Interview Transcript

Disclaimer: This interview is for informational purposes only and should not be relied upon as a basis for investment decisions. In Practise is an independent publisher and all opinions expressed by guests are solely their own opinions and do not reflect the opinion of In Practise.

A significant part of this call will relate to how Amazon views competitive or differentiated digitally native e-commerce solutions, particularly those that differentiate based on their upstream relationships with suppliers. Examples of this would be Temu or SHEIN. Are you familiar with these?

Yes. I worked in Amazon Retail for six years and now I mostly work with investors. I help them understand Shopify, Amazon, Fulfillment by Amazon, the growth of third-party platforms, selling on various platforms, the roles of Walmart e-commerce, and so on. I hope I can add value to this discussion. Let's see where it goes.

Let's dive right in. Based on your experience and time at Amazon, how did Amazon perceive competitive business models like Temu or SHEIN, where their relationship with upstream suppliers differs from Amazon's? How would you define those models? What makes SHEIN or Temu stand out, and how did Amazon view that?

I'll start with a somewhat generic answer, but it's a real one that sets the stage for Amazon. Amazon historically didn't focus much on their competitors, across logistics, fulfillment, and so on. This is quite different from my experience at Delta Airlines, where we knew everything about our competitors like United Airlines, American Airlines, and Southwest, including their policy changes. We would conduct focus groups with customers to understand why they chose one airline over another. This wasn't the case at Amazon.

Amazon's approach was more about customer studies. They focused on what customers wanted in terms of speed, cost, and product selection, rather than spending time on what Temu was doing and how to counteract it. They would look at consistent customer data and identify what the customer wants or what they're not getting. For instance, if customers weren't getting low-value, delayed products from China, Amazon would try to understand how customers wanted that.

This might sound generic because every company claims to be customer-obsessed. However, in my experience in transportation, logistics, operations, and retail at Amazon, there weren't many conversations among leadership teams about doing something because Walmart was doing it. That's not how Amazon typically operates.

Amazon's high-level goal was always product selection, availability, and cost. As long as these three factors were considered in the decision-making process, that's what got funded. If you could improve selection, speed, or cost, that's what got funded, rather than trying to combat a direct-to-consumer approach like TikTok's. That wasn't Amazon's culture.

There weren't many instances of retail leaders coming to work and saying, "TikTok Marketplace is doing this, so we need to launch this to combat it." This approach is different from what I've seen in other companies, where the mentality is, "Our competitor is doing X, so we need to do Y to combat it." That wasn't the culture at Amazon. I'll pause there.

Could you elaborate on your understanding of the Temu or SHEIN model? They're somewhat different.

I've placed around a dozen orders from them for myself. I did this because I wanted to observe which carriers they use operationally and how quickly the items arrive.

Generally speaking, the key difference between Temu and Amazon lies in consumer trust. With Temu, there's uncertainty about the quality of the products. Yes, they have a star system and rating systems, but Amazon's unique selling point is customer trust. This is where Amazon believes it differentiates itself from the Temu business model.

Speed is not a factor because the delivery times from Temu, which fulfills orders from China, are very different. Temu is generally cheaper than Amazon due to its business model, which will always be the case given tariffs, import duties, taxes, and US fulfillment. In the long run, Amazon cannot compete with Temu on a true cost margin. This is a well-understood fact, regardless of how much robotics are involved.

A significant part of Amazon's positioning is consumer trust. When you purchase something from Amazon, it's not counterfeit. It's not covered in lead paint, theoretically. Whether that's true or not is another matter. There are many instances where things slip through the cracks when you have hundreds of thousands of suppliers in China selling on But generally, the standard is high, and consumer trust is what I would consider the main differentiator.

Could you explain why it's well understood that, both now and in the long term, Amazon will not be able to compete with Temu on cost? What is the cost advantage of Temu?

There are two fundamental costs. This is evident across other non-Temu direct-to-consumer fulfillment from China. I've advised one company on this. Neither Temu nor SHEIN have been my clients, but I have had a client whose entire business model was China cross-border ecommerce. One cost advantage in the Amazon model, or any US retail model, is that duties and taxes are not paid for low import clearance. If you order something, I assume you know this, but let me know if this is a significant point. If you haven't studied it in depth, I apologize.

Are we discussing the de minimis exemption?

Yes, the section 321 clearance.

Please explain it as if you're speaking to a child.

When you import anything into the US, there's a duty or tax, depending on the classification of the goods. This is common worldwide. If you import something into Japan, for example, you pay an import tax. In the US, many goods are taxed at 20% to 25%.

If you have an ocean container of widgets coming into the US, you're going to pay 25% in duties or taxes to the US Government for importing them. However, if you ship one or six widgets, and their total value is less than $800, which is typically the case when you're shipping directly to a consumer, you're not paying that 20% or 25%. So, if you bring it in an ocean container, you pay a 25% tax. If you ship it to a consumer, you pay a 0% tax. This is one structural advantage.

