EP321 - Amazon Q3 2024 Earnings Recap
In today’s episode of the Jason & Scot Show, Jason and Scot dive into the latest developments shaping retail, tech, and consumer trends heading into the 2024 holiday season. Here’s a breakdown of the topics covered in this jam-packed discussion:
Amazon Q3 2024 Earnings Release
Amazon Q3 2024 Earnings Call Transcript
🎉 Holiday Season Sentiment & Retailer Anxiety
• With the holiday season shorter than usual due to a late Thanksgiving, retailers face the challenge of fewer shopping days. Consumer behavior trends indicate a shift toward essential purchases over non-essentials, creating mixed expectations for holiday spending.
• The impact of the election is expected to influence consumer sentiment, media spending, and holiday promotions, with Amazon and Walmart predicted to perform above market averages.
📈 Amazon Q3 Earnings Highlights
• Retail & GMV: Amazon’s retail revenue surged by 7.2%, with U.S. gross merchandise volume (GMV) rising 9.9%—nearly three times the industry average.
• Efficiency Improvements: Amazon’s focus on fulfillment efficiency under CEO Andy Jassy is paying off. The company’s same-day delivery options and regionalized inventory system have led to a 25% improvement in fulfillment cost efficiency.
• Growth in Essentials: With increased demand for everyday essentials, Amazon is capturing market share from traditional pharmacies, offering same-day delivery for prescriptions in select cities.
🏢 Amazon’s New Store Concepts
• Whole Foods & Amazon Grocery: A new Amazon grocery concept opened in Chicago, catering to convenience items like packaged snacks and sugary drinks, which contrasts with Whole Foods’ health-conscious inventory.
💸 Amazon’s Profit Engines: AWS & Advertising
• AWS: With a 19% increase in AWS revenue, Amazon is now operating at a 38% margin. Demand for AI-powered compute continues to push AWS growth, even as it faces GPU supply constraints.
• Advertising: Amazon’s advertising revenue reached $14.3 billion, growing 19% year-over-year. With a nearly 70% estimated gross margin, advertising may soon outpace AWS in profitability.
🛒 Rise of AI-Powered Search & Perplexity’s Native Checkout
• Perplexity’s Surge: Scot shares his switch from Google to Perplexity as his primary search engine, noting the emerging competition from OpenAI’s ChatGPT search. Perplexity now includes shopping links, allowing users to check out directly through Amazon, hinting at a new era of AI-driven shopping.
• Impact on Retailers: Retailers need to rethink SEO strategies as search shifts toward AI-powered “answer engines” that may fundamentally change how products are discovered and purchased online.
📡 What’s Next for Alexa?
• Amazon’s next-gen Alexa, powered by large language models (LLMs), faces delays into 2025. Scalability and usability challenges highlight Amazon’s shifting internal dynamics and potential headcount reductions in its Alexa division.
🔍 Is Google Search Under Threat?
• Perplexity’s and OpenAI’s expansion into search could spell trouble for Google. With monetization still in the early stages for these answer engines, the retail industry is watching closely to see how they’ll shape online shopping behavior.
⭐️ Tune In, Subscribe, and Leave a Review!
If you enjoyed the episode, help us reach more listeners by leaving a five-star review on Apple Podcasts. Thanks for tuning in!
Episode 321 of the Jason & Scot show was recorded on Monday, November 4th, 2024.
Join your hosts Jason "Retailgeek" Goldberg, Chief Commerce Strategy Officer at Publicis, and Scot Wingo, CEO of GetSpiffy and Co-Founder of ChannelAdvisor as they discuss the latest news and trends in the world of e-commerce and digital shopper marketing.
Transcript
Jason:
[0:23] Welcome to the Jason and Scott Show. This is episode 321 being recorded on Monday, November 4th, 2024. I'm your host, Jason Retail Geek Goldberg. And as usual, I'm here with your co-host, Scott Wingo.
Scot:
[0:38] Hey, Jason, and welcome back, Jason and Scott Show listeners. Well, Jason, you've been out there traveling more than I have. What are retailers saying about the holiday? Are they lots of excitement? Are they more worried about the election and they'll worry about the holiday? what's going on out in the retail end.
Jason:
[0:54] Yeah i feel like the the like webster word of the year is going to be anxiety obviously we're a day before election day and everybody's anxious half the country is going to be depressed no matter what happens there but i feel like retailers are also anxious about holiday it's a weird thing we won't we don't have to get into the whole thing but it's the shortest holiday season we get uh thanksgiving super late so it's five days shorter than last year. So there's fewer days to spend on stuff. Consumers have been spending slowly, trading down the lower priced goods, buying more everyday essentials and foregoing non-essential purchases for a while. And then the election is a major distraction. It also is really hard to buy media and break through all the noise. You live in a swing state. So I assume you haven't been able to turn on a tv or look at your phone for several days um yeah.
Scot:
[1:54] Wall-to-wall and three inches of mail every day it's crazy.
