Wayfair: Demand Forecasting

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.

I'm specifically interested in the importance of demand forecasts for suppliers' success through CastleGate. I understand that this was a significant focus of yours while you were at Wayfair. To start, I'd love to hear about your background before Wayfair, your main roles and projects there, and what you've been doing since you left.

I'd be happy to share. Over the last decade, I've been involved in supply chain operations. My roles have typically involved running operations, strategizing on how supply chains should be designed and evolved, and identifying the technology requirements needed to bring those designs to life. My experiences at Amazon, Wayfair, and now Walmart have largely been at the intersection of strategy, data science, technology, and supply chain operations.

At Amazon, my focus was primarily on e-commerce operations. I managed large fulfillment center operations, introduced robotics, and ran automated fulfillment operations. I then transitioned to ultra-fast fulfillment, which involved shifting from the typical two-day shipping to one or two-hour shipping. This was when Amazon was piloting a program called Prime Now, which offered immediate delivery on select items in specific markets. I launched this program in several US markets and then internationally in Europe. After moving to the UK to oversee these operations for about a year and a half, I returned to the US.

Back in the US, my focus shifted to last mile delivery optimization at Amazon. This was before Amazon had its own last mile delivery service. Now, you see many of those grey vans on the streets, but back in 2015, most of the last mile delivery operations were outsourced, and Amazon didn't have much control over the quality and timing.

As Amazon started to insource many of the last-mile delivery optimizations, my role involved integrating data science to optimize routes. Additionally, I was responsible for identifying the technological needs to ensure we had the capabilities to run these routes, thereby achieving cost savings by insourcing the operation. After five years at Amazon, I moved to a digital freight brokerage in New York. There, I was tasked with creating better matching and pricing optimization algorithms. Once again, my role involved using data science and technology to make transportation operations more efficient for shippers.

During this period, market volatility was quite high, creating ample room for spot freight. This was a time when different production seasons triggered various supply-demand crunches in the transportation industry. I moved there to learn more about this. About a year into this startup, Wayfair reached out. They were looking to transform their supply chain with a focus on artificial intelligence. They wanted someone with supply chain knowledge but not necessarily with demand forecasting experience. This is because demand forecasting is one of the more advanced use cases of artificial intelligence deployment in the supply chain.

Wayfair hired me to establish an internal capability for demand forecasting. Before my arrival, Wayfair had been using third-party technology but had realized that demand forecasting could be a source of competitive advantage and wanted to insource it.

My role at Wayfair was to determine what this capability should look like, how we should build it, and what was needed to build it. Then, I was to go ahead and build it. I was at Wayfair for about four years, during the pandemic, where there was a significant focus on demand forecasting and inventory optimization. In short, I spent four years at Wayfair, establishing an internal demand forecasting engine and expanding it to various other aspects of supply and demand. This included thinking about variability and deciding what should be in CastleGate versus not. I also focused on the strategic value proposition of CastleGate analytics.

After four years, the company was not performing well, and investments in my area were expected to remain static. I found this unexciting for my career growth, so I started looking elsewhere. That's when Walmart approached me. They were trying to operate at the intersection of supply chain operations, strategy, data science, and technology, similar to my previous role but with a slightly different problem focus.

To clarify what you said earlier, when you joined Wayfair, they were beginning to realize that demand forecasting was a source of competitive advantage. However, the third-party solutions they were using were not sufficient. Is that correct?

Yes, there was significant nuance in the catalog. Many of the third-party tools that Wayfair was using did not provide the value they were seeking. Additionally, these tools were quite costly. As a result, they decided to phase out the third-party tools and establish an in-house team to develop this capability.

Why did you perceive demand forecasting as a competitive advantage during your time there? To clarify, when we discuss demand forecasting, are we referring to it at the SKU level, the supplier level, the category level, the entire company level, or perhaps all of them? At what level is it primarily?

We're talking about all of them. It encompasses all levels. In any demand forecast, whether it's for Wayfair or other retail companies, it's necessary to go down to the SKU level forecast because that's what the customer purchases. The more granular you can get, the better. However, as you get more granular, the variability also increases. So there's a balance between what level of detail makes sense. Often, this level is determined by variability and the resulting forecast accuracy. But to answer your question, it was at all levels.

Did you view this as a competitive advantage for Wayfair? Or was this a perspective held by the company that you perhaps did not agree with?

Indeed, I have always viewed it as a competitive advantage, even during my time at Amazon. Retail operating models vary. At Amazon, it's even more crucial as Amazon invests heavily in it.

Wayfair typically operates CastleGate, a third-party network. The inventory is owned by suppliers, not by Wayfair. Therefore, it's essential to know what and how much to bring in to optimize warehouse space and guide suppliers accordingly. This prevents flooding the warehouses, which would necessitate building new ones, duplicating the inventory, and wasting resources.

In this case, it was vital to ensure high capital utilization, and that's where demand forecasting played a significant role. This was a time when the CastleGate business had a higher share than it did during the pandemic, as CastleGate was growing. It was crucial to optimize the capital investment. However, as a growing retailer, you don't want a situation where customers visit your website and can't find products. Amazon has trained customers in the retail space not to expect out-of-stock goods. It's a trust buster when a customer wants to buy something, and you don't have it in stock.

From an out-of-stock perspective, it's also important to ensure adequate supply. This starts with demand forecasting. If you don't have the right demand forecast, you won't have the right supply, and you'll have out-of-stock items. That's the importance of demand forecasting.

Regarding insourcing, many third-party tools are generic and take time to reflect best-in-class algorithms, such as demand forecasting algorithms. The industry is moving quickly in terms of algorithms and what works for different types of datasets, and accuracy improves as a result. When my team members were evolving the algorithms compared to what was used in the previous third-party tools, there was a noticeable difference in accuracy. This shows that it's a competitive advantage if you can evolve your demand forecasting algorithms faster than the industry.

Additionally, Wayfair has always been more data science forward compared to other retailers that generally rely on third-party tools. It's more the Amazon route to say that we don't want to use third-party tools because our growth would be somewhat normalized in the industry as everyone else is using the same capabilities. If you want to outpace the competition, you need to insource those capabilities that you deem strategic for your business. Instead of trying to adjust a third-party tool to your needs, it might be cheaper and faster to build something more customized for your business in-house, where you have more control. That's the route we've taken.

I understand that during your tenure at Wayfair, a significant part of your role was enhancing the profitability of the CastleGate operation. Apart from demand forecasting, what were some of the major initiatives you focused on to boost CastleGate's profitability? Or was demand forecasting the primary focus?

Demand forecasting was indeed a key part of it. However, to clarify, my role involved demand forecasting for Wayfair globally, irrespective of whether it was for CastleGate or not, as the models were adaptable. Although CastleGate had more data and history, which naturally improved the accuracy of demand forecasts compared to items not in CastleGate.

Other areas under my purview included safety stock optimization, which is linked to demand forecasting due to variability. When demand forecasts are more variable, you essentially need a higher safety stock to manage that variability.

We also had to consider seasonal items, which have longer lead times than other items in the catalog and require substantial in-season demand shaping in case of inventory imbalances. This involved strategizing about in-season sales and identifying levers to drive up inventory sales.

Post-pandemic, we had to consider which items were best suited for CastleGate. Before the pandemic, Wayfair was trying to grow CastleGate without many constraints on what suppliers could bring in. However, post-pandemic, due to the supply bullwhip effect, suppliers brought in more supply than Wayfair had forecasted, leading to overcrowded warehouses. Consequently, a part of my role was to strategically decide which items should be in CastleGate and which should not. For items that should be in CastleGate, we needed to ensure we made room for them. For items that were not a good fit, we worked with suppliers to explain why and discuss strategies to make them a better fit. If suppliers were unwilling to adopt these strategies, we considered disposal mechanisms to ensure that the inventory could flow out of the CastleGate network and not occupy unnecessary space.

I can only imagine how challenging demand forecasting must have been during the pandemic and over the past five years. The retail sector, and specifically this category, has seen dramatic changes. For instance, Wayfair was growing at 20% year over year, then suddenly it surged to 100%, and then started to decline. Now, they've been experiencing contraction for several quarters. While some of this is specific to Wayfair, a lot of it is due to how the category evolved during the initial lockdowns and as the world reopened. The task must have been incredibly difficult.

The situation has indeed been challenging. However, there are positive signals. The key is to identify these signals and ensure you have robust models that can detect them, primarily macro signals.

The question is, how do you ensure that you have the correct aggregate number before you start breaking it down to the SKU level? As you mentioned, it's about understanding the overall segment and Wayfair's share in that segment.

Once you've clarified that and arrived at the correct aggregate number, the next task is to break that down into SKU level numbers. This ensures that each SKU has the correct share of the total aggregate.

It's a complex process, undoubtedly. But it's not impossible. There will always be some missing signals, but you'd be surprised at the impact that key signals can have.

That's intriguing. What determines whether an item is a good fit for CastleGate?

There are two factors to consider. The first is your sales history. The more sales history you have, the more information you have about what sells.

Sometimes, best-selling items may not be in CastleGate because the supplier already has a parallel CastleGate network. They might think, why should I send it to CastleGate when I can handle it myself?

Yes, I've heard that from numerous suppliers over time.

Primarily, it's about finding a balance with smaller scale suppliers who have high-quality products. It's important to consider demand signals from across the industry, not just from Wayfair. Even if Wayfair doesn't have the market share yet, others might. So, how do you identify the characteristics of a good product? This could be from the perspective of sales history, customer reviews, growth in sales, and sales velocity. All these factors create a level of consistency that allows for better demand forecasting. That's one aspect.

Another aspect is considering the product's positioning relative to the customer's house in e-commerce. Amazon has trained customers to expect fast shipping, which many customers value. Even for large, heavy, bulky items, customers still care about fast shipping. Part of my role was to identify the relationship between demand and speed, and the demand lift from faster shipping. We found value in CastleGate in terms of speed. Even if you're a large supplier with warehouses across the country, those locations may not facilitate fast shipping to the customer. There's still value in leveraging CastleGate to see the sales lift from being positioned closer to the customers. This isn't necessarily product-specific, it's more of a halo effect from supply chain design.

The challenge is separating whether a product is selling due to its intrinsic characteristics or because it's closer to the customer. These are the interesting aspects of data science that can be applied to solve the puzzle.

Over the years, I've spent a lot of time discussing these issues with suppliers. Wayfair tells suppliers to expect a significant uptick in demand when they put items into CastleGate. I've heard a range of experiences from suppliers, from seeing no uptick in demand and pulling their products out, to doubling their sales for all SKUs put into CastleGate within a month. What kind of demand uptick did you typically see internally? And how variable was it?

That's a good question.

I'm referring to an item that goes from drop shipping from the supplier warehouse to CastleGate.

My team conducted that analysis, and while I won't share the exact numbers, I can confirm that it is indeed variable. It varies by product category because not everyone wants the same level of fast shipping across every home item. There are certain items where customers expect faster shipping, which adds to the variability.

Typically, smaller parcel items, such as houseware tabletop items, are expected to be delivered quickly due to the instant gratification aspect of online shopping. On the other hand, larger items like sofa sets and bedroom furniture often require scheduled deliveries because they necessitate white glove service. Customers might not have moved into their house yet or need someone to assemble the furniture, so they prefer specific time slots.

In these cases, it's less about getting the item the same day or the next. While there are instances where customers want fast delivery, the nature of the product usually doesn't demand it. Consumer behavior doesn't mandate the same lift for these items.

Another aspect that many suppliers fail to understand is the competitive dynamics in play in each segment of the catalog. When customers are shopping, they are looking for a range of similar products. Home goods is one example. Customers don't usually come to the site knowing exactly what they want, so there's a good amount of browsing involved.

When customers browse, they create competitive effects. These effects can be influenced by factors like imagery. A product might be good, but if the 3D imagery on the website isn't appealing, or if the supplier hasn't participated in Wayfair's 3D imagery program, the customer might move past it.

Items closer to the customer are typically surfaced earlier in the sort ranking to encourage customer conversion. However, if the product isn't enticing, the customer will move down. This is a classic case of a good product stored in the CastleGate warehouse with faster shipping, but poor customer conversion due to inadequate 3D imagery.

Another potential source of noise is price. A product might be well-positioned and have the right imagery, but if it's priced higher and the supplier isn't willing to reduce it, the product will move down the sort rank when customers sort by price. This can create a downward spiral for the product.

In most cases, in the retail industry, the popularity of a product is a function of what other customers are buying. The more hits a product has, the better it is perceived. Over time, if a supplier doesn't reduce their price, their product keeps moving down the sort ranking, and they don't see a sales lift because others have taken market share away from them.

Generally, these are the two major factors that many suppliers fail to fully comprehend. They significantly influence customer conversion, and the market is dynamic. The sorting algorithms are dynamic as well. If customers purchase other products that are cheaper and better presented, the ones that are not as well presented or priced will continue to fall in the sort order, even if they are well-positioned. Over time, these products won't experience the expected sales increase because they didn't have the right price or images. In Wayfair's case, suppliers control this. Wayfair can guide the supplier, but ultimately, the suppliers control their destiny.

I completely agree with your point that suppliers often fail to understand this. Over the years, I've had numerous conversations with suppliers who believe their product is the best option at a given price point and style.

That's confirmation bias.

They place an item in CastleGate and are surprised when it doesn't sell. These conversations have led me to a theory that I've been developing. I believe that on platforms like Wayfair, the competitive advantage often lies at the SKU level. Some SKUs have excellent imagery, numerous customer reviews, customer images, and every possible product detail a shopper could want. If a coffee table with all these features is competing against a table with no or few reviews, even at a lower price, customers will choose the one with 65 reviews. It becomes challenging to compete with these perfect SKUs. So, the competitive advantage often isn't at the supplier level, but at the SKU level. Do you agree?

It could be a combination of both. You're correct that a product with reviews will rank higher than a product without reviews. However, consider new products or a new category where Wayfair is just entering, or there's no clear market share leader due to varied distribution across categories. I believe it's a matter of how eager the supplier is to grow. Some suppliers are extremely ambitious and may significantly cut prices compared to others in the category, allowing them to gain market share. Price elasticity in home goods is high because it's a discretionary item. Often, customers will choose a lower-priced item, regardless of whether it has reviews or not. If it's a great deal, they'll take it. I believe that's where you were heading with your point.

That's a valid point. A better way to express my previous statement, which is in line with your comments, is that suppliers often overlook the competitive dynamics on these platforms. They tend to believe that their product, given its quality and price point, should naturally sell. However, they fail to consider the various competitive factors such as price, reviews, and imagery. I've heard success stories about CastleGate, but also about suppliers who claim that Wayfair persuaded them to stock large quantities of their products in CastleGate, which then failed to sell due to inaccurate demand forecasts.

How does Wayfair assist new suppliers entering CastleGate? I've heard it's challenging to gain visibility. It seems difficult to forecast demand for items that haven't been sold on Wayfair before. Additionally, there's an issue where unsold items remain in CastleGate, accruing storage fees. This is detrimental for the supplier and inefficient for Wayfair. Could you explain the onboarding process and the initial demand forecasting challenge?

Let's start with new items. How do we forecast demand for a new item? It might be a new SKU, but not necessarily a new item. The catalog might contain other items similar to the new one. When I was at Wayfair, we developed several algorithms to identify similarities across different products in the catalog.

Based on which product or products the new item is most similar to, either visually or based on the product description, or both, we could predict the potential trajectory of the new item. This process also involves human judgment to estimate the growth trajectory of the SKU.

However, suppliers often believe they can keep everything static, based on a certain price point. If the price changes, the demand forecast becomes invalid. If the inventory isn't delivered at the right time, the demand forecast becomes invalid. This isn't a "set and forget" type of business. Demand forecasts are refreshed every month to account for the latest market share dynamics, price dynamics, and sort dynamics, which then inform the forecast.

The issue is that many suppliers, accustomed to operating in the brick and mortar world where things are static, struggle to understand the complexities of e-commerce operations and market share dynamics. They believe they can set it and forget it. However, if a competitor cuts their price by 50% on the second day of the product launch, and the original supplier doesn't react, their product won't sell because the competitor has taken their market share.

Do you believe the success of Wayfair's platform is dependent on having a sufficient number of suppliers with a wide range of products that can be delivered quickly and at the right price? I'm trying to understand if it's essential to maintain a certain number of suppliers to keep the business momentum going, even if there are other suppliers who may not be satisfied with the partnership.

Over time, you'll notice that there are e-commerce savvy suppliers who are gaining market share from traditional brick-and-mortar suppliers. This is largely because these e-commerce savvy suppliers understand the dynamics of the online marketplace. They've been operating with Amazon for some time, and Wayfair is not trying to reinvent the wheel. The Amazon marketplace operates similarly. The only difference is the maturity level of the marketplace on Amazon versus Wayfair.

There are plenty of suppliers who understand this dynamic. In fact, I believe that e-commerce savvy suppliers are taking market share from traditional brick-and-mortar suppliers. The latter often blame the marketplaces for their decline, failing to recognize that their market share is being taken by e-commerce savvy suppliers.

There's been a significant increase in e-commerce savvy suppliers, particularly in Asia. Consider the product category; it's white-labeled. While there are specialty items, they're not necessarily branded. They're all manufactured in the same facilities across China, Vietnam, and Indonesia. Even the high-end brick-and-mortar suppliers are sourcing from the same factories.

These suppliers are more lean and understand how to navigate the marketplace. They're making a significant impact and, in some cases, are even replacing traditional suppliers. If you look online, you'll see that e-commerce savvy suppliers, particularly from China, have gained significant market share on Amazon.

These suppliers have expressed their frustrations with Wayfair, stating that it's more difficult to work with than Amazon. This feedback indicates that there are enough suppliers to keep the business momentum going. However, it's also important to ensure a diverse product range, including non-generic, visually appealing items that attract customers.

There's a strong focus on evolving the catalog to ensure a balance between new, unique, and innovative products, as well as standard products.

I'd like to revisit a point you made earlier about how Amazon has conditioned consumers to expect fast delivery. In the home category, the importance of fast delivery depends on the item. For example, if it's a houseware item needed for a party on Saturday, fast shipping is crucial. However, for larger items like a couch or a mattress, the delivery speed might not be as important. Generally, smaller parcel items require faster shipping, while larger ones do not. Doesn't this directly impact the value proposition of CastleGate? As I understand it, Wayfair's rationale behind CastleGate is that it allows for quick delivery of large, bulky products, something most other companies aren't equipped to handle. What are your thoughts on the value proposition of CastleGate and its competitive advantage, or lack thereof, compared to Amazon, Walmart, Home Depot, and Target? Do you agree with the investment case behind it?

Your articulation accurately captures the value proposition. My earlier comment was more about the relative sales lift. It's not that large furniture items don't provide any sales lift. They do. However, there's a range. Small parcel items might be at the upper end of that range, while larger items might be in the mid to lower end. But overall, that range is still higher than what a standard supplier or marketplace can achieve. In summary, there is a clear sales lift.

Do you have an opinion on the cost aspect of this? Specifically, how much cheