who our customers are become even more important. So this ability of just a few years ago to have a very broad reaching, not really targeted Amazon listing is not going to serve you as well as something that is highly specific to Who you know to truly be your target customer. And it's not just because of AI. I think that this also has a lot to do with the way that social commerce is also influencing how people think about and learn about and choose products. And it's all becoming much more fragmented. You know, we don't have those Single sources that are influencing culture in the same way that we used to. Instead, we have all of these microcosms of worlds that exist online of hyper specific sets of interests that Our influencing fashion choices and lifestyle choices and all the different decisions that we make about how we evaluate whether a product is good for us or not. And so it does require a much deeper understanding of who our customers are. And then from there, really building out our content in a way that is going to be able to speak to them. Beyond just this very general. Michael Erickson Facchin: So I think it allows a lot of Amazon sellers and Amazon marketers to do things that perhaps were very labor-intensive but do them very efficiently because dropping in your product listing into an AI and asking it for feedback like, who does this seem like it is appealing for? Is it pre-answering people's questions? Is it descriptive? Did I leave anything out? To be able to converse with AI about your product listings is a really powerful way to sort of cover those bases. And I would imagine through the process of making your product listing more properly detailed and more descriptive and better answer and anticipate those people's needs, it will eventually become better for I'm Rufus as well, which is interesting. Emma Schermer Tamir: Yeah, or any of the AI that's just bypassing. The sales still might end up converting on Amazon, but a customer might not go to the search bar or the search results page at all. They are on Perplexity, or ChatGPT, or Gemini, or whatever other tool they're using that gives them a compelling enough reason. They're like, yeah, that looks good. Click into the product page. Mm-hmm. I think that suits what I'm looking for. Let's do that. Spend the next three hours evaluating all of my possible options and getting completely confused by 20 different products that all seem sort of the same but different on Amazon and I can't actually differentiate between. Right. Michael Erickson Facchin: Can you read all the reviews for all these products for me? Right. Yeah. So it's really it's it's really interesting. Like what is your next call look like with a client? Like client comes to you and says, hey, I need help with my product optimization. I feel like I'm getting crushed out there. I made some Rufus searches and I'm never showing up. What would be your action plan for them? Emma Schermer Tamir: My action plan would be a few fold. It would be, first of all, to make sure that you are really taking a critical eye to your listing to clarify that And I know it sounds so basic, but that all of the must know information is there and easy to find and organized in a clear way. And that it's also very visually compelling and gives a clear pointed Reason why a customer should choose your product specifically. Michael Erickson Facchin: So this person who writes into you and says, hey, I keep making roofier searches. I never show up for them. Would you describe their existing page as maybe like thin? Like how would you what would you expect their existing page looks like? Emma Schermer Tamir: My expectation would be that it's not necessarily thin, but I'm imagining like boxy old like 2017 A plus modules are just really messy looking and like not great pictures and like look like strings of Keywords that probably aren't even that relevant anymore because they haven't updated their listing in a while and poorly edited images where it's so obvious that the product is not actually In the picture, very long bullets that are kind of gibberish where you're like, wait, what does this even do? Michael Erickson Facchin: I just looked at a page like this earlier today. Yeah, the boxy A-plus content gets me. Emma Schermer Tamir: I know. It's like, oh, come on. You would never have a website that looks like this in 2025. Why are you like dismissing this as like anything other than critical? Michael Erickson Facchin: Yeah, I suppose the takeaway from all of this is like product page optimization still matters a lot and like the clarity. Unknown Speaker: is not this. Emma Schermer Tamir: You still ultimately have to convert the customer. So the only way that Rufus is going to serve up your products is the same way that you're going to rise in the ranks on the search results page, which is that it is not enough to have The keywords there, or in this case, maybe those questions answered and that basic content there. But if you are not able to drive sales And drive sales that people want to keep, then you're just going to be treading water. You're not going to achieve the growth that you need. So you need to be so conversion minded. It can't just be this obsession over optimizing your main image, but then failing to actually have good content once people get into it. It can't be this obsession over your keywords, which do still have an impact. And also, you know, keywords can be a great insight into what it is that customers really want out of a product, because they're not always a straightforward Description of your item or maybe it's a description but with the context of how they're going to be using it or what a particular need is that they have. And so those keywords can actually be great clues to use to think about how can I optimize not just the content, not just what I'm writing, but like, okay, if I know that a lot of people are using my product When they're traveling and none of my photos suggest that somebody is traveling. Then that is not really strongly communicating to a customer that this is going to be great when they are on the road. And so it's really important to be looking at all of those layers so that when a customer arrives at your product page, no matter how they arrive at it, that they feel reassured that your product is going to suit their needs. And, you know, one of the reasons that people shop on Amazon specifically is because they have such fast shipping. Like I was going on a very last minute trip a couple of weeks ago, and I was ordering things at 1 p.m. the afternoon before needing to leave for the airport at 6 a.m. the next day, knowing that those SD cards were going to be at my door by 6 p.m. Michael Erickson Facchin: Amazing. Yeah. So, I mean, I'll be honest. Before you came on the show today, I was reviewing a lot of our old videos that we made together. And I feel like it's still thinking of your customer. It's still think of your customer is like the biggest thing to do here. Emma Schermer Tamir: Yes. Michael Erickson Facchin: So, I thank you for our annual reminder of this very fundamental piece. It's always great to have you on the show. What's new for you? Where can people find you if they wanted some more Emma in their life? Emma Schermer Tamir: Yeah, I recently bumped up to two videos a week on YouTube, so you can get a whole lot of Emma ranting about all these things. And nerding out about all the latest developments in e-commerce and AI and marketing and how to make sure that you're ahead so you stay successful. My channel is Marketing by Emma. And that's really where I'm spending a very big chunk of my days. In fact, after our call, I'm going to record another three or four videos. Michael Erickson Facchin: Awesome, Emma. Well, congratulations on everything. Thank you so much for, again, it's so helpful to have on The PPC Den. I know that our listeners, I know that I am very appreciative of it as well. This is sort of moments to sort of get out of the campaign manager and get into product page optimization. And not just product page optimization for pure SEO purposes, but product page optimization for discoverability and conversion. It's always so great to have you on the show. Emma Schermer Tamir: Well, thank you so much for having me. I loved speaking with you as always and I guess I'll see you next year to remind everyone. Michael Erickson Facchin: Thank you. We need it. All right, have a good one. Emma Schermer Tamir: Bye. Unknown Speaker: No bad mistakes. I've made a few. I've had my share of rocky ones. But I've gone through. We are The PPC Den, my friends. And we'll keep up the business. We are the PPCs and we're talking about Amazon. No time for Medicoms because we fix the gambit of the world.