The Retail Roundup - Rufus Gets "Agentic", Walmart Gets Fast & Sellers Feel the Squeeze - Episode 433

Pat: This is the E Commerce Brain Trust, a podcast about building momentum online for established consumer brands. Join our hosts and their expert guests for high level conversations about e commerce strategies, trends and innovations. Access our brain trust and boost your brand's E commerce potential.

Julie Spear: Hello and welcome back to another edition of the Retail Roundup, where we recap the biggest news and trends shaping retail media and marketplaces. And we've got our crew back together. We've got Armen for all things operations and organic in retail marketplaces. And Pat focusing on the retail media side. And as always, my trusty sidekick, Jordan Ripley. With that, we have a lot to dig into today, so without further ado, shall we round up?

Pat: Let's round it. But let's not be associated with Roundup because there's a lot of leads.

Jordan Ripley: Yeah, yeah, yeah, yeah. There's no Monsanto sponsorship.

Pat: Okay, perfect.

Julie Spear: I had a feeling my intro might be not that great.

Pat: I'm just glad you guys picked up on the deep cut Roundup. So it's already off to a good start.

Jordan Ripley: We all have season. We know what that is.

Julie Spear: Uh, all right, so we're bringing back for the second month our AI corner. That's where we're going to kick things off. Because in our space, in everyone's space, in our lives, AI is front and center. So we must start there. And in the world of retail media and marketplaces, when we're talking AI, let's start with my favorite AI of all, Rufus. The name just makes it so endearing to me. And Rufus is proving to make our days more and more interesting as well. So I'm going to start with you for one of the latest Rufus updates that we saw in the past month. Pat, we have a couple of ways that we're going to be talking about Rufus today. One is the way that it's impacting how customers use it, but one is the way that it's impacting brands. And as of March 25, Rufus sponsored prompts moved into general availability for brands. And also with that, we're seeing that brands are now getting charged a CPC when customers click on some of the Rufus generated questions. Can you talk us through that update a bit and what it can mean for brands?

Pat: Yes, for all things that AI is new and Rufus is new. The one thing that you can always hang your hat, um, on is Amazon monetizing retail space on their PDPs. I feel like that's a bet we can make forever. Yeah, you spelled it out for the listeners. What's really changed here is that Rufus was In open beta, many, uh, advertisers had access to it, but now it's gone ga general availability, which means it's available to everybody and it's defaulted to on what you call that in the open, I actually think is important because we think of Gemini and ChatGPT and Claude over in this bucket over here and then we think of Rufus as something different. And it is different, but it is essentially still Amazon's own LLM in the same way that uh, asking questions and engaging with it. What we're seeing, like observationally what we've seen and this is across accounts, brands, advertisers of different sizes, also different categories, is we're not seeing a ton of volume yet. So not a lot of clicks, not a lot of spend yet. Now that what I would say my recommendation is, if I were a brand right now, I would still want to go into my seller central, go to my ad console and download the Sponsored product prompts reports. Because the biggest opportunity that exists right now is learning, is we're at the beginning of the beginning. But understanding what prompts is Amazon surfacing? Because that's what Sponsored prompts is. It's the questions that preload on a PDP that Amazon has used and it's been informed by reviews and content and other questions that have been asked in the Q and A section of the page. It's valuable information just to understand which prompt has Amazon surfaced. Then the second question I would ask myself is how does my product value proposition respond to those questions? Is it microwave safe? Can it be frozen? Does it work on the moon? Whatever the questions might be that might be relevant to your given product, you can turn them off. And my recommendation is not to. And the reason for I would recommend turning them off if they were burning 25% of budgets, they weren't converting, they weren't driving performance. And you'd say, hey, there's a lot of ways to spend going out the door here for not much juice. But right now there's value in just learning, understanding what's going on there. This is a really tiny percentage of budget that we're seeing. Yes, that might change over time, but it's good to get into the habit to start to roll these reports into what you're regularly tracking. I think LinkedIn prefers to be fast than right a lot of the times, and I think it's important in the language when we think about sponsored product prompts and Rufus optimization, they're two different things. Sponsored Product prompts, um, is powered by Amazon's Rufus engine. But optimizing a PDP for Rufus is different. And that's because if you pull up, um, Rufus on your Amazon mobile app or on the desktop, uh, you're engaging with Rufus and the responses that it gives you when you're talking with Rufus, that is not sponsored product prompts. Sponsored product prompts is just what preloads on that detail page when you pull it up. And so, yes, they're both connected to Rufus, but because it's new and confusing, I think I'm seeing just conversationally in our space, those two things are getting conflated. But if I'm an actual brand that I want to optimize, those are two separate tasks.

Jordan Ripley: This feels very much days of yore when we did not have the search query, performance reports or brand analytics and I was like my best proxy to actually understand what people are searching is just look on the ad side of the business. And I feel like we're a little bit in that space with Rufus words. I'm not going to get all the data of, like how people are conversing with Rufus and understanding my products. At least give me the sponsor of these prompts and give me at least that little slice and I'll pay some money for it.

Pat: And one of the things we're all trying to understand is to what extent do people care about, uh, these preloaded prompts. Do they click on them, they look at them. And this finally at least peels back some level of transparency into understanding. But to what extent are people using these to inform their decision making?

Julie Spear: I use them. So there's one

Pat: zero.

Jordan Ripley: She's an important data point. Oh, that's a good lead into our next section here, which is really more on the shopper side, on the customer side of like how customers are using Rufus and how that's changing. And a lot of this, I would say, is based on the evolving capabilities of Rufus. We've seen a whole slew of updates in the last few months here, many of them previewed on the last earnings call. But just Rufus is becoming smarter, more intelligent, more adaptable in terms of interacting at various points in the shopper journey. We've seen for a while that any time I look at that native price tracker that Amazon unveiled on PDPs, it's putting me right into Rufus to have a conversation about pricing that has naturally led to some additional new features that I'd love to get your take on, Armand. One of those being auto buy. Right. I can set some pricing parameters and trigger automatic purchases. Second, being this whole Concept of scheduled actions. I might not be ready to purchase now, but let me set up some future purchases based on interacting with Rufus. All of this seems like a shift from uh, traditional oh search and buy. I'm in the browser bar. I'm going to figure things out to a more set and forget and expansion of what shopping looks like with Rufus. Armin would love a brief take on like how these functions actually work in terms of capabilities and what that means for brands.

Armen: Yeah Jordan, as you said it's getting smarter and smarter. I would say now it graduated from high school about to start smaller. So uh, yeah, getting there.

Jordan Ripley: What a Friday.

Armen: The first thing I want to talk about is the historical price information we now have within Rufus. It's been around a while. You can view it as a trend line and just see how the prices goes or you can just talk to Rufus and say is this a good price? Has this been cheaper recently? And stuff like that. The way I see it, one Amazon is leaning into this price transparency move. But then also a lot of shoppers used to do this research elsewhere using tools that are tracking prices like Keepa for example. So this is replacing that as well. So Amazon doesn't want you to leave, right? You're there and you can just find the information. And then this also ties back to what we've seen especially now when Prime Day not announced but like when we have the rumor now when it will happen, how they approached on the brand side, what you can do in terms of how much you need to discount will even get the be eligible for the promo for Prime Day. And this is all tied to how Amazon is tracking the price history. The other one. So you said auto buy. Very interesting. So it's you can just set a target price and say buy this product when it hits the specific threshold and when the price drops, Rufus basically takes the action and buys the product for you using your default payment method. The way I see this, sales are not happening in real time anymore. There can just be loads and loads of sales sitting there and waiting for a price trigger. And then out of not nowhere but like when you trigger a promotion, you get a mini Black Friday there and a lot of your units move. It's also a double edged sword in my opinion where yes it's great shoppers will have this feature of I will set a uh, threshold, I want to pay for this product. But then does that mean I'm losing shoppers more? Or if for example a lot of shoppers visit my product page and they just leave because the Price is not right. Will this tool actually be a better tool for conversion? Because now instead of leaving I can set, let's say, I don't know. M. If this drops by 10 bucks, I will actually buy it and then you just auto set it. So still TBD to see how this goes. But what I would love to have as somebody who manages accounts is having that data. What are my customers actually setting as thresholds? That data will be gold because then you can actually be strategic about it and trigger promotions at a time, know what the price points are, the ones that customers want and uh, stuff like that. And then the other one you mentioned scheduled actions. So this is a fun one. It's you just use Rufus as any other AI to be your assistant and help you either shop or predict something you would need. It's like a natural language subscribe and save version as well. And a step forward as well is the shop direct integration. So you can even have Rufus buy stuff off of Amazon and find find deals. I would say this is more of a play with this feature kind of thing. I don't know. You have an author that you really alert me when there's a book from this author when it comes out. I have a birthday for whoever on this date. Remind me of the birthday and give me a gifting recommendation. Why not? Everybody loves that. So a very welcome change, I would say.

Jordan Ripley: Yeah, I feel like every thing I've read about this is just trading on the fact that we live with 4% guilt that we're forgetting someone's birthday at all times. Because every one of these examples was like, what about a birthday? You could finally plan for that? And I was like, okay, I guess that's compelling.

Pat: It's interesting on the TBD piece about how target price, my first thought might to. I don't know if people use like E Trade or Robinhood or whatever, like your stock buying apps where you can buy essentially you can sell buy limits at certain stock prices. This isn't entirely different. Obviously we don't know. But I was thinking what's the average threshold or like the minimum threshold for deal discounts that you need to offer to make a Prime Day deal compelling? And it's basically 30%. Like anything lower than 30%. Customers on Prime Day go, that's not really exciting. I wonder if somebody would be more prone to setting, uh, oh, I wonder if I can get this for 10 or 15% off. And then could you potentially as a brand actually make that conversion at a less steep discount into your margin than that person who was otherwise going to wait for a 30 or 35 on Prime Day. There's three levels of speculation in that question. But it could be good.

Julie Spear: With all these updates, some things become more efficient and then hopefully our insights also get great. That's the goal. That's the goal. Anyway, let's move away from our pal Rufus, but still stay in our AI corner. We're not ready to come out yet. There were a couple other updates this past month. One was Amazon doubling down on its collaboration with Anthropic with an additional 5 billion investment. And then also OpenAI launched its own ad manager. From reports it looks like very similar to Google Ads. Pat, moving out of the world of retail media a uh, little bit. What do you see these bets and these updates? What are they telling us? What is this going to mean for investment in media through 2026?

Pat: Yeah, there's a lot to chew on in this one. And it's interesting because the sheer size and amount of money that is sloshing around, especially between the big players, has become normalized. Where we talk about 5 billion or 20 billion, we lose that running for the bus kind of thing. This is not the first time Amazon and Anthropic have entered into a partnership. They have had a partnership since 2023 as you mentioned. They're deepening that collaboration and part of that is a commitment from Anthropic to spend more than 100 billion. Again, just no big deal over the next 10 years. Most of these are tied to AWS. This is hitting Amazon's AWS side of their business. Really what this is differentiator for Amazon. And Google has this to some extent as well. It's built around their chips. They're making their own custom chips. For example, Amazon isn't going out and buying chips from AMD or from Nvidia. They're making their own chips, Trainium and Graviton. And please don't ask me any detailed questions about those two products other than that I know they're high powered chips that Amazon makes. They provide superior performance and so Anthropic is leveraging Amazon for that. I will get to your question, but I think related to this is at the same time also in the last week Meta uh and Amazon announced a pretty big deal and traditionally we think of Meta uh and Amazon as pretty fierce competitors, especially in the ad dollar space. Same thing Meta said who's got the chips that's going to power the next generation of our AI capacity? And they signed an agreement to deploy AWS Graviton processors as well. There's a long standing relationship between Meta and Amazon to power their AI. Same with Anthropic. Your question is what does this tell us about what does this mean for the future and what does this mean about where performance budgets are going to go? Obviously if I had for sure the answer to that question, that would be amazing. But, uh, the takeaway really is that performance budgets are going to follow where eyeballs go, where attention goes, where that attention is measurable and where that attention is incremental. The theory of the case here, and we're at the beginning of the beginning and seeing this, is that uh, there is, there are for sure eyeballs and attention that are going to the cloud, going to gemini, going to ChatGPT, and folks are doing research there to some extent, some level of buying there. As you mentioned, with OpenAI launching its ads manager, they're making the bet that we have traffic here and we can monetize that traffic. They also have a financial imperative to do just the amount debt that they've taken on. In order to finance some of these deals, they have to monetize that traffic. This is an annoying answer, but this is very much a stay tuned because what we don't have is still the precise amount of data to actually measure who is going where to do what at what stage of the customer journey. If in fact the reason that they've gone to that chatbot, uh, is for commerce to begin with. And what is the value of that relative to the value of that traffic that's already on Amazon or on Meta or getting through Google or influencers or whatever it might be. Obviously there's a lot of eyeballs on here because AI in real time is changing how people use the Internet to make all kinds of decisions. And with that attention, budgets flow.

Jordan Ripley: I also feel like there's some incentive here of OpenAI for such a long time was like a technology leader, a capabilities leader, and a lot of that has eroded from competitors. But Sam Altman and OpenAI have always been pretty cavalier of embracing the more capitalist instincts around this and the fact that they could be pushed to market with an ads manager, I think tracks with like, okay, we need to show some innovation and we need to show monetization and they're happy to do so.

Pat: And we talked about this last month too. Their challenge will still be how do we use this in a way that doesn't dilute or make voices? The model itself, Those ads almost need to live separate in a way that adds value. Um, and ads are not by their nature bad. Relevant ads that help answer a core customer question are good. And part of how effective this will be is less based on the model and more based on just like how the quality of their execution. And do they make the ads additive to the experience as opposed to diluting it?

Julie Spear: I appreciate you resisting the urge for hyperbole and just throwing out real predictions, but instead you're waiting on the date Google is dead. Okay, uh, we need that at some point. We're talking AI. It's all the hyperbole and huge exaggerated predictions of what will come.

Pat: So April 30, 2026, the beginning of the end of Google print.

Julie Spear: It now came for a hard left out of AI corner.

Jordan Ripley: Yeah, this seems like a good time to scamper out of there and we'll go out of the corner into the weeds with you. Armin. We, um, a trio of updates that I would think have caused a fair amount of grumbling. There was even, I think I saw an advertising boycott TBD on if that brings any real lasting change here. But there were a few things that rolled out in the past, I would say handful of months here, one being something called DD+7, which with a name like that how could we not get into it? The second one here, fuel logistics surcharge, third one, ads now being funded via proceed deductions versus credit card payments. All of these seem dry at firsthand, but I think have real meaningful impact for the people that are staking their livelihoods as third party sellers on Amazon. Can you talk us through what these mean, their impact and why there might be some grumblings happening?

Armen: Yeah, let me tell you, Amazon has always been so innovative when it comes to finding ways to squeeze more margin. But then when it comes to acronyms they're terrible. So obviously we know where their priorities lie. Yeah, DD +7, um, payout delay. So mid March or so, Amazon moved most sellers to this delivery date plus 7 days reserve. So instead of getting paid when you ship, you get paid seven days after the customer actually receives the item. And depending on shipping speeds, this can add 10 to 14 days to your cash cycle. So you're essentially giving Amazon an interest free loan while you wait for your uh, own revenue to clear the other one. So fuel and logistics surcharge effective just two weeks ago, it's a new surcharge that was added to fulfillment fees. So FPA multile channel fulfillment and also Buy with Prime. This was reactive. So it's triggered by the spike in global oil prices following the uh, conflict in Iran. But Amazon calls it temporary. But at the same time they haven't said anything about end date. It averages around 17 cents per unit. It sounds small but 3 to 5% of net margin for some of those high volume, low margin goods. So it can be meaningful to some brands. And then the uh, proceeds deduction for ads. So Amazon is moving ad billing from credit cards to proceeds deduction. Instead of getting a credit card bill later, Amazon just takes the ad money out of your sales balance before you ever see it. So this gets rid of the buy now, pay later benefit that a lot of brands relied on, having those 30 to 60 day delay to keep the extra cash on hand that brands used to use to grow or just improve their cash flow. Now Amazon takes that money immediately, the safety is gone and yeah, the brands are left with less cash for their daily bills. I think looking from Amazon standpoint, In a shaky 2026 economy, Amazon obviously wants to be safe and profitable. They don't want to be stuck with the bill for gas, high gas prices or shipping delays. So they added the surcharge also holding onto your cash longer. They're just moving the risk from them to the sellers. They stay profitable and then sellers take the hit. Also just by taking the ad spend directly from the sales balance, Amazon is ensuring they get paid immediately. So this is removing the risk of bad debt or just credit cards that kind of cannot be charged. Just they get to be flexible with cash flow instead of the brands. There are a few ways brands can work with this and just they have to adapt. So one of them is just remapping their cash flow. So for example, if you used to reorder inventory every 30 days, you might now need to do 40 days of capital on hand to account for the DD7 delay and loss of the uh, ad spend float for the advertising. There's also an opt in into pay by invoice. So this gives you a 30 day window to pay for your ads instead of Amazon taking the money from your sales immediately. So this is the only way you can get some um, leeway and some cash to work with and keep it longer on your bank account and then for the surcharge. I think the inevitable will eventually happen, especially for those smaller, tighter margin products and brands that sell low ASP products. You'll just have to adjust your MSRP to account for this.

Julie Spear: I think we're all just waiting. We know it's coming, the price increases as a result of the gas shortage and all of that. So that one makes sense. Let's move on to out of surcharges and maybe to uh, a little bit lighter and maybe more interesting. I don't know. Surcharges are very compelling. But in the world of Instacart, we've seen a lot of updates recently. The creative AI tool. We've also seen occasions and also a pilot for bundles on Instacart. Pat, talk to us about this. Is this Instacart building out more and more options for campaigns? What does this mean?

Pat: Broadly speaking, every retail media network is looking for ways that they can leverage AI, uh, to remove barriers that have historically and continue to exist for advertisers to open up their budgets. And as we all know in this space, creative and creative generation has always been a friction point. It's sticky, it takes a long time, it's hard to create at scale, it's hard to a B test because you need iterations. And so what, uh, Instacart has done, as you mentioned, they're invited a small group of brands to participate in early. It's called AI Creative Tools now. It's not lost. I made that, um, at Amazon Unboxed in October of 2025. They highly celebrated their AI Creative Studio and Creative Agent, which included the AI Ads Generator. But anyway, through this pilot with Instacart, a brand's team can submit an existing image or baseline creative. So essentially, okay, I have assets on hand. Maybe I've used them for other creative, maybe I' used them for vertical video, maybe I've used them for other digital platforms. And the Instacart team, using their AI tool will resize, restyle, enhance using their own internal tool so that they're ready to run and fit the specs specific to the Instacart ads ecosystem. You would suspect that Instacart is also probably using with nanobanana or using some other tool that would probably to me make more sense than having them have them build their own. So they can resize existing assets into 16 by 9 for shoppable display, for recipes, for occasions, bundles. As you mentioned, they can do object removal or additions like cleaning up images and adding elements. You can go in and say, oh, I want to kind of add this badge here, or can I see what this looks like on a nice uh, wooden table out in the forest on a cabin, something like that, highlight your products and then I think a big one is restyling for seasonal moments. So it's okay. Let's say I sell a fire pit or something like that. Do I want to have it for a summer Barbecue, do I want to show it outside when the leaves are turning for fall, is it wintertime when people are all bundled up, et cetera, et cetera. So restyling product images for seasonal moments. And overall, again, I think this is Instacart is fighting for budget the same way as other, uh, retail media networks are. And I think it's particularly important in for shoppers who are on Instacart and the type of purchases that they're making, that it's really compelling. Creative is like the key conversion driver there. And so the point of this is to make that easier and so that brands, if brands are like, hey, I'm optimizing creative for Amazon and for Google and for Meta and for whatever. And then by the time they get to Instacart, maybe that's the fourth or fifth retail media network lit on their list Even after their D2C, Instacart might not be getting as much love in terms of creative generation and this unlocks that.

Jordan Ripley: Yeah, that all makes sense. I want to pivot over to one more retailer here while we have the time, which is this announcement that we saw from Walmart where traditionally we've had the marketplace sellers and that massive assortment for via 3P, we had the sort of assortment that you could buy online, pick up in store or offered 1P or directly in stores themselves. Those have existed in kind of separate islands for years and years. This wall has started to crumble a bit in that we now heard, at least they're testing this in Dallas, that there will be some assortment that offered through WFS of 3PMarketplace SKUs that I can then pick up in the store and actually being stored in the back of brick and mortar locations there. So Armin, like, curious, how big of a shift does this indicate in your mind, like what's the benefit for customer experience and how does this position Walmart?

Armen: Yeah, when I read this, like, finally this is like the biggest moat for Walmart is like so many locations, so many stores, like, why are you not utilizing that? Glad to see they're pulling the trigger. Hopefully this expands as well. So if you look in the past, if a customer bought a marketplace item on Walmart.com, it ships from, um, the typical fulfillment center using WFS or the seller's own warehouse. So it was mainly E commerce, but now Walmart is staging a, ah, marketplace inventory in the back rooms of local stores. So popular items from FS are moved from big distribution centers to the back room of specific local stores. So when a local customer orders that item. It doesn't have to come on a FedEx truck from three states away, just picked up by a store associate and delivered by a local driver. This means better speed and also better ability to scale because as we said Walmart has so many stores across the country. I read this fact where it says uh, like 90% of Americans live within 10 miles of a Walmart. That's huge and something that I feel Amazon can't even replicate at least right now. So it's the biggest advantage Walmart has. And yeah, as I said, just finally glad that they are pulling the trigger. We'll see how it goes. Basically when it comes to brands, I only have two recommendations right now. One is used wfs. So in order to basically qualify for this before selling on Walmart is number one. But then use WFS because this is your ticket to actually get into that local delivery game. And then of course Walmart will only focus on high velocity SKUs. So you cannot expect that your slow mowers will be taking space in store backrooms. Just pick your top five to 10 bestsellers and just go all in on WFS to trigger these local delivery badges.

Julie Spear: Totally agree. The longest time coming update I think we've ever shared kind of wild. Let's wrap this up and do a little hype meter action and let's do it rapid fire because we had a lot that we dug into today, both Pat and Armin. We'll start with you Pat. What's your high? What would get a 10 or at least net promoter score like an actual promoter? And what would get ah, a uh, detractor.

Pat: What's not hypable Rufus overall would be my 10 out of 10 hype meter. I think it could have long term as big of an impact on the space as the invention of Amazon. DSP1 is the least hype meter. I think there's a lot with the different multi hundred billion dollar deals going back and forth between the chip makers and the platforms, I think it's going to be winners for them. How much is going to actually impact the customer retail experience? Not so much.

Julie Spear: That's a bold assertion that Rufus might be bigger.

Pat: But you added bigger. Did I say bigger?

Julie Spear: That's what I heard.

Jordan Ripley: Okay, 18 out of 10 from that

Pat: because I believe we're only seeing a tip of the iceberg of what Rufus is going to be. I think what we think of as Rufus on April 2026 is nothing compared to what we will see Rufus as six months from now.

Julie Spear: That's the hyperbole I was looking for so thank you Answer the klaxon Armin how about you?

Armen: I think Pat wanted to say bitter not bigger but okay Rufus for me as well and yes I see like how it's evolving and I like how it's advancing but I like how it's fun it's just making things easier things integrate it's easy like how you win right so that's a plan for me I love that the Walmart update as well but I don't love Walmart as much as Amazon my lowest hype is just the Amazon squeeze 100 in three years I don't know how many they've had but yeah it's been it's enough

Julie Spear: really I appreciate all the hot takes mid takes all the takes that were shared and your insights and of course we'll have you back next month where we will dig into more I'm sure with Rufus AI and hopefully not more surcharges

Pat: thanks all.