Jason:
[1:56] Yeah yeah and so like like even if you're not someone that's super focused on the election like it just like the black friday ads just don't break through all that noise and it's more expensive to put that by the media for those ads and all of those things so you know there's some there's some historical precedent in previous elections that you tend to get a little bounce in spending after the election, but no election's been this polarized before. So I'm less confident that history tells us exactly what's going to happen here. And then there's this huge vibe session thing still going on where, you know, people are generally pessimistic and anxious about the economy and, you know, laser focused on the price of goods and spending less, even though honestly, like most of the macroeconomics are actually like pretty solid. Like there's a lot of economic good news that you would ordinarily expect to translate to consumer confidence and more spending. Now there was a slight uptick in consumer confidence for the first time in a while this month. So there's plenty of tea leaves, however you want to interpret them, but there there's very little certainty. My I look at all this and I say, on average.
Jason:
[3:15] The biggest retailers that are, you know, best at really being highly efficient and low cost. So the Amazons and Walmarts of the world are going to do reasonably well. They'll outperform the market. They'll have lower margins and a bunch of specialty retailers and big box retailers and people that aren't quite as efficient are going to have a really hard time. But, you know, I hope I'm wrong.
Scot:
[3:39] Yeah. Wow. Lots of wasn't there. So we got through the the longshoreman thing. Right. And then isn't there a big budget thing where the government could shut down or are we is that been kicked into.
Jason:
[3:53] It's kicked till post-election. It may still be in, there may be another budget thing in December. But again, like, depending on what changes in the House and Senate, like, that could be like a trivial additional continuing resolution or, you know, someone could go nuclear and try to shut down the government. So I think it is a risk.
Scot:
[4:17] Yeah. Okay. I don't know. We'll see. So lots of mixed messages out there for retailers to try to parse there.
Jason:
[4:25] Yeah, exactly. But we at least know what happened to Amazon in Q3.
Scot:
[4:36] Amazon News. Your margin is their opportunity. Yeah, yeah. So on the eve of Halloween, Amazon announced their results. And coming into that report, Microsoft had reported kind of slowish cloud growth. Turns out they're having a hard time getting enough GPUs and they're really having to stretch to handle all their customers' needs, which you would think would be positive. But Wall Street is kind of because they also see this whole mixed signal thing. All they care about right now is short term signals. So they were kind of down on Microsoft. And then everyone and then Meta also reported and they had a great quarter, but then they said they're going to really ramp their spending on GPUs going into next year. So there's this really big kind of wall of worry around all this money on these GPUs, which are for AP.
Scot:
[5:28] Artificial intelligence or ai and they're mostly nvidia chips so nvidia is off cycle so that i think they report in the 20s like right before thanksgiving so i'm going to be watching that one carefully so you know amazon is largely driven these days by aws so that was kind of feeding into that there's a lot of concern that amazon was either going to miss for a similar kind of thing like microsoft or they were going to announce they're just going to like buy an ungodly number of GPUs. So there was a lot of, I believe the stocks sold off going into the report, which was interesting. Usually Amazon kind of lifts a bit the day of the report. So those ended up being unfounded. They achieved an 11% overall revenue growth and operating income hit 17.4 billion and beat forecast by 18%. And that was the seventh consecutive quarter of operating income above guidance, which is good. So Jassy kind of like reset it low enough and has gotten better now that he's CEO of kind of managing expectations is kind of how I read that. And then the looking forward guidance was very well received. So looking at Q4, they kind of gave a band of seven to 11% growth. When you have 11% and you're going into Q4 and you're kind of showing 11%, that's good because a lot of people worry that they're going to project a slowdown Because when they do this, they've already got a month's worth of data. So in a way, Amazon being Amazon, they're pretty good at reading these tea leaves and kind of figuring out how the holiday is going to shape up.
Scot:
[6:58] A couple of things I thought in the report I just wanted to call out for listeners. We've talked about this a little bit. This kind of really part of Jassy's era has been efficiencies. And he's really dug into the fulfillment part of Amazon. and it's been interesting to see the results. So they had a lot of progress on the cost of fulfillment and they attributed it to three areas. Number one, we've talked about this before, but they had this new regionalized inventory system where they're just getting better at spreading inventory and predicting demand across their fulfillment centers. So if they can predict, hey, it's getting cold in Chicago, we're gonna kind of balance some cold weather stuff there or anything in that regard. Obviously, that saves the money versus having to ship stuff across the country. So they've improved that over the last year, they've improved this system 25%. And it's interesting, I triple checked this because I thought I was hallucinating. I thought I had LLM bring, but they've, you know, many of these, these have resulted in 25% improvement. So this is not a typo.
Scot:
[8:04] The, so when you get the product closer, not only is it lower distance to the customer, therefore lower costs, it's like in a closer quote unquote zone, but it also increases the number of products they can put in a box. So there are a lot of customers like you to get a daily Amazon box. And the more they can put in that JSON box every day, the better off there.
Scot:
[8:25] Then they have really expanded the same day facilities, which results in more of the products being available same day. So they now have over 40 million customers a quarter benefiting from same day delivery. That's up 25% year over year, that number of customers that are having access and getting products delivered same day. And to deliver on this, they've built a bunch of new facilities and a lot of new processes increased. And that ends up really increasing customer satisfaction, conversion rates, and speed to reorder. And then the last one is robotics and automation. There was a lot of talk on the call and then some various notes and in the Q&A from Wall Street around this one fulfillment center in Shreveport. And that's what they call a quote unquote 12th generation system. So I think it's like their showcase where Amazon has put all their best automation ideas into this fulfillment center, and it is showing really promising results, which gets Wall Street excited because if they do it one, they can scale it across their network of north of 200 fulfillment centers. It'll take a long time, but these are the types of improvements that are pretty material. It also saw that fulfillment center and the things they do in there, which is lots of robotics and new ways of automating. It had a 25% reduction in the cost for fulfillment center cost per unit.
Scot:
[9:43] That's pretty material. So let's say it costs $6 to ship one of these things. You can knock a dollar or two of that off. And this is Amazon and you're shipping some billions, tens, if not hundreds of billions of products. That is just basically margin that kind of falls to the bottom line.
Scot:
[10:03] So together, these did help with that earnings beat. And they caused a somewhere in the four to 6%, let's call it 5% underage in what Wall Street was thinking shipping, total shipping spend would be. So, you know, the good news is they're making progress on these things. And it feels like it's still early days for, I think the regional, they're probably kind of at the end of that one. I think same day facilities, they have a lot more they can do there. And I think robotics, they're very early. So it feels like they've got another two or three years of runway for improving the economics of fulfillment.
Jason:
[10:39] Yeah, I think in the press call after they released their earnings, the CFO even said that they believe there's headroom in all three. So they think they're going to continue to get additional efficiencies. And they talk a lot about those three things together, like fundamentally reducing the cost to serve is kind of the language for talking about all this stuff and what the average price is to get a box of goods to Jason's house.
Scot:
[11:09] Yeah. Another theme that you'll kind of hear as Jason, I go through some of the pieces is this has really dramatically increased consumers ordering replenishables. So, so we'll talk about that.
Jason:
[11:20] Yeah, yeah. So obviously the most interesting and important segment of Amazon's business is the retail part of the business. The numbers were all pretty favorable. So they hit $61.4 billion in revenue, which is a 7.2% year over year growth. It's a significant acceleration from the 4.6% growth in Q2. And it's a significant beat over Wall Street's expectations. It's like 3% over. So I think the revenue number was very encouraging. And a number of other retailers had reported and were saying that there were a lot of signals of the consumer slowing down. So this was, I think, surprising and well-received. Of course, listeners of this show will know you probably shouldn't care too much about revenue because revenue is a blend of the cost of goods for 1P sales and the take rate for 3P sales. So when we kind of are thinking of them as a pure retailer, the number we really like to know about is their GMV, their gross merchandise value, which is the value of all the goods they sell, whether it was on the marketplace or first party.
Jason:
[12:31] And they don't tell us what that is, but…
Jason:
[12:34] Citibank has released their estimates. Morgan Stanley's released their estimates. A couple other sites have released them. And in my pull of polls, the Morgan Stanley numbers are kind of right in the middle and most believable. And they have Q3 GMV in the United States going up 9.9%, so 10% essentially. In that same quarter, we know from the US Department of Commerce that core retail went up 3.6%. So Amazon grew almost three times as fast as the industry average, which, you know, Amazon's number one or number two are just retail out there. So that's super impressive. That's slightly lower GMV growth than they had in Q2, which was 11%. The number we'd most like to compare this to would be Walmart's number. Walmart doesn't report Q3 till November 19th, I want to say. So, so it'll be super interesting to look there. I would expect Walmart to come in somewhere between Amazon and the retail industry average. So above the retail industry, but below this 10%. So that in and of itself is super interesting. You know, Walmart probably beats the industry average. Amazon triples the industry average. Then we still have Timu Sheehan and TikTok shops out there growing faster than anyone's ever seen before. And so you've got, you know, Those five horsemen eating up most of the growth in the retail industry.
Jason:
[14:03] One thing that did negatively surprise people a little bit was the mix of, which you alluded to, is the mix of 1P and 3P sales. So ordinarily for a long time, 3P sales are creeping up as a bigger and bigger chunk of the overall mix every year. I think it was 61% last quarter, maybe. And it dipped down this quarter to 60%. So that's the first time 1P sales have gained share over 3P sales at Amazon. And the management tells us the reason for that is increased demand for everyday essentials, right? So that's the cleaning products for your house, the, you know, affordable skincare products, the things you use in your bathroom and shower. Or all of those products are more often sold 1P than 3P on Amazon.
Jason:
[14:57] And, you know, both because of the vibe session and the economic situation that I talked about earlier, and because Amazon's better at same-day delivery, they're winning more of these everyday essential trips. And that is shifting their mix slightly to 1P. In theory, it should also be eroding their margins because these items are lower margins. But Amazon did so well on the efficiencies and the cost to serve that you covered that those efficiencies more than made up for the lost margin from the slightly less profitable mix of products that they're selling with everyday essentials. So that's a super interesting trend. When you hear management talk about it, they're like, we love these everyday essentials, even if they are slightly lower margin because those customers are way more sticky and we get way more wallet share in the long run and the the more those orders we get the more volume we have to drive our.
Jason:
[15:59] Cost to serve down and the lower our cost of serve, the more we can profitably fulfill all of this everyday essentials demand. And so this is a sort of a new flywheel for Amazon, if you will. But like one tangible proof point on this is they said something like when a customer sees a one-day delivery promise or same-day delivery promise, conversion rate is 20% better than when it's a two-day delivery promise. So kind of, you know, giving evidence to the fact that like consumers have an insatiable appetite for faster and faster delivery and the faster and cheaper they can deliver, the more they can sell.
Scot:
[16:37] So- Where do you think they're taking that share from you? That's like Target and Walmart or like, you know, do you see any evidence of it?
Jason:
[16:44] Target is totally possible based on the last couple of earnings calls, some grocery, the super vulnerable place that we'll talk about a little bit later is the traditional pharmacy. So Walgreens, Rite Aid, CVS, and then all the independent pharmacists. They're for sure taking share from those guys. And they talked about that in the earnings, which in pretty funny language. So I'll get to that in just a second. But it doesn't feel like it's coming completely at the expense of Walmart, because again, both Amazon and Walmart are growing faster than the industry average, but it's pretty much everyone else that sells everyday essentials that's dipping below. Side note, we're now seeing these Chinese direct-to-consumer companies start to lean into everyday essentials. So Timu has a ton of everyday essential goods on there. So the world has noticed that that's what consumers have an appetite for spending on right now.
Jason:
[17:40] Interest rates are still a little high. People aren't moving houses. So people are not buying as much home improvement goods as they normally do. They're not buying as much, except for you and I are not buying as much consumer electronics as they normally do.
Jason:
[17:52] And they're spending more on the center store grocery assortment. And so that's, you know, Amazon through great insight or luck or whatever combination you want to attribute it to, that shift to regionalization of their fulfillment center and this like laser focus on cost to serve has really positioned them to take advantage of this trend, where arguably four years ago, they wouldn't have been well positioned for this trend. They would have said this, that's all stuff that craps out that we can't realize a profit on and that we're discouraging sales on. Now they're encouraging sales on all this stuff. So that's interesting. And then the slightly, if there's a negative sentiment in the industry regarding all of this, it's, hey, 3P sales ticked down slightly, but the revenue that Amazon is earning from 3P sales went way up because Amazon has found a lot more fees to charge sellers than ever before. And so, you'll hear a lot of noise from the 3P seller community that they're, The overall, not the raw take rate, but the overall cost to sell on Amazon when all these fees and marketing services are factored in has never been higher and makes it harder than ever for three-piece sellers to truly be profitable.
Scot:
[19:14] Yeah.
Jason:
[19:14] So that's the story on online sales. I mentioned pharmacy another trend happening in our industry is you know the way pharmacies normally work is the traffic to a pharmacy is to get pick up your prescription and then while you're there you go oh I need paper towels and you buy paper towels by design they're the worst place to buy paper towels they don't have a great assortment and they they don't have a good price but you're already there to pick up your Lipitor prescription so you know you just want to save yourself a trip and you pick and you take them home from there. So the front of house of a pharmacy only works because of the back of house. And the back of house is under huge stress. There's like three companies called pharmacy benefit providers, and they get to dictate the price that a pharmacist gets reimbursed for every prescription they fill. And these pharmacy benefit providers are reimbursing the pharmacy less than the pharmacy is able to collect from the consumer. So on a lot of popular prescriptions, pharmacies are now at negative margins and all the chronic prescriptions, if you take something every day or every week, instead of getting a 30 day supply from Walgreens, that pharmacy benefit provider is making you get a 90 day prescription from mail order.
Jason:
[20:35] So people are going to the pharmacy to pick up prescriptions less, which puts more stress on the front of the store. And then, of course, their lower overall profit, that 10% loss of traffic has a huge impact on their profits. And then they start freaking out about shrink and they lock up all the products and that chases even more customers out of store. And the pharmacies are kind of in a negative flywheel from all this. And Andy Jassy summed it up perfectly in one sentence on the earnings call. He said, brick and mortar pharmacies often require customers to make trips to forlorn physical venues with much of the selection behind locked shelves, wait in line for meds, and only to find out about pricing at the point of purchase. The largest mail order pharmacies offer delivery in five to 10 days. We think customers deserve better, which is a clear shot across the bow at the pharmacy industry. Amazon has doubled the number of cities that they do same-day pharmacy delivery to. But I actually think the big story here is not Amazon's prescription business. It's Amazon getting all this everyday essential traffic that used to go to the pharmacy in conjunction with the prescription business.
Jason:
[21:52] That has other healthcare implications in the long run. a lot of people in America go to a pharmacist to get a vaccine. And if all these pharmacies close, you know, we got to figure out how we're going to get those vaccines and things like that. They, of course, do break out their physical stores. Physical stores for Amazon is almost exclusively Whole Foods at this point, and maybe a few Amazon Go stores left. And physical stores grew like 5% year over year, which again, on itself, that's better than the industry average. that's better than most other big grocery stores are growing. And that was also a beat versus the analysts that thought they were gonna be at 5.16 billion.
Jason:
[22:32] Same story there. The mix is shifting to more essentials. And what's fascinating and interesting and controversial here is Whole Foods is a unique grocery concept. They have all of these food quality guidelines that dictate what products they will carry and what products they won't carry. And so they won't carry products that have artificial colors in them. They won't carry products that are sweetened with corn syrup, for example. And so the assortment in a Whole Foods is very different than the assortment in a Kroger or an Albertsons. And a lot of the Everyday Essentials volume is traditionally those products that Whole Foods chooses not to sell. So the fact that they're doing better at Everyday Essentials is interesting. But what's super interesting is in Chicago, where I live, Amazon opened yet another new grocery concept this month.
Scot:
[23:25] Which is now like the fifth or sixth one.
Jason:
[23:27] Yeah. So, yeah. So now called Amazon grocery. So they have, of course, Whole Foods, they have Amazon Go, they have Amazon Fresh. A few of these, they have multiple iterations of Amazon Fresh used to mean an online grocery store, and now it means a physical grocery store. And now they have this new concept, Amazon grocery. And the one and only Amazon grocery store is downstairs from a Whole Food in downtown Chicago, and it carries all the unhealthy stuff that Whole Foods isn't allowed to sell.
Scot:
[23:59] So you go down for your Rice Krispie Treats.
Jason:
[24:01] So it's literally, they carry like 3,500 SKUs and it's mostly Kraft and Hostess and Coke and Pepsi, right? And so it seems superficially, it seems very explicitly like go upstairs and get your organic produce and feel good about your purchases there, and then stop on your way out and get your guilty pleasures.
Jason:
[24:25] From Amazon Grocery on the way out, right? And so, It's unclear. Did Amazon already own this real estate and it's a convenient place for them to test it? And Amazon believes that the world needs another, you know, small format grocery store with 3,500 SKUs. That strains credulity a little bit. Or is this an overt, you know, first step at eroding the Whole Foods quality promise that was, you know, part of the core mission when John Mackey started the company and the Amazon said they would respect when they took it over? Like don't know it's super interesting i would say you know i got back from grocery shop three weeks ago amazon was on stage at grocery shop and they talked about like you know how much they're leaning into grocery and they and how synergistic all these grocery offerings are and how they're unifying them all and then the week later the ceo of grocery resigned at amazon and they and this they revealed this new concept and it seems totally muddled to me so i'm not saying Amazon can't win at grocery, but it does not seem that there's a crystal clear hypothesis from Amazon as to what a winning grocery concept looks like at this point.
Scot:
[25:39] Have you been in it?
Jason:
[25:40] I have, I have. And in fact, I'll put a link in the show notes, but I uploaded a little video tour on LinkedIn, so you can, you can go check it out. It's, I would call it most similar to an Amazon Go store without the tech. So there's no just walk out. There is Amazon.
Scot:
[25:55] It's like a convenience store kind of stuff.
Jason:
[25:56] Yeah. It feels more like a 7-Eleven. There's a cafe, a small espresso bar in there. There's an Icy machine, so you can get your non-Swerpy branded Icy. And yeah, all the Twinkies and Apple Jacks you want.
Scot:
[26:12] Nice. You have to hit that first before you go out to Whole Foods because by then you'll spend your whole paycheck. You don't have any money left for Rice Krispies.
Jason:
[26:20] I would feel personal shame dragging that stuff around in my cart on my Whole Foods shop. So I feel like it's more likely to be on the way out to avoid the shame. But in fact, I took all these funny pictures because there's all these pictures about the ethical principles at Whole Foods. And it's like giant wall-size signs that say, eat good food. And then you go downstairs and it's like, buy more, get more Twinkies.
Scot:
[26:49] Maybe it's a higher conversion rate for the people that are cheaters, Whole Foods cheaters.
Jason:
[26:55] Yeah. Yeah. Again, there's probably a small cohort of people that Whole Foods is their own grocery store, but I feel like there's a lot more that are like my household where Whole Foods is definitely in our mix and there's a lot of products we only get from Whole Foods. But I have a nine-year-old son and like a lot of his preferences do not exist in the Whole Foods environment.
Scot:
[27:13] He's not eating tofu potato chips.
Jason:
[27:15] No, no. He wants the good Pringles.
Scot:
[27:20] Cool on the aws side revenue rose 19 for 227.5 billion which kind of met expectations but they were just a little bit higher so it was like right in line so that wall street was thinking 20 to 21 but they came in at 19 so they didn't really get much heat over that because they did talk about the strong demand they've they've started talking about a backlog or kind of a bookings backlog so So that was interesting. I saw a wall sheet analyst kind of do this whole thing around it and felt like there's a lot of positivity in the number that they gave there.
Scot:
[27:55] And yeah, so the margin there was very nice. So 38% operating margin. That's not even a gross margin. That's a net margin. And they saw their own benefits from cost efficiencies. They're getting more life out of their servers. Imagine what's going on here is those workloads that are not GPU bound. They're probably flat right now because everyone's obsessed with moving everything over to GPUs. So they're probably having to spend a lot less on traditional CPU type things and getting more life out of the machines than maybe they initially had thought they would. Let's see the they were also supply constrained they did call out they have their own chip that can do some inference and they're they're deploying those as quick as they can and they're also supply constrained on the nvidia chipsets they did say they're gonna spend you know they prepared wall street to buckle up for more capex that were on this one time in a lifetime demand curve so that was interesting and but no one really freaked out about it so that was good, what'd you see on the advertising business?
Jason:
[28:58] Yeah. Well, I would just add that, that 38% margin that was up from 30% the quarter before. So that, that was a.
Scot:
[29:05] Yeah, that's material. Yeah.
Jason:
[29:06] Uh, very, very meaningful. And, and I think nobody freaked out over the CapEx because they're like, Hey, we, we thought like regular AWS was like a pretty good demand curve. And, and the AI AWS demand curve is ahead of where AWS was at this point. And so I think they think there's a lot of headroom. It was interesting. They used to talk a lot about how we really only have scratched the surface on cloud compute that like, you know, some ridiculously low percentage, like only like 5% of all the workloads are in the cloud at all. So, you know, traditional AWS has all this headroom. And I assume they still believe that's true, but they don't bother even talking about it because there's so much demand for all this AI compute at the moment. So advertising, I always love talking about right after AWS, because I feel like there's this common notion that AWS drives Amazon. And you said it earlier, like all the analysts think that Amazon's at AWS with some annoying side projects. So the advertising business also had a good run. Exact same growth rate, 19%. The difference is 19% is an acceleration of their advertising business, where it's a site deceleration of their historic AWS rate.
Jason:
[30:26] So that 19% growth gets them $14.3 billion this quarter. So if you look at their last four quarters, they're now at a $58.7 billion run rate, or not even run rate, trailing 12 months. So nearly $60 billion in ads. To put that in perspective, the trailing 12 months at AWS is like $100 billion, $102 billion. So AWS is a bigger business in terms of revenue, no question. But the ads business is growing fast. They're adding a lot of advertising products, like almost all this advertising is still just in support of search. It's mostly sponsored search ads on the website. But of course, Amazon is winning more eyeballs with their media, right? And their NFL stuff is going really well. Their original programming is going really well. And they're increasingly putting ads in all that content. So like this really is like $60 billion of search ads. And there's still a lot more headroom in, you know, monetizing all of this other media that Amazon is increasingly succeeding in. And so that's interesting. Unlike AWS, which is highly CapEx intensive and, you know, constrained commodity because of these silicone chips, the ads are very low CapEx, right? And so they don't tell us what the gross margin is for the ads, but let's call it 70% gross margin. There's no cost of goods, right?
Scot:
[31:55] Yeah.
Jason:
[31:55] Let's call it 99%.
Scot:
[31:57] Okay.
Jason:
[31:58] Very popular. Yeah. They pay a lot of sales, guys, right? Like that's the one, that's, that's the one cost. And I think those are the only sale guys that don't get fired if they don't go to the office, by the way. But so 60 billion, 58 billion in revenue at 70% gross margin is 41 billion in earned income, operating income that they've made over the last 12 months for the advertising business. If we go back and apply the 38% gross margin, which is the highest gross margin they've ever achieved and not what they actually achieved over the last 12 months and apply that 38% growth margin to the last 12 months of revenue on AWS, it earned $39 billion worth of operating income. So like in the most conservative version of this, advertising is contributing more dollars of profit to Amazon than AWS's and growing faster.
Scot:
[32:48] Yeah. I 100% believe that because the cost structure for AWS is pretty, pretty big.
Jason:
[32:53] Yeah. Yeah.
Scot:
[32:54] And it's going to get worse, which is amazing that they popped up to 38.
Jason:
[32:59] Yeah, yeah. I mean, it's a testament. They do a lot of stuff really well. So that is the advertising story. Scott, I know the future of Alexa is very near and dear to your heart.
Scot:
[33:12] It is. And I've had some interactions with Amazon in like the last year. I won't go into specifics. And I came away thinking, man, the company seems like different. Like they're just like not as engaged and the people you meet are very… Company kind of feeling, which is way different than, you know, when I was at Channelvisor dealing with Amazon, you would meet a VP and he or she would know exactly how everything at the company worked. And it was like freakishly, you know, freakishly informed about everything. So, you know, case in point, they're coming out. They've been working on this for a long time. At first, it was going to be kind of, you know, late summer, fall, and then they moved it to kind of late to catch holiday. And now they just announced that they're delaying till 2025, that they're not going to have any kind of LM type technology on Alexa. They started with having, saying they're going to do their own in-house kind of a model. Then they moved to Anthropic. But what they're doing is they're finding some problems with it that in the beta testing, I read an article that said it likes to show off. So instead of just answering a question, it will give you all this extra expository. It just kind of like basically has the Jason version of the LLM.
Jason:
[34:26] All right. You were thinking it.