Let's pause for a moment and consider Amazon's third-party marketplace, where a merchant in China might ship directly to a customer in America. Why doesn't Amazon benefit from the same de minimis exemption, assuming they qualify?

That aspect is not a significant part of Amazon's business model. Ebay, for example, used to offer a wide range of low-value goods that could be ordered from China and delivered to the US at no cost. However, when you consider the billions of units that Amazon sells, a very small percentage of Amazon's marketplace is fulfilled directly from China. This doesn't mean there aren't sellers from China, but they don't typically fulfill orders from there.

Amazon has become synonymous with the Prime membership, which guarantees two-day delivery. It's virtually impossible to fulfill orders from China within this timeframe. The vast majority of Amazon's sales, although not publicly disclosed, are to Prime members. Prime membership is well-known, especially in the US, and it promises fast, quick delivery.

People don't usually go to Amazon to order inexpensive items from China that will take 12 days to arrive. While this option is available on their platform, it doesn't constitute a large portion of their business. Amazon is known for fast, free shipping, particularly for Prime members. If Amazon decided to dedicate a large section of their website to goods fulfilled from China, it could potentially damage their brand in America. If a Prime member were to search for a backpack on and the first page of search results were items fulfilled from China that would take 12 days to arrive, it would be a stark departure from Amazon's core brand.

Amazon's core brand is built on trust and the promise of local availability. Customers expect to be able to order a backpack and have it delivered by 08:00 tonight. Right. So is it available? Yes. Is it available and used at a high rate?

I'm still not clear on something. What prevents Amazon from benefiting from the de minimis exemption? You mentioned that although there are Chinese sellers, the products are not fulfilled from China. Could you explain this in more detail? Perhaps that's the part I'm not fully understanding.

Amazon, for instance, handles a significant amount of global logistics for its sellers. If you're a seller, you can produce goods in China, move them to the US, and then place them in Amazon's system. Amazon will even pick up the goods from your factory.

If you're shipping goods via ocean freight, you're typically moving about 3000 widgets in an ocean container, which will cost more than $800. Almost 99% of goods in America are transported via ocean, and they don't qualify for the de minimis exemption. De minimis, which is like mail, applies to goods worth less than $800.

Walmart, for example, doesn't move anything into the country via de minimis. In theory, Walmart could package something for less than $800, but in practice, it's not feasible because they import more than $800 worth of goods per day.

If you have an ocean container full of widgets, it's always going to have more than $800 worth of goods. So, the de minimis exemption is mostly applicable to direct-to-consumer shipments from China.

If you're an importer and you order 25 different orders from Temu, 25 from SHEIN, and 25 from the eBay marketplace, and then order $700 worth of goods from a friend's shop in India, in theory, you're supposed to pay duties and taxes on all of that because you can only import $800 worth of goods per day.

However, in practice, the enforcement is on a per shipment basis. It's also important to note that most direct-to-consumer orders from Temu or SHEIN are usually worth less than $800. But if you're importing ocean containers of widgets, businesses have a higher compliance standard than individuals.

So, Temu doesn't handle the logistics at all. Amazon, on the other hand, manages the ocean containers with the freight to fill its warehouses so that the goods can be ready on demand. When you purchase a product from Temu, it's fulfilled by the seller. It may still come via ocean freight, but it can go through the de minimis exemption. Is that correct?

Typically, anything coming via ocean is in bulk, and bulk shipments generally don't qualify for the de minimis exemption.

So how does a product come from Temu? Is it not transported via ocean?

No, it's shipped by airplane because it's an individual package. The clearance is also individual. For instance, if you ordered a flashlight and a shirt from Temu and they were packaged together, the shipment would consist of two units inside one package going to one consignee or customer. That's how the clearance is done. The 321 clearance, or de minimis clearance, applies to that one shipment. It might contain two individual units, like a flashlight and a T-shirt in the same bag, but it's being imported to you as the consumer and it's under $800.

However, when you consider an ocean container, the law applies to the import value of the importer or the consignee for the day. Typically, you wouldn't have an ocean container with 800 individual customers. Parcels don't usually come over in ocean containers. They come in bulk pallets, which means it's 800 widgets, and once it gets to the US, it's available for purchase. Only after someone clicks 'buy' on does it then have a lower value per unit.

But the aggregate, most retail like Walmart, everything comes in bulk in ocean containers. So everything's over $800 as soon as Amazon and that's the model. That's where that 20% or 25% advantage comes from.

I was just going to mention the other structural piece, which is the cost of labor. As much as Amazon uses robotics in its fulfillment centers, the fulfillment cost per unit will always be structurally higher when you pay an American. Amazon pays health benefits and starts wages at around $17 an hour plus health benefits on day one. This is in contrast to what labor in South China would cost for packing a shipment.

These are the two structural differences we'll discuss. Of course, air freight offsets that somewhat. But we've been talking about 321 the whole time, and the original thought was the cost of labor and duties and taxes. These are the two main points to consider.

I've spoken to a few investors who ask at what point does Temu become so large that it has to set up US fulfillment? What's interesting about this question is that if Temu sets up US fulfillment, its value proposition is eliminated. It would then have US labor costs and would pay in bulk for imports, not to individual consumers. Then it would have a hard time competing against Amazon.

Currently, Temu can compete against Amazon due to its two structural advantages. But when investors ask me, a question I often get, about the size Temu needs to become to set up its US fulfillment network, I think that breaks the model. That's just one advisor's opinion, because then you're challenging the business model. It's not a question of scale at that point.

We've discussed the fulfillment piece and the cost of labor, as well as the potential cost of having these fulfillment centers themselves, developing them, and leasing them. We've also talked about the de minimis exemption. When we consider cross-border shipping, as you mentioned, some of the benefit from the de minimis exemption is offset by the fact that air freight is much more expensive than ocean freight. We can delve deeper into that later.

But I also wanted to take a step back and discuss the actual procurement of the goods, the cost of goods sold. Do you know if Temu has any cost advantages versus Amazon in this area?

I'd like to comment on my experience with global supply chain procurement and sourcing, even during my time at American Eagle. Companies like Amazon, Apple, Nike, and others set a high bar for sourcing, especially in terms of compliance. In my work in global logistics, I've noticed that large companies like Nike are very sensitive to issues such as child labor.

If you buy a pair of sneakers from an unknown brand, there might not be any standards in place for the factory settings where those sneakers were produced. However, we can probably agree that a pair of Nikes was not assembled by child labor. This is also true for Amazon's first-party marketplace. If you buy an Amazon Basics computer monitor or office chair, it's the same as buying from Nike. There will always be companies that can produce shoes cheaper than Nike or Apple, but they may have less compliance. Take Foxconn, for example, which Apple uses.

I'm more interested in the comparison between the price that a third-party merchant on Amazon could get for a product from a manufacturer versus the price that Temu could get for the same product from a factory. How would those two compare? Does Temu have a structural cost advantage over third-party merchants, and if so, why?

In the scenario you're describing, where it's third-party ownership from factory to production to sale on either the Temu or Amazon website, I don't believe there's a structural advantage between third parties and Temu. They're using the same factories. For example, in China, a large portion of battery production comes from the same factories. They produce batteries for Energizer, Railvac, and even Costco's brand.

The biggest advantage usually lies in the size of production, as that's the main driver of production costs. I don't have enough information about the volumes of Temu or the merchants on Temu to determine if there's a structural advantage. The structural advantage would come from the size of production. If Temu is moving two million units of batteries and an Amazon third-party seller is moving 10,000, then that's where the structural advantage would come in. Most origin manufacturing is volume-based.

Could you provide some insight into the landed cost of a parcel from Temu that is carried by air freight when it arrives at its destination?

I can discuss the individual costs of the components.

Considering the labor rates in China, the fulfillment cost per unit for Temu is likely to be about half of that of a traditional US retailer engaged in ecommerce. This is not in comparison to Amazon, which has made significant strides in robotics, but rather to traditional retailers like Target or Dick's Sporting Goods.

The air freight rates have changed significantly in the last three years. The cost per kilogram, on an all-in basis, is roughly $3.50 from airport to airport. If you add another $0.50 for warehouse-to-warehouse transportation, it comes to about $4 per kilogram. This rate fluctuates during peak season.

For a one-pound shipment, which is about half a kilogram, the cost to transport it across the ocean via air freight is between $2 and $2.50. This is under the assumption that Temu only uses this model and does not use UPS or FedEx.

Temu does not produce something in a fulfillment center in Hong Kong and then hand it over to FedEx, as that would make it a $40 package to deliver to the consumer. Instead, Temu buys bulk air freight on airlines like Cargo Luxe or China Eastern. They fulfill the order, move it to a warehouse near the Shanghai airport, and then to the airline's warehouse for shipping.

Once the shipment clears customs, the cost per package for clearance is around 5 to 10 cents. This is relatively low because it's essentially a filing with customs that includes the consignee, the contents of the goods, and the amount paid for them. It's mostly a paperwork exercise.

If they're benefiting from de minimis, does that change the calculation or not?

When we talk about de minimis, we're looking at costs like five cents or ten cents per package. This is assuming it's de minimis. For a formal clearance, you're not only paying duties and taxes, but the clearance itself might cost you $125 because you're paying a customs broker to manage that entry.

When it's de minimis, we're talking about negligible costs per shipment. It wouldn't surprise me if Temu has already purchased a customs broker in the United States, and the cost is simply their fixed cost of running that customs broker.

The challenge with the Temu model is the use of a variety of last-mile carriers, which could cost anywhere from $5 to $8 per shipment. You have your cost for fulfillment, the cost to move it to the airport, and the cost to move it to the US airport. Then, you're injecting it into a company like OnTrac or perhaps even the US Post Office, depending on the shipment characteristics. UPS, FedEx, there are many reasons why you'd use a variety of different carriers, but that part is not cheap. They'll use certain carriers for certain package sizes, certain carriers for certain geographic areas. If it's less than a pound, typically the United States Post Office is the most cost-effective carrier to deliver a shipment.

