Is Retail Media Ready for AI Ads?

Is Retail Media Ready for AI Ads? Speaker 2: Hello everyone and welcome back to Better Advertising with BTR Media. Today I have on Kiri Masters. If you do not follow Kiri, if you're not subscribed to her newsletter and if you do not know about her, you are missing out. She is a top 100 retail influencer by Rethink at age 40 under 40, which is an incredibly exciting one that I've always wanted to be a part of. And just an incredible analyst, commentator, influencer from within our industry. Is there anything I missed on that, Kiri? Speaker 1: Thank you for the overview. I mean, I will add we used to be competitors, but we were always still very friendly even back then. So I appreciate that. Speaker 2: We were just having a fun conversation of, I feel like the key influencers or reporters of our industry are not the same people that they were seven years ago, but we're still going strong. We have head down, stuck with it. Speaker 1: Don't have anything better to do, right? Speaker 2: Now that may be the problem. I'm going to find something to do other than focus on all of the new updates because that is what has kept us in the industry. I would say is the industry changes and evolves incredibly, incredibly quick. And you are one of the first people to report on all of the change and excitement across the board. Speaker 1: Thank you. I appreciate that. There's a lot going on. I always feel like even though I'm publishing a lot, relatively speaking, I am constantly behind. So if you also feel behind, please know it is my job. I don't have anything else that I'm doing. I'm just trying to keep up with the news and I'm struggling myself. Speaker 2: I have gotten into an unhealthy habit of benchmarking my content to yours and it's terrible because I mean, you, you are so ahead of the game right now that I'll see a new post and it gets sent around. Like we have this little e-commerce friend group and it's like, Kiri's there again. I'm like, dang it. So my, my middle ground was like, I have to have her on the podcast. I need to learn from everything she's putting out lately. Speaker 1: Well, I'm not running an agency anymore, Destaney. You've got a whole other job that you have to do. Speaker 2: I will take that. That's the little piece of confidence I needed today. But one of the most recent articles that you wrote was actually for The Drum and I was excited to chat about it. The title of the article is Why OpenAI's Ad Announcement Should Worry Retail Media Networks. If you haven't read it, we'll link it in the comments. It's a fantastic read. But Kiri, do you want to give us a quick high level of the premise of that article? Speaker 1: Yeah, so I think that this is a existential moment for the retail media industry as we look at OpenAI, you know, now launching an ad capability. This is not a huge surprise to anyone. It was always a question about how these LLMs are going to monetize. We have other LLMs that have monetized this way for a long time, and this is like one of the Best kept secrets out there is that it's not even a secret. They've just not done a particularly good job of publicizing it. But Microsoft has had ads inside of copilots right since the beginning, back when it was called Bing Chat two plus years ago. So ads have been part of the ecosystem for a while, but just it happened that the biggest player had not really rolled out an ad capability, even though we're all expecting one for a while. So what does this mean for retail media? That's really the interest that I have. And I think that, look, it's not like retail media is over. It's the funeral for retail media. We're absolutely not there yet. But I think it does raise some existential questions for the current setup that we have in retail media where Most of retail media is on-site sponsored product ad units and on-site banner ads with a little bit of off-site retail media sort of in the ecosystem as well and then tiny but sort of fast-growing area of in-store media as well. I think the first two have an existential threat that's being lobbed at them. With ads in LLMs and yeah, we can get into all of that. Speaker 2: When you wrote this article, had they yet announced the approximate CPMs for ChatGPT? Do you have any thoughts on the CPMs that are, are they just projected or are they actually live expected CPMs? Speaker 1: I think it was the information that got those figures. As of the time we're discussing this, I'm not sure if it was confirmed by OpenAI. Speaker 2: I'm not sure either. What I saw is, I believe, along the lines of $60 CPM for ChatGPT, which is the highest out of all the media, with the second below it being linear TV, potentially at $50. The reason I bring it out is it's been a pretty big shock, I would say, for everyone who has been buzzwordy around AI, and there's been a lot of conversation of, are brands going to risk moving their budgets over for a CPM that high? But I think the part that I'm actually really intrigued in, the part that I think maybe people are potentially missing is if OpenAI does things well, they have so much more context and relevancy than I would even consider Amazon having. And that's what's made Amazon so successful is their relevancy is incredible, so conversion rates are inflated. My chat knows everything about me from my blood work to how I feel about I'm a cat to my thoughts on retail media. So when I ask the question, it gives me a really aligned response. And if we consider the aspect, you know, putting agentic aside, how much my chat, because I use ChatGPT, knows about me, I would assume I'd be willing to pay those CPMs if I had really incredible conversion rates. Speaker 1: Fellow cat ladies of LinkedIn, unite. Yes. I'm here for that. Absolutely. And I think that this is the question. Just to quickly address the format and the fact that CPMs as of right now, I don't think that we take what OpenAI has confirmed or announced or started out with as any real indication that that's what ads in LLMs are going to look like going forward. I have a Wildly optimistic for you, perhaps, that these ad units inside LLMs don't necessarily have to look like what we've become accustomed to with sponsored products and the blue links. There is an opportunity with this surface to We provide ad units that are immersive, that are multimodal, that are much more relevant to our context, as you mentioned. And I'm hopeful that we see something genuinely new and interesting that comes along and that it's not the, to use a very technical term here, insidification of things that we've seen in the past where the ads So I think, you know, that there's a, the example I would call out is, is Instagram being generally like an additive experience with the ads because the algorithm knows this well and, and I'll, I'll find stuff on there. I didn't know that that existed. I didn't know I wanted that. That's actually great and very relevant to me. So I think when done in the right way, Ads don't have to be disruptive. They don't have to take away from the experience. I don't see an ad on Instagram and think, oh, they're trying to influence me and get into my mind. I'm like, it's an ad. And is it something that's relevant to me and interesting? Yes or no? And I don't really sort of judge the platform for it. It is a free service. So I think there's two arguments there. One is maybe this is a chance for some really great new advertising experiences where we can do virtual try-ons with avatars that we can hear about. New product drops from our favorite brands proactively and like there's so many interesting things that could be done. And secondly, because of the context, like you mentioned about just how much they know about this, algorithms know about us. There's also a chance for these ads to be much more relevant to us as well. Speaker 2: I couldn't agree more and leaning in on your point of they don't have to be disruptive. The word I've been using is I can see an avenue where they start taking up a lot more discovery. Right now, I would say the majority of discovery is really being driven by TikTok, Instagram, affiliates. That's how I'm discovering new products. By the time you make a search, the digital shelf has started to reflect the retail shelf. Your top 20 biggest players in any category are spending the most on ads. They're showing up in the top of the sponsored placements and they're ranked the highest. Amazon started slipping in discovery in my opinion. Walmart has not done an incredible job, but it is a little bit more difficult in store. If, you know, chat or all of these AI models start leaning into the advertising experience from a discovery perspective in that I am asking chat about my cold and flu and it just happens to work in the perfect medication based on my symptoms. And XYZ, I will discover new products that way because I'm not going to a platform and typing a very high intent search. And I think that could really change how all of the retailers play together, especially when you have things like the OpenAI and Walmart partnership, where if chat starts driving discovery a little bit like TikTok has, and then driving that traffic to Walmart, I think that could be very powerful. What are your thoughts on that? Speaker 1: Yeah, I kind of see an opportunity for a collaborative bidding model as well between the retailers and brands such that let's talk about like, you know, big CPG brands that don't typically have much of a focus on D2C. They're primarily selling through retailers and they're happy with that model because it's expensive to ship. Bags of cat food. So to that end, it would be interesting for cat food brand to be collaboratively bidding with Walmart, with Petco, et cetera, within an LLM for a purchase to be made, even if it takes place sort of over on their retailers app or on their website. I think also what we're, you know, there's There's a lot of debate and discussion about like a truly agentic shopping experience within an LLM. And are we going to be checking out within ChatGPT and Gemini? I think eventually we will. Right now, the experience is pretty suboptimal compared to using a And we're going to talk a little bit more about that on the Apple website, but I don't think that we need to get to that point before we start to see some big disruption to how retailer.com's operate and the data that they're able to collect on search and browse behavior and research stages. That's actually where a lot of the data comes from to enable great targeting and interest groups and things like that. Upfront discovery behavior happening on a retailer.com, then that takes away the onsite ad inventory and it takes away some of the data signals that they're getting, that they're using to create these audiences for brands to activate against. Speaker 2: I couldn't agree more. Everyone's leaning in too heavily on agentic isn't a great experience, but I don't think that's the disruptive part for right now because it takes 30 seconds to click the link from my ChatGPT and buy on the retailer site. The part that's been disruptive for me is posting a photo of my cat and saying what food is best for my 50 pound Maine Coon, not actually 50 pounds. But being able to have that high alignment, the other day I was designing part of the podcast studio and I took a photo of the space and said, can you give me recommendations on the chairs and table that will look best in the space? It gave me 10 links. I'm not going to use it less for discovery because I cannot check out immediately because, in my opinion, from all the platforms I've used so far, the discovery and the context to my design and aesthetic style is so aligned through one app, I can then go Apple Pay anywhere else. So yes, agentic may not be amazing right now, but to your point, it's not the disruptor. It's going to cause it to be a little bit more seamless and frictionless in the future, but it's really not that high friction to swipe over on my app and push double tap on the side of my iPhone. Speaker 1: Right. I'm not sure if you've been following, there's a new sort of setup that you can do called Claude Bot and they just rebranded as well. Have you been hearing about this where people set up like that? Yeah. Speaker 2: It's another area where everyone's posting about the incredible efficiency they're driving with Claude and I'm like, I can't. Speaker 1: I can't. I'm with you myself, but I think I completely agree like we don't and I'm optimistic about agentic commerce being a thing I would personally love for my groceries to get ordered automatically. Yes. And for my agent to go find the best deal on my cat food. But as it stands, it's not, you know, really there yet. But we are seeing that disruption, I think. Just for listeners who don't have this context with Claude, but now it's been renewed. I can't even remember, but basically there is this, we're able to now start using AI, not like I need to go onto the chat GPT app now and search for something and do it. These custom setups have been designed and they're very technical right now. Don't stand a chance doing this myself. But people are interacting with Claude using like the signal messenger and voice messages and stuff like that and getting it to go and take actions on their behalf by setting up a virtual private server or a different computer and just sort of using this I'm an assistant in this way. So I think things are just changing so quickly that we could see that really come to life sooner than later. I actually want to show you, Destaney, one, and I'm not sure if this will work, but on the, just as a point about the multimodal possibilities here, have you used the Zara shopping app recently? Speaker 2: I have not. Speaker 1: Let me show you something cool. So on Zara, now you can actually have a AI generated model of you and do virtual try-ons in the app. Virtual try-ons have been around for many years with makeup and things like that, but that's not particularly new. But I do just want to show you because this, I think they built this technology themselves as well, but just to show you how Great. This can actually be, I'll show you, this is my AI. Unknown Speaker: Wow. Speaker 2: That is incredible. Speaker 1: Isn't that amazing? Speaker 2: Wow. Speaker 1: And then I've got, so I can sort of like change the look, um. Speaker 2: And I'm assuming it's pulling in accurate measurements for all pants size. Wow. Speaker 1: So look, this is, we're at the start of something like this where we could have, to your point about then, you know, even these products that are very, um, uh, you know, they, they, they, You'd classically think of them as something you'd want to go to a store and sit on a sofa or see a sideboard in real life or like to your point about home decorating, apparel, these categories that they're kind of fun things to buy as well. So we've got this interesting dichotomy of like the tedium of shopping, of reordering things I always buy, of like, okay, I know I want to cook these things. Now I have to go and individually add. Search for it. Select how much. Do a price comparison for every single item on our grocery list. It is tedious. So that's one sort of Great use case that we all imagine is like automated grocery ordering. Awesome. Can't wait for that. But then also there are these shopping missions that we have that are more experiential. The journey of it is kind of fun. And I think what got lost is like, well, I don't want ChatGPT to like go buy my jeans for me. Well, no, of course you don't. But isn't it great if you had like a little bit of help with that and like a styling assistant? I would love that. Couldn't it be fun as well? Speaker 2: I can never find clothes that fit me. So having an experience that has all of your measurements down to a tee. I mean, I'm such a fan and maybe I'm naive to be so excited about all of this, but When something has the appropriate context for what you're trying to achieve, it can do it faster and better than you can. Absolutely, save me some time and energy to go do this so I can post as much content as Kiri. It's like, at the end of the day, the time savings is incredible. I mean, what, five years ago, I don't actually know the timing. I should not be blanking on this, but COVID's a major catalyst to the older generation starting to buy into more online grocery and grocery delivery. And that was incredibly disruptive. Now we're sitting here saying, I can get my groceries in three hours from the majority of the provider. It's incredible. But now we have an opportunity to even take out that little piece of the puzzle and have your groceries shopped and price matched directly to your house without you having to really lift a finger. Speaker 1: Mm hmm. Yeah, absolutely. So I think this is the thing that the criticism I hear is like, well, what? This self-selecting population that we have that's like typing all this up on LinkedIn, well, of course, we're going to be the early movers on this. I completely agree. Like there is a lot of stuff that we're talking about now that the general population has got no awareness of, let alone any inclination to try. But this is the way all technology works. There is an adoption curve. It's not that great at first. I shared this example on I did another podcast recently, but I was thinking about what was the first thing I bought online. And it was, it would have been about 1999. And I bought this MP3 player on eBay. And so I was so excited. I did all this research. I got this really cool one with a screen. It was so big and chunky and heavy. And I bought it on eBay. And then I waited three weeks and then I went to the post office and I handed over my cash to the clerk at the post office and I got it cash, you know, I paid for a cash on delivery because I, you know, I was, I was also a teenager, but like, we didn't really buy things on the internet back then. Right. So we did cash on delivery. Then after that, I was like, you know, that that worked pretty well. Maybe I will use my debit card to buy the next thing on eBay. And that's how it gets started. So I think if, you know, as people have positive experiences here and the technology gets better, that it starts to move. Of course, we're right at the start of when it's kind of like a You know, it's a hit or miss experience, but just at the very beginning. Speaker 2: Yeah, it's interesting. You hit on your five points for being fundamental of is something going to stick or not. Jay Richman recently posted similar, and he said that he's the VP of product for tech and design at Amazon. I've worked closely with him on creative studio launches, and he posted beginner's mindset over expertise right now. And he said, I've been giving more career advice and a lot of people are struggling with AI, but it's not the people that are lacking talent. It's the one that are constrained by experience, not maybe being willing to trust a platform or not knowing how to give it context. I've seen the same thing. All of my college students that I'm bringing in teaching ads are already creating headline search ads or leaning into AI to help with forecasting and all these incredible things because they don't know what they don't know. So they're naive and they're able to use the tools in very unique ways. Those with experience have certain ways in which they like to do things. And I think that's pretty important because you kind of mentioned like the younger generations adopting and like the learning curve we're seeing. AI has moved so quickly and been so disruptive that the people who are still like naysayers over agentic commerce or how this could be so disruptive to even retail media advertising, I think are wrong. But I would love for you to dive into kind of the five key points that you've used to make decisions and fundamentally say like, yes, this should be concerning or brands need to lean into it because it is here to stay. Right. Speaker 1: Yes, absolutely. I think this is where frameworks are helpful. And so when I've been looking at the examples of new technology or conveniences that never really stuck versus the ones that stuck around, I think of five different criteria. So the first is pain. Is there pain in the current It's a great way of doing things. So we don't, you know, if you ask most people, like, do you think that shopping online is painful? They wouldn't say that. But when you step away from the experience, we're talking about You know, like this ocean of products that all look the same, fake reviews. I just don't like when I'm buying something on Amazon, I'm looking for a new, like a hair styling tool. And honestly, when I go on Amazon, if it's not a brand that I recognize, I'm very skeptical because I know, as you do, Destaney, how things really work when you've got a very aggressive brand that wants to build traction on Amazon. I look at these five star reviews and I'm very skeptical. And now it's my job to like figure out what's like legit, what's a legit brand or not. The return, like dealing with returns. So there's a lot of there's actually a lot of pain in the process when you step back and look at what it's like to shop online. So that's one element. Is it solving a real pain point? That's necessary. The next one is frictionlessness. So does it slip into our everyday life? And I think that this is one area that, for example, the metaverse What failed is like, OK, you need this headset, this big expensive headset that's like really heavy and you try out some games and you put it away and you don't use it again because it never really like integrated well into our daily lives. Whereas with with AI, we've got our phones in our pockets. We're using them all the time. So that's that's another aspect. Culturally, this is a big question. Is it is it do we have trust in it to to to work? Do we have trust in the security of it? Lots of things tied up here. Is it something that people are just kind of willing and excited to participate in? So the way I think about this is Our relationship with AI has evolved over the years from like, it's an intern. We ask it to summarize an earnings report, but like do it in the style of a, you know, 90s hip hop artist or something, you know, like. This silly stuff and we had to like we had to watch it and manage it. Now I think like a lot of people would look at their LLMs as a best friend like we tell them. Speaker 2: It's scary. Speaker 1: Yeah. Yeah. We tell them a lot of stuff and now it is more rewarding because the LLMs have much better memory than they used to. So they remember things about us and our preferences and what we, you know, When I'm looking at a hairstyling tool, like what are my, what are my priorities? Am I a value buyer or am I, um, you know, focused on other attributes? So that's another area is like this cultural adoption of something. The technology itself needs to deliver on the promise. And that's where I think like if we're talking about truly agentic shopping, we are not really there yet because the capabilities are so limited. And then finally, economics. So is this a good deal for me as a consumer? Do the other parties have an incentive to get involved? And I think this is an interesting part because a lot of the, you know, the economics of this ecosystem are being worked out. OpenAI just announced that they're sort of the transaction piece that's going to be a 4% cut that they're going to take. And so there's some people are of the view that, hey, that's a bargain. Like that's better than what you can get on the Amazon marketplace simply to get an acquisition, like a transaction. And then other people are saying this will never work because it's just another tax on these transactions. So there's diverging opinions there. I think ultimately the The ecosystem gets pulled where the customer is incentivized. Right. So if we think about. Do retailers want to get involved with this? Well, they kind of need to be where customers are really wanting to transact. So, of course, it can't be substantially a loss for them to do this, but the economics need to be, if not perfectly aligned for all players, at least it needs to make economic sense. Speaker 2: I love it. I think that gives a clear foundation to the why this needs to be talked about, why we need to be considering it with our planning, our budget distribution, with our discovery methodology. Do you have any last tips or recommendations on how brands can start incorporating and or planning for the future? And, you know, originally when I ask these questions, it's usually, you know, planning three years out. I don't even think that's a realistic world to live in right now. What should brands be doing to set themselves up for success? Speaker 1: A lot of the action that can be taken right now is really on the part of a retailer in this area. So I think that, look, there's a baseline, and I'm not an expert in this, but there's baseline things to understand about About GEO and AEO and these things, which is, you know, the more I learn about it, the more classic SEO really is. So look, do your due diligence there on what is vaporware and not, but that's a foundation for brands. But a lot of the access, especially if you're a brand that primarily sells through retailers, a lot of your ability to show up with the product cards, In these LLMs comes down to how invested your retail partners are and what they're doing and what is their posture around it. So I think it's really important for brands to have these conversations with their retail partners. Hey, what is your philosophy around adjective commerce and how are you working with these platforms and how can we help To participate and have the right product data available for you and working collaboratively with their retailer partners to understand what they can do to be supportive. As it pertains to the ad side of things, it's also another area to be asking retailer partners about how are they thinking about this from a existential angle. I will say, you know, some retailers will be more resilient than others in this regard. Some retailers, their value prop is Not so much on the sponsored product ad units. They're much more about in-store. They might be operating in a regional area. They might be... It might be a little bit different. So this is where context is important. But for brands, I think it is really just trying to understand how their retail partners are Thinking about this and who is leaning in versus who is sort of holding back on where this is all going to go. Speaker 2: I love that. I know our brand's been asking a lot of questions, and I think that is the most scalable solution is to kind of lean on the retailers. And the blanket advice I'm also giving everyone is, look, the LLM, trash in, trash out. It's being built off context from somewhere, whether it's your consumers giving it insights and information or even from the brand perspective. I mentioned I was sick a few weeks ago and I was asking for recommendations and something around like a migraine. It wasn't even specifically related to the sickness. It said, you know, based on your other symptoms you have, I'm recommending elements and liquid IV. And I was thinking a lot about that. Both of those brands have done a fantastic job of building kind of awareness and discovery off platform. So the LLMs had enough context and information to pull in. So even your conversation or your aspect of regionalization, it's for any brand, even on D2C, are you being very vocal around what stores you're in? Are you putting context online for the specific problem that your brand solves? Because the LLMs are going to be pulling in all of that information. So just make sure you have a strong foundation. We're giving the LLMs and the retailers the right context. At the end of the day, it's another algorithm. It's traffic and conversion that we're optimizing for, so making sure that I think we're doing everything we can to power both of those levers. All we can do at this point because it's changing so fast. Speaker 1: Yeah, absolutely. And I think as it pertains to advertising as well, I mean, it is interesting to see what ChatGPT is doing. They're rolling it out to a very small number of advertisers. It's CPM only. It's frankly not really a point where it's super attractive to most brands, I think. Or accessible. So it is, I think, a wait and see on that side. I think if you want to start getting your feet wet, start looking at Copilot and other platforms that already have some capabilities here and understanding how these work. It might not look exactly the same. I hope that we have more sort of interesting ad units, like I said, but You don't have to be waiting around. There really are opportunities to get your feet wet today. Speaker 2: Absolutely, even as simple as Amazon-sponsored prompts. We're not seeing a lot of traction with them yet, but we do have prompts live and it's as simple as moving away from a keyword intent to audience-based kind of more natural language targeting. So we're playing around with it, but I think the key theme to this is this is something that you need to be paying attention to. There may not be a perfect move just yet, a perfect playbook, but that's, I think anyone who works in the e-commerce industry know there's not a lot of perfect playbooks. It's really contextual and I think building the right solution for where you're at, at any point in time. Amazing. Well, thank you so much for joining. It has been an absolutely incredible conversation. Is there anything that you want to shout out or maybe give some more context on your newsletter? Speaker 1: Yeah, well, I mean, this has been interesting. This is a lot of the stuff I talk about in my newsletter, Retail Media Breakfast Club. It comes out Monday to Friday, sorry, Monday to Thursday. I had to drop Fridays. Monday and Thursday as an email at 6 a.m. or as a podcast at 6 a.m. It's just 10 minutes long. Or if you're more of a LinkedIn junkie like me and Destaney, you can also follow on LinkedIn and you see those posts pop up there. Speaker 2: Amazing. I will make sure to drop everything in the comments and thank you again so much for joining. Speaker 1: Thank you, Destaney. Great to see you.