Scot:
[34:31] Alexa, what time is it? Fascinating you asked. The nuclear standard for time began and then it like, it can't understand basic things. So it's gotten, it's better at answering questions Alexa currently can't ask, like, you know, certain facts and stuff but it it's not good at turning your lights on and off and those types of things so they're having a really hard time with it and they even got someone they weren't on the record but they i think this is in what's the one that's in seattle starts with a v i want to say vibe that's not it anyway they they got some amazon folks to say it's because because we have a management layer that's bloated and no one can make any decisions above us and they're constantly jerking around what they want the product to be.
Jason:
[35:12] It's not a two pizza team.
Scot:
[35:14] No, no. The two pizza team thing is not happening on the Alexa product. So that's pretty interesting that they seem. And then I think Jassy just basically nuked the leader and put a new leader in there. And they're doing all this return to office stuff, I think, as a way of kind of thinning out the employee base. I think next year, we're going to see them get pretty serious about reducing headcount in a lot of groups. So it's going to be a lot of companies have gone through that. And I don't think Amazon's done as much as they probably could based on how slow they're moving on some things.
Jason:
[35:48] Yeah, yeah. It is going to be interesting. I'm a little disappointed that the devices haven't gotten better faster. I think you and I both expected, like with all the amazing things that are happening in the outside ecosystem of LLMs that like, we can't figure out a way to make these devices better. But, you know, I think we talked about it in the last earnings call that there's this kind of more advanced concept of what AI can do called agent-based AI or agentic AI.
Jason:
[36:22] And there was a little bit of a hint in the earnings call when Andy was talking about this ecosystem. And he said, you know, all these LLMs out there are amazing. They're super good at processing information and summarizing information. But what we haven't seen yet and we think is going to be a huge part of the AI capability set is systems that can take action on that information. So actually do stuff for you, not just tell you stuff. And he optimistically said, and we think we're going to be pretty good at that. And so that, you know, my interpretation there is, hey, we at Amazon are going to lean way into agentic AI. They kind of have to because they've already been beaten in the first generation of LLM. So, you know, it kind of makes sense at this point to try to leapfrog. And of course, the guy that invented the concept of agentic AI, the super credible AI scientist, Andrew Ng, is on the board of Amazon now. So like none of this is particularly surprising.
Scot:
[37:29] Yeah. Yeah. So you, you wanted to bring up something that a lot of clients are talking to you about, which, which is interesting.
Jason:
[37:36] A lot of my clients have noticed this, this intersection of three very scary events that happened in the world of retail last week. So event number one is that Scott Wingo deleted the Google icon from his dock in his iPhone and replaced it with a perplexity icon.
Scot:
[37:57] True.
Jason:
[37:58] So, Scott is all in on perplexity as his primary search engine. Like, I'm less relevant, but I totally agree. Like, it's amazing. It's my daily driver for search.
Jason:
[38:11] And I think a lot of other early adopters have all now found it to be far, far more useful than traditional Google search. So shortly after Scott announced to the world that perplexity had beaten Google, OpenAI launched their first true kind of search competitor called ChatGBT Search. For me, the jury is still out on if ChatGBT Search is better or even at parity with perplexity. It's way better than any other OpenAI at doing search type things. I still think at the moment I prefer perplexity, but if you do 10 searches, you're going to prefer the answer for both of them for different things. And so this was the second big announcement was OpenAI getting directly into the search. And then in response to that, perplexity unveiled a new feature. They unveiled a public feature, which is shopping links in the search. So now, you know, you do a perplexity search for best Thunderbolt monitor to add to my new MacBook M4, and it comes back and gives you a bunch of answers. And at the bottom of that page, it gives you links to go to product detail pages of various retailers that are selling those monitors.
Jason:
[39:29] So that's interesting. But then in beta, they've added the ability to buy the product right in the perplexity interface without ever leaving. So what we would call native checkout. And I'm not in the beta, so I haven't gotten to do it hands-on yet. But I've been watching a lot of videos from people that are in the beta. And not all, but a disproportionate amount of those links are to buy now from Amazon. So they very clearly have some kind of partnership with Amazon. And, you know, if the world were to move to these kind of AI searches instead of Google, the two major things would happen. It would totally break SEO and keyword research and all the things that are a core part of marketing today. And number two, you know, it would be super important to win that buy box and that pull position in these new search engines, this new shopping surface that didn't exist for most of the world a month ago. So this is fascinating. I feel like this is just the first evolution of a battle we're going to see play out, at least for the next year.
Jason:
[40:37] Couple of years, but if you're a retailer and you weren't thinking about, you know, what happens when all this, this search traffic moves off of Google and onto these new, new surfaces, you know, it's, it's, it's time to start putting some contingency plans in place and at least think about how you're going to participate. And if you're a product or a content provider, you need to start thinking about how, you know, what your monetization strategy is and how you're going to optimize, you know, so your products show up at, you know, for their fair share of voice or better and all those sorts of things so it's.
Scot:
[41:08] Yeah it's called the aeo so the ceo of perplexity he's a fun interview i've watched a lot of his interviews now and he calls it an answer engine instead of searching in which is really appropriate that's why i like it because it just like gives you an answer and then so instead today google gives you like some things you have to figure out you go to 10 sites you open up 10 tabs three of them are garbage seven are okay then you kind of get to an answer after a lot of work this gives you the answer and then says well here's the supporting stuff if you want to dig in further so yeah so then to optimize it he calls it aeo and they're actually seeing people do this and they're they're really tweaking you know the sites to make sure that it's more llm friendly and and whatnot which is fascinating i wonder how the the pay on perplexity works because where are your payment credentials being stored is something i'd really want to understand so that's why i'm desperate to get.
Jason:
[42:01] In the beta because those are all.
Scot:
[42:02] That Like.
Jason:
[42:03] Normally when companies that aren't in commerce space first launch a native search, There's a bunch of deficiencies.
Scot:
[42:12] Yeah.
Jason:
[42:13] When will I get it? And what are the return policies? And how do I bundle multiple products? And what about, can I use my digital wallet? And all these things. There are now people that have figured it out. TikTok Shops has a pretty decent experience for native checkout. And so, yeah, I'm super interested to dive in all that with perplexity. I have not.
Scot:
[42:35] That's apples and oranges. They're like using it up at their lair, though. like you're actually selling on TikTok. Like it's the, it's like a marketplace. But this one, they're like sending the order down into Amazon or wherever. Like that's, they've got to be doing that. You know, how, is it Anthropic? Yeah, Claude can now take over your screen. I wonder if they're doing that. Like they're storing your credentials and then insert them, like crawling it in.
Jason:
[42:55] I doubt it. Could be, I kind of doubt it.
Scot:
[42:58] Super not PCI compliant.
Jason:
[42:59] No, and to be honest, the Anthropic demo is cool because it can do it at all. Like it doesn't do it well enough to solve a real world problem yet. And so I'm kind of assuming all the examples of checkouts I've seen are retailers that I know syndicate APIs, right? And so I actually think this isn't using TikTok Shops native checkout, but TikTok has a native checkout in TikTok ads where they already have a Amazon API integration. And it seems most likely that that's what they're leveraging is that they're, they've gone out and said, what retailers are, you know, enabling checkout in, in native ad formats? And we'll, we'll, you know, position ourselves as another ad platform for those, those checkout engines. Yeah, I'm with you. It's to me, the analogy I like to use is perplexity gives you a meal, whereas Google gives you a recipe.
Scot:
[44:00] Yep. True.
Jason:
[44:01] And I feel like, you know, the world has expanded from 10,000 SKUs to 800 million SKUs, like the recipes are now too complicated.
Scot:
[44:11] Yeah. Cool. I agree. It is. And there's even worse than that. There's bad ingredients, right? Like the, there's so much spam inside of Google now that it's just like really hard to find anything good a lot of times.
Jason:
[44:22] Yeah. I will admit like you've just hit my, my greatest fear about all of this. I feel like the perplexity search experience is so awesome right now, but they're not monetizing it. I mean, I, I pay 20 bucks a month, but you really don't need to. And I think back to every other cool technology that you and I have adopted, and they all started before monetization and they were all cool. And over time, the owner of that technology has turned up the dial more and more on monetization to the point where it really arose the experience, right? So, search, social, all these things were great pre-monetization. And, you know, now, you know, there's no organic content on social. There's no organic product listings that you're going to see on the first page of Amazon. Like, they've all been overwhelmed by the monetization. And so my fear is, as great as perplexity is right now, that when they really land on their, you know, that it's kind of the first hit of crack for free. And that when they eventually land on their monetization model, it'll be inferior and more noisy, as has happened in the past.
Scot:
[45:33] Yeah, we'll see.
Jason:
[45:34] Yeah. And one last side note on that. I'll tell you who wins here is Jeff Bezos because he has a lot of Amazon stock. He has perplexity stock and Amazon has a big investment in Anthropic. So it kind of feels like, you know, whoever wins here, Jeff Bezos is going to do okay.
Scot:
[45:53] Jeff would not cry to see Google taken down. A lot of people in tech would not be upset to keep Google taken down a peg.
Jason:
[46:00] I know you know this. Like many of our listeners may not know that Jeff made billions of dollars on his initial investment in Google. He was one of the angel investors in Google.
Scot:
[46:11] Yeah, he famously identified him as a threat early on and wouldn't let anyone at Amazon talk to them or share any kind of – they never installed the Pixel or anything. They wouldn't tell them what – they wouldn't buy ads on there for a long time.
Jason:
[46:25] Yeah, now they're the largest advertiser, $18 billion.
Scot:
[46:28] Yeah.
Jason:
[46:29] It's a crazy world, Scott. But that seems like a great recap for this week. Hopefully you found value in it. And if you did so, there is still room for more reviews on our iTunes page. So Apple has called us. They've said, hey, you're an enormously popular podcast. You have so many great reviews that we're expanding the capacity so that more listeners can jump on iTunes and leave us a five-star review.
Scot:
[46:55] Thanks for joining us, everyone. Hope you enjoyed the show.
Jason:
[46:58] And until next time, happy commercing.