Once the shipment arrives, you would be paying a third-party carrier for both the middle mile and the last mile. They'll need to put it in some sort of sortation center and then send it out to a delivery center to get it to the front doorstep.

Typically, Temu or SHEIN only come into a few airports. The only middle mile they're really doing is launching from, let's say, the Dallas airport. Temu or SHEIN may run a Dallas to Denver truck themselves and then inject it into someone in Denver. But they're not doing a lot of middle mile, as far as I understand.

To some extent, OnTrac, LaserShip, and the regional carriers also have somewhat of a network. There's a diminishing return on how much you do versus letting UPS or the Post Office handle the middle mile. Given Temu's size, I wouldn't be surprised if they're only coming into maybe six different airports in the US every day.

They might land in Atlanta and run their own truck down to Orlando, injecting into OnTrac or LaserShip's last mile delivery network there. But if it's going to Miami, it's not a Temu truck that brought it there. Getting at scale, getting as close to your final mile carrier allows you to lower costs.

With UPS or any carrier, you always have a rate card. The concept is the longer they carry it, the more money it costs. But if you have a lot of volume going a decent distance, you can run a truck more cheaply than paying UPS for those packages. It's a math problem on why people do their own middle mile and hand it off further downstream to a final mile carrier.

Let's return to the concept of Temu. Once the package has arrived here, they need to get it to the customer. This could be referred to as the middle or last mile, and it's a bit of both. You mentioned that this costs in the high single digits.

Yes, the cost is around $5 to $8. This depends on the distance the package has to travel. For example, if a package lands in Los Angeles, is picked up and goes to a warehouse in Los Angeles, and is then delivered within the Los Angeles metropolitan area, it can probably be done for $5 per package. However, if the package is going to Oregon, it won't be delivered for $5. A package going to Portland is likely to cost $7. The same carrier might handle both deliveries, but the cost varies.

In package delivery, there are often many surcharges tied to the delivery area. If you live in Los Angeles or within 30 miles of Los Angeles, you pay the base rate. However, if you live further away, there might be a surcharge. I'm sorry, I don't know where you're based. I could have used your location as an example.

I'm located in East Bay, California.

The whole Bay Area has the same delivery price. But if you go to a place like Tracy, California, which is about 40 miles east of you, there might be a surcharge of $1.50 per package. This needs to be taken into account. So, while we'd like to say that everything is landing in Los Angeles and being delivered to customers there for $5 a package, some amount is going to places like Tracy, California, in line with the population size of Tracy. These would all be $7 packages. So when I say $5 to $8, the average might be six and a quarter for the last mile.

Now, let's do the same exercise for Amazon, considering their use of ocean freight and their scale, without de minimis exemptions. What would be the cost? We can then move on to the second part.

Regardless of whether you're Amazon, Walmart, or Target, the process is similar. However, Amazon has a unique step in its supply chain. When goods arrive at ports such as Los Angeles, Long Beach, or Oakland, they are typically moved into a transworld or cross-dock facility near the port, then to distribution centers. For Walmart or Target, the goods would then go to a store, but for Amazon, they go to fulfillment centers, which are equivalent to Walmart's distribution centers. This is the additional step before the goods reach the consumer, which is a key difference between physical and online retail.

When comparing cost structures between Walmart and Amazon, you have to consider final mile delivery versus in-store pickup and the cost of store real estate. However, the cost of ocean freight has also changed significantly. During the peak of the pandemic, the cost of a 40-foot ocean container increased from around $1,500 to over $10,000. This had a significant impact on landed costs, particularly for large goods.

For example, if you're shipping large screen TVs and you're now paying $13,000 for the ocean container instead of $1,300, you're talking about an additional $30 per TV to get it across the ocean. This drove inflation and increased the cost of every TV at Target by at least $30. So, landed costs have changed a lot, but for this discussion, let's use 2019 and current data and exclude the pandemic.

There are other costs to consider as well. For instance, someone has to load the ocean container at a facility in China, which means turning on the lights in that facility. Then, the goods have to be picked up from the factory. However, when you're moving goods in bulk, the unit cost can be as low as six to eight cents per unit. This is because the goods are moved on pallets by forklifts, so individual packages aren't being handled.

The landed cost for ocean freight, including the facilities on both sides of the ocean, is highly dependent on how much physically fits inside an ocean container. If you're shipping cell phone cables, you could fit 10,000 cables in a container that costs $1,300, which equates to 12 cents per unit for ocean freight. In contrast, for larger items like an Instapot, you might be looking at $3 per unit.

To simplify, let's average it all out and view it as a typical ocean shipment, which includes a mix of large and small items.

Yes, your costs will definitely be less than a dollar for each operation, from filling the container to distributing it in the US.

Does this include customs?

It depends on the value of the goods. Customs are typically considered in landed costs. Your question about the landed cost is correct. However, in transportation logistics, we often exclude the tariff because small goods can be valuable. For example, watches are small but valuable. So, you would have to make an assumption about your average landed cost. Surprisingly, I don't know what the average value of an import on Amazon is. You could probably determine this from public data or by dividing Amazon's total sales by the number of units they sell. If the average unit is $8, then it's 25% of $8, but that would be the retail price. You would have to assume some sort of margin on the good because you're paying the duty or tariff on the cost of the import, not on the retail price.

To clarify, I'm trying to compare the landed cost of an item from Temu versus a comparable item landed via Amazon. With that in mind, perhaps we can answer.

I would suggest you go back and try to find out how Amazon talks about delivering five billion of its own packages. This has recently become popular in the Walmart marketplace. So, take Amazon's revenue, divide it by the number of packages they deliver. Let's say Amazon delivers 80% of its own packages. If Amazon is going to deliver five billion packages this year, let's say they actually deliver seven billion packages. Divide your total top-line revenue by seven billion and you probably get $6 per unit or something to that effect. Then apply a 20% or 25% duty or tax on that. If it's a $250,000,000,000 company and they deliver seven billion packages, I'm not sure what that would be. No, that's not correct either. That wouldn't be the right math.

What we want to do in that case is to think about the landed cost. It would probably be more relevant to look at it from a cost of goods sold perspective rather than a revenue perspective. That's tough because you have a sense of what their first-party cost of goods sold could be. But then there's a lot of other stuff mixed in there, like the ads business. Actually, no, the cost of goods sold will be the same. It's the cost of goods sold divided by the number of packages.

Yes, it's the cost of goods sold divided by the number of packages. I shouldn't even guess, but on average, one of the challenges is that Temu tends to sell lower value goods than Amazon. I don't think we typically turn to Temu to order our new 55-inch TV or to buy luxury goods or products. Anything that's $200 plus is probably not coming from Temu. Temu tends to sell low value, low cost of production goods. Let me pause because I'm probably getting too far outside of my threshold of knowledge. I'm just guessing at this point.

Let me share my thoughts. Correct me if I'm wrong, but let me get through this. The price that I, as a customer, will pay on Temu for the same good as I would pay on Amazon depends on the cost of that good, the cost of manufacturing it, the cost of shipping it to this country, and the cost of last-mile delivery. We can ignore the last-mile cost for now and focus on the cost of the good and the cost of cross-border fulfillment. The retail price is determined by how low you can get the cost of the good and the fulfillment to this country. I'm trying to understand the cost of getting the same good per package, per good here to this country, given Temu's current business model and Amazon's current business model.


Let me clarify one last thing. If Amazon uses third-party fulfillment, the cost of getting the good fulfilled cross-border to this country is covered by the merchant. Is that correct?

Yes, the merchant pays and remains the importer of record. Amazon does not take possession until someone buys it at the checkout.

So, it's different in a first-party marketplace where Amazon would assume the cost of bringing it here.

That's correct.

Amazon would bear the cost of importing it.

Yes, in a first-party scenario, similar to Target or Walmart, they decide where they take possession with their suppliers. Typically, companies like Home Depot, Amazon, Target, and Walmart find it cheapest to take possession at the factory. If you're buying in bulk, the best bet is to buy it at the factory in China and manage the supply chain yourself. However, you can also sell any good, like a Stanley hammer, and have Stanley deliver it to your fulfillment center in the US. Stanley would import it, place it in their distribution center, and you would only buy it from Stanley when needed.

The landed cost is what determines the retail price, regardless of whether it's covered by the merchant or Amazon. So, the question is, can Temu achieve a lower landed cost than Amazon in the long term? Without speculating about the future, let's consider the current situation. We previously discussed a landed cost of about $2 for a widget via Temu, thanks to the de minimis exemption. How does this compare to Amazon's landed cost, considering ocean freight and the absence of a de minimis exemption?

The key factor in answering that question is Temu's scale. Temu's business model relies on airplanes. The question is, at what point does Temu need its own fleet to fly goods from Shanghai or Hong Kong to cities like Seattle, San Francisco, Los Angeles, and Minneapolis, in order to minimize the final mile delivery cost?

If Temu's network is fully utilized and they have a large volume of goods, they could potentially compete. However, Amazon's strategy for reducing costs is quite different. Amazon is focusing on moving goods in bulk closer to the consumer, thereby reducing last mile delivery costs.

The key cost factor we're discussing here isn't just the landed cost, but the total cost from the factory to the consumer's doorstep. Amazon's strategy is to get goods in bulk closer to the consumer, which reduces costs significantly. Moving a pallet of goods on a per-unit basis is much cheaper than moving individual units.

Amazon's long-term plan involves moving bulk goods, such as paper towels, cutlery, or widgets, closer to the consumer, thereby minimizing landed costs. The ultimate goal is to deliver goods to the consumer for around $3 per shipment. This is an advantage that Temu will never have. Temu's costs will always be in the range of $5 to $8 per shipment.

So, you're saying that because Temu won't be building distribution, fulfillment, and delivery centers in America, they will always have to outsource the middle and last mile delivery to a third party. Therefore, they won't be able to achieve the same economies of scale as Amazon. Is that correct?

It's impractical to land a jet daily in Miami, Orlando, Atlanta, Charlotte, and so on. If we're considering billions of units, I believe the government would question Temu's environmental impact. For instance, flying 300 airplanes daily from China. Considering Temu's growth rate, what would it look like in 12 years? I doubt any government would be thrilled about 300 daily flights with carbon emissions from China. I'm not suggesting there's a regulatory issue, but public sentiment towards the environmental impact of these flights is worth considering. However, this is beyond my area of expertise. At some point, this approach may lose its practicality. While we often discuss landed cost, the last mile is the most significant chunk of the cost in this entire transaction. Amazon is essentially trying to place more goods next to large population centers to reduce their last mile cost from $5 to $4 per parcel.

My focus on landed costs is based on the assumption that, even in the long term, Temu may not be able to compete with Amazon on middle and last mile. If this is a fair assumption, then the only area where they can gain an advantage is landed costs. I'm trying to understand whether they currently have a landed cost per unit advantage over Amazon. We're comparing two different models here, bulk and de minimis package.

In terms of labor and de minimis, Temu currently has about a $0.50 advantage per unit in fulfillment cost. This is where I become cautious about landed cost because I'm unsure of the average value of a Temu order. Let's assume it's the same, say a $6 unit at 20%, which would be $1.20. So, if there's a 50 cent advantage on fulfillment cost, and if the unit cost is $6 or $8, the value of that 20% changes quite quickly.

Could you elaborate on where this benefit comes from?

The $0.50 advantage I mentioned isn't necessarily related to landed cost. When I consider landed costs, I include Temu's fulfillment cost per unit, which is fulfilled in China. This $0.50 advantage is in comparison to a US ecommerce retailer's fulfillment cost per unit. Even though landed costs typically include different elements, Temu has to factor in fulfillment costs, unlike Amazon. When discussing Amazon's landed costs, we usually refer to the cost up to the point of injection into a fulfillment center, not the cost of fulfillment. However, in my opinion, these costs need to be considered together.

So, you're saying that Temu's model involves fulfillment in China and air freight with de minimis, whereas Amazon's model involves ocean freight and fulfillment in the US.

When we talk about landing costs, we usually refer to the cost of getting the product into the country. At Amazon, landing cost typically refers to the cost when the product arrives at the doorstep of the Amazon fulfillment center. However, in the Temu model, the product has already been picked, packed, and shipped. Therefore, it would only be fair to charge Amazon for that cost if we are making a true comparison.

To clarify, in Temu's model, the product is picked, packed, and put in a package. In Amazon's model, the product is sent, and then we also need to factor in the fact that it's not just at the doorstep of the center, but it also needs to be picked and packaged within the center.

Yes, and it needs to be available to ship. However, if you ask an Amazon retail leader about their landing cost, they will give you the cost to put it inside their Amazon fulfillment center. They don't include the cost of picking and making it available to ship.

That's an excellent point. So, we're essentially talking about everything except the last mile. Let's compare those two for Temu.

The calculations will vary based on the value and size of the goods. The fulfillment cost per unit is cheaper. It might be easier if we pull up an Excel spreadsheet and run through a few different size and value products. It all comes down to your assumption on the average value of goods. For instance, for an $8 item, you're talking about $1.60 per unit in tariff costs. If you're talking about a more valuable item, like a $30 watch, saving that 20% becomes a different calculation.

So, it's always 20%?

It varies based on the commodity. Clothing is taxed at a different rate than electronics, but roughly 20% is a good round number to use without getting into product specifics.

I'm happy to pull up Excel. This could be very illustrative. It may make sense that for certain goods, Temu and SHEIN will always have an advantage, and for other goods, they will not. It would be helpful to know the extent in terms of vertical market share.

Let's discuss the fulfillment cost per unit. This is a challenging area, particularly when comparing different products. For example, the fulfillment cost per unit for shipping cell phone cables is significantly different from that of televisions.

When comparing these two models, it's difficult to generalize because Amazon and Temu operate in different spaces. Temu primarily deals with low-value, small goods, which generate a lot of volume, but they don't handle items like office chairs or TVs. It simply doesn't make sense to ship these larger items via air freight due to the high costs.

Air freight is generally expensive. Consider an airplane flying from China to New York. The round trip operation cost, including the salaries of the pilots and fuel, can reach up to $475,000.

If we consider the cubic foot of a TV, for example, the math simply doesn't work for shipping large items on airplanes. However, for smaller items like watches, cell phone cables, and wallets, which weigh on average half a kilogram, the math can work.

I would be surprised to find items like air fryers on Temu's website. The shipping logistics for such items are not based solely on weight, but also on volumetric weight, which considers the space the item takes up on the airplane.

Ecommerce tends to deal with bulky items. Even if an air fryer only weighs two pounds, the box it comes in could be dimensionally equivalent to a 12-pound box, which is how shipping charges are calculated.

If you're selling a $50 air fryer and the air freight is $3.50 a kilo, but dimensionally, it calculates out to be a six-kilogram package, you've just spent $26 on air freight. This model doesn't work for air freight.

There's a limit to the size of the product that can be sold. I suspect that Temu doesn't sell large items. If you can buy it at Costco, Temu probably doesn't sell it.

Let's try to understand the cost of goods sold for Temu versus Amazon, excluding the last mile. This would involve comparing ocean freight versus air freight with de minimis. Can we walk through a mathematical example, covering every aspect from fulfillment to cross-border and the de minimis versus non-de minimis?

When I mention having an advantage, consider a $15 T-shirt that weighs around six ounces. There's going to be a structural advantage for Temu. However, for an air fryer, it probably won't have a structural advantage. So let's use whatever size we think is appropriate. I don't want to guess the average sizes, but maybe you have a better idea of what you want to use for your model.

Let's do one of each. One where Temu does have an advantage today and one where they may or may not. We can try to figure it out together.

If we start with the cost of fulfillment at an Amazon fulfillment center, let's call it a large ecommerce fulfillment center in the US, you're probably looking at eighty-five cents per unit. In a Temu fulfillment center in South China, I have to be careful because a lot of ecommerce fulfillment actually happens in Hong Kong Island. Hong Kong island is very expensive, so many South China-based ecommerce businesses fulfill in Hong Kong. There's a difference in fulfillment costs between Hong Kong and South China. There are a lot of nuances to consider.

All models are wrong, some are useful. I understand all the nuances. Let's just try to be rough.

Think of the fulfillment center cost in South China as half of whatever you plugged in for the US. So, from eighty-five cents, it becomes forty-two cents.

Forty-two cents per unit for Temu. Okay.

Let's do the T-shirt example. It's a half a pound, so it's eight ounces. If it's a $15 T-shirt, the air freight cost for Temu for a T-shirt is going to be about one fourth of that $3.50. This is because it's a half a pound, which is a quarter of a kilogram. So, the air freight for a T-shirt is likely eighty cents to the US. If you can do some sort of customs clearance in the US, in the Temu model, so section 321 clearance, and if it's something that's less than a pound, they're going to use the US Post Office. For last mile, if it's less than a pound, the best value is USPS. If you're shipping something, or Temu is shipping something, or Amazon is shipping something, if it's less than a pound, the post office is the way to go 100% of the time. You're looking at about $4 for that parcel.

Amazon fulfillment is $0.85. How much is cross-border ocean freight?

Cross-border ocean freight is probably only going to be, in this case, probably $0.40 in an ocean container.

Then what's the clearance?

Assume a higher 20% of that $15.

So that's $3.

Indeed, Amazon's last mile cost in this case is significant. For example, for a T-shirt, Amazon is likely to use the same post office rate.

Suppose they manage it on their own.

Yes, they could potentially have a slight cost advantage on their own. However, even Amazon, with their large vans driving around, can't always match the cost-effectiveness of the post office.

What else do we need to consider?

We should consider the post office rates.

Do we need to factor in Amazon's cost of fulfillment in the US, such as pick packing?

That was the $0.85 we included in the initial calculation. I can't see your spreadsheet, but that $0.85 would be the US cost. This is applicable for an apple, for instance.

Does that include the cost of getting it to the delivery station?

No, that would be a middle mile cost. You should probably add another 30 cents for that.

So, Amazon has an additional 30 cents plus a few dollars to get it to the actual doorstep.

Yes, that's correct.

If we do the math, we can estimate that it costs Temu about $5.50 to deliver this T-shirt to the customer. For Amazon, using the same last mile cost, it would cost them almost $9. Now, let's consider a more challenging item. For example, something that Americans love now, air fryers.

Yes, in the Temu example, the cost would increase due to the direct correlation between size and cost. However, the fulfillment cost would still be around $0.50. Let's say this air fryer is worth $45. What do you think an air fryer is worth?

Let's say it's $50. Temu seems to have them priced from $50 for mini versions to $89-$100 for larger ones. Let me check Amazon. Amazon has air fryers for the same price. Let's estimate it at $100.

In this case, whether you're Temu or Amazon, your cost is...

Wait a moment, an air fryer weighs six kilos.

It's six kilos, but the box that the air fryer fits in might not be six kilos. Air fryers might be dense, I'm not sure. Six kilos seems right, but typically in air cargo, for air freight, there's usually a 1.2 multiplier. So if it weighs 6 kg, you're going to be paying to move 8 kg because of the box. You're paying to move an eight-kilogram box on air at $3.50 a kilo.

So it's $3.50 a kilo, and it's eight kilos. So, $28 for the air freight?

Yes, for the air freight. And the clearance is still going to be 10 cents or whatever we put in there. It's very cheap to clear because it's still less than $800, so it doesn't really matter.

And then we have the last mile cost.

For the last mile delivery of an eight-kilogram box, you're likely paying UPS around $13 to $14.

What about Amazon's fulfillment costs?

The fulfillment cost for Amazon, let's estimate it at $1. The real savings come from ocean freight. For ocean freight, you're probably still looking at around $1 per unit. To illustrate, a 40-foot container that's 10 feet tall will cost you $1,300. You can likely fit around 1,300 air fryers in that container. So, it's essentially $1 per unit.

Is that for one-way or round-trip transportation?

That's for one-way. Ocean freight is typically calculated one-way. If Temu continues to grow, they can't just buy one-way freight. They would need to have their own airplane bringing goods to the US, and that airplane would have to return. With ocean freight, you generally pay for one-way, and the containers find their way back. So, you're looking at a dollar per unit for ocean transport. Now, you're only paying 20% of the cost. If a product retails for $100, the cost might be $70. So, it's 20% of $70, which is $14 for duties and taxes.

For the last mile delivery, Amazon uses its own network. Let's assume it costs $5 to do that.

So, in this scenario, Amazon's total cost is $21, while for Temu it's $42.

That's why I believe Temu will always excel at delivering cheap, low-value items from China. They can compete with small items, but anything larger becomes a challenge. I'm not an expert in manufacturing, but if you consider a $100 product that costs $40 to deliver to the consumer, I'd have to believe that returns pose a significant challenge for Temu. In ecommerce, it's not uncommon for 10% of goods to be returned. So, what does Temu do with 10% of goods that are returned? They can't send them back to China or the factory. This becomes a problem. When dealing with low-value items, it's fairly easy to just tell the customer to dispose of the item if it's not right. But if you're shipping a $100 air fryer and the customer decides to return it, that presents a problem. This is a challenge for both Amazon and Temu, but I'd say it's a more difficult one for Temu.

Is it similar with returns, where increased volume reduces per unit cost? Or do returns become more expensive and problematic with scale?

Returns don't necessarily worsen with scale, but they are typically one-off. They go from the consumer to a return center. It's not uncommon to have a return center in Ohio and another in Las Vegas. The aim is to minimize the distance of returns because you have to pay UPS or FedEx, or the post office, which we know charge more.

The last thing you want is to deliver an air fryer to someone in California, only for it to be returned to Ohio. You're usually covering the return cost by providing a shipping label. If you're paying UPS $26 to get that air fryer from California back to Ohio and this happens with 10% of your shipments, the costs become significant. This has to be factored into the calculation.

Let's conclude with the possibility of another call focusing on the concept of speed and regionalization within America. This is a significant change that Amazon has implemented. We want to explore whether faster delivery times, such as next day or same day, directly correlate with increased costs. Or, if there's enough volume or due to optimization and regionalization, does speed not necessarily result in much of an incremental cost? Would you be comfortable discussing this topic in depth?

I've extensively discussed this topic, spending no less than 30 hours with 20 different investors. The short version is that speeding up the network initially costs a lot of money. However, over time, optimization occurs. When you move from two-day to one-day delivery, it involves multibillions of dollars. Then, as your network evolves, you can gradually reduce that cost. Essentially, their one-day network now is what their two-day network was four years ago, in terms of delivery cost.

Is that because the volume increased, and you have to assume that demand will continue to grow? Or is it because you're actively changing and optimizing the cost structure for some reason?

The strategy involves both an increase in volume and a regionalization approach, which involves building more fulfillment centers closer to the consumer. This means goods have to travel less distance. When you move things over a shorter distance, it can be done more cheaply without increasing speed. For instance, if you have to move a package from Ohio to a customer in Miami, it takes two days unless you want to fly it. However, if the goods are already in Orlando and need to go to a customer in Miami, that reduces your speed because your speed is being bought by placement, not by transportation speed. There's a cost of speeding up the transportation. But what Amazon has been able to do is not just speed up transportation. As they regionalize their inventory and improve inventory placement, they effectively transport fewer goods over longer distances, which means they don't have to trigger faster speeds.

So, would it be accurate to say that the more you can shift the middle and first mile and the less burden you place on the last mile, the better it is overall for cost?

Yes, on a macro level, it's cheaper to deliver something if you can place it within 30 miles of a customer in Miami who is going to order it. It doesn't matter if you deliver it in two days, one day, or the same day. You're starting to get the same benefit as long as that truck is full going to that customer's neighborhood. The key point is where the package starts its journey. If the package starts its journey and doesn't need a first mile or middle mile and all it needs is a last mile, you can speed it up without incremental cost. You get cheaper deliveries through scale with incremental deliveries. So, volume has helped them achieve this, but placing things closer has also increased their volume.

Regarding Amazon FBA unit economics, which includes the merchant fee, advertising revenue, and the cost that the merchant covers for warehousing, and then there's the last mile delivery cost. Amazon has been clear that they've focused on the profitability of the actual products over time. When they do that calculation, do they include the revenue from the Prime subscriptions or do they exclude that? Is that part of what they're contributing to get to the profitability?

I'm not sure how they report on it. I know they consider it internally. I've never looked into where the Prime subscription revenue is factored into their profitability.