Groceries, Aggregators, and Agents
Just over two and a half decades after Walmart was founded, the company entered the grocery industry in 1988 and then in just thirteen years after getting into this business, Walmart became the largest grocery player in the US. Grocery leadership was also a major unlock for Walmart since not only a good chunk of discretionary spending goes to groceries for US household, it also drives frequency of visits to the store. Such frequency also creates familiarity and loyalty to a specific retailer since nobody likes to try grocery shopping different chains every week and figure out the maze of grocery aisles.
However, the nature of grocery shopping is going through a fundamental evolution ever since the pandemic has forced a lot of people to give online groceries a shot. This good deep dive on Amazon’s grocery ambition by “The Jason & Scott Show” (h/t Scuttleblurb) really drove this point home to me. Let me quote from the podcast (emphasis mine and slightly edited for clarity):
“I have been saying for a long time, just talking about digital commerce in general, that we used to go shopping and now because of digital, we always are shopping, right? And that was a kind of fun throwaway line. But think about how that impacts grocery in a very unique specific way over everything else, right?
When you go grocery shopping, you buy 60 to 100 items, right? And so traditionally in the U.S., grocery was a weekly shop, right? Because you can’t make it a monthly shop because you can’t predict exactly what you’re going to consume and there’s perishables and things that expire and milk only has a 21-day shelf life and all these things. So it turned out to be about weekly that you’d go and you’d do a big shop every week and if you screwed up and forgot something, you would do a fill-in shop between that weekly shop. When you always are shopping, if you suddenly can add one item to a cart and have it show up in three hours, in one hour, in 30 minutes, and you just fill in as you go, like it literally replaces this big shop with all these ad-hoc little shops.
And in our current environment, if you’re a busy time-starved family and you maybe don’t know how many times you’re going to sit down for dinner this week, it’s more convenient to be able to buy the items you need for today and tomorrow than it is to plan a week in advance. And if you have economic uncertainty in your life, there’s even more reason than you might want to shop in smaller bites so that you have less waste, you have less breakage, you have less out of pocket at a time, you can manage your cash flow better.
So for all of these reasons, the advent of digital commerce and the speed of service has meant that we’re fundamentally re-teaching people how to shop for groceries. And that big stock-up trip is largely going away for a lot of American families and it’s being replaced by these ad-hoc trips, which interestingly was more common originally in, for example, Europe, where you might have lived really close to a local market and you might have gone to that baker three times a week to get fresh bread or the butcher two times a week or whatever the case is. Yeah, so inside of that, we used to have kind of like your one grocery you would do your big stock-up at. And now industry research shows that the average consumer has three to four trips, but more than half consumers don’t really have a primary grocery.”
I understand this may be obvious to many, but somehow I didn’t fully appreciate it until listening to this podcast. I am actually a bit surprised that it didn’t occur to me because my wife herself is a prime example of this changing grocery behavior. We used to go to Target for our weekly grocery shopping, but ever since our son was born, we mostly moved our grocery shopping online. Since then, I have noticed my wife ordering from Instacart almost every other day. The entire weekly grocery shopping has essentially been replaced by a series of fill-in deliveries in our household.
This behavioral evolution also made me update a bit on DoorDash. There is one positive and one negative read for DoorDash compared to how I used to think about their bet on groceries. The positive read is DoorDash seems to be on the right side of this trend as they are already the market leader among the marketplaces in fill-in trips, so if fill-in trips are indeed the new dominant behavior among consumers, DoorDash will be able to ride that secular trend. On the other hand, the non-linear economics of the grocery basket makes the math less compelling for fill-in trips than the larger weekly baskets. But you cannot really fight the consumer behavior, so if companies such as DoorDash indeed dominate the fill-in trips, they may ask for better economics from the grocery chains. We are still somewhat in the early phase of grocery delivery and my guess is end state economics will be favorable to the demand aggregators (with one caveat I will discuss later).
One interesting implication for such rising fill-in trips is that while it has expanded TAM, the competitive intensity is also commensurately rising. From the same “The Jason & Scott Show” episode (emphasis mine and slightly edited for clarity):
“…the thing is that everyone’s TAM has greatly expanded. If you’re Walmart, you used to just sell the 40% of calories people consumed at home. Now you can sell 100% of the calories. But at the same time, you used to just be competing with three other grocers. Now you’re competing with all the fast casual restaurants and the QSRs. And so it’s become the wild west. Everything is fragmented. You have to win each food occasion as opposed to winning that aggregate food opportunity.
…So the company that exclusively focuses on groceries is only the third largest grocer in America, which goes to this aggregation story. That people want to get as much, are favoring the people that can solve multiple problems as opposed to only solve one problem.”
Indeed, aggregation story is exactly why marketplace business models can have compelling economics at scale. Even Walmart’s economic drivers are changing in front of our eyes and in several sell-side sessions over the last couple of weeks, Walmart management really wanted to hammer this point home. Dave Guggina, President and CEO of Walmart US, highlighted that Walmart’s e-commerce business would be profitable even excluding advertising:
“…we have these three magical businesses that are helping reshape the P&L within e-commerce. That’s our membership, which is incredibly important. That’s our Walmart Marketplace, which I’ve already mentioned, and that is our advertising business. All three of those saw double-digit comps this past quarter and have a lot of momentum.
We could not be more happy with the trajectory that we are seeing in profitability. But I would also mention two other things we have talked about, and that is our Marketplace and that is advertising, which is absolutely contributing to the profitability of e-commerce. However, even if you strip out advertising, our e-commerce business was profitable in Q2.”
That is quite notable given that in another sell-side session, Ryan Mayward, Senior Vice President and General Manager of Walmart Connect US, quantified advertising margins to be 70%+ (emphasis mine):
“We have a great high-margin advertising business, over 70% margins compared to 5-ish percent for the core retail business. And this profitability is incremental to Walmart. I think some of the retailers out there are growing their advertising businesses through sort of a pocket shifting, giving a break in one sort of trade investment area in exchange for investment in ads. That’s not something that we’re favorable on, which we don’t do that at all. And so the profit from the business is purely incremental to Walmart. And we see a lot of headroom to grow this business.”
Of course, the big question mark for such high margin advertising business is the rise of consumer agents. Interestingly, Guggina shared some compelling stats and use case around Sparky, which is Walmart’s own shopping agent for consumers. From Guggina at GS Global Consumer and Retail Conference (emphasis mine):
“We launched a new capability with visual shopping with Sparky, so you can now take a photo and Sparky will help you shop depending on what your mission is with that photo. When customers utilize visual shopping, we see the conversion rate jump by 57% versus text-based shopping.
One of my coworkers had this great example that he did recently and shared with me. He went out to his backyard, and I guess he doesn't do a great job taking care of his grass. He takes a photo of his lawn and says, "I need some help." Sparky, just with that context, was able to recommend seeds, lawn growth fertilizer, and help him solve that mission, build a basket of unique items that he may not even thought of to help solve that mission. As a result of these enhancements, customers are responding. We've seen just quarter-over-quarter weekly engagement with Sparky grow over 60%. When customers engage with Sparky, their average order value jumps 40%. That's because of what I just spoke to. Maybe you are having friends over to grill for the summer, and you just say, "Sparky, I've got eight folks coming over. I want to grill beef, I want potatoes, and I want other things. I need food for eight people." Sparky can build that meal out for you.”
I am very, very skeptical that Walmart’s Sparky will be able to do a better job than Muse (and Apple, OpenAI, and Google’s Muse-like agents). The more I use Muse, the more I realize broader horizontal-layer agents like Muse will have the best context of my query. As computer use gets much better and much faster than humans over time, it is very conceivable to me that Muse can give me the best suggestion from the entire corpus of alternatives out there, not just Walmart. ChatGPT was lot clunkier to use for these use cases which is why vertical agents still seemed compelling to me, but my skepticism around vertical agents has increased by an order-of-magnitude after using Muse over the last week. So, in that sense, it is good news for Walmart that their e-commerce is profitable even excluding advertising but the terminal margin question remains very much alive for Walmart and Amazon’s e-commerce businesses. Amazon will likely try to compensate for the lost margins through higher margins from their logistics and fulfillment operations. Walmart’s 3P business is still sub-scale compared to Amazon, but that may be a new hill to climb for Walmart in the coming decade.
Of course, the same logic could apply to DoorDash, but I still think their food delivery business is more insulated than the broader retail. Since you don’t always know beforehand what you want to order and actually need to browse the app to choose your meal, the browsing flow and the ad inventory that comes with it still seem largely intact to me for the core food delivery business. Grocery and retail orders are more exposed to agents, but there DoorDash is mostly delivering for retailers that have nothing resembling Amazon or Walmart's logistics and fulfillment operations. Even if an agent ends up deciding where the order goes, those retailers will likely still rely on DoorDash's network to get it to the door. And if agents do erode DoorDash's ability to generate compelling economics via ads, my guess is DoorDash would simply ask for a greater rake from the retailers, just like Amazon leaning on logistics and fulfillment to potentially make up for lost ad margins.
Subscribers get the daily journal and five+ years of Deep Dives, i.e. full-length analyses with financial models on 70 companies. The daily is just how I think out loud between the Deep Dives!
Current Portfolio:
Please note that these are NOT my recommendation to buy/sell these securities, but just disclosure from my end so that you can assess potential biases that I may have because of my own personal portfolio holdings. Always consider my write-up my personal investing journal and never forget my objectives, risk tolerance, and constraints may have no resemblance to yours.
My current portfolio is disclosed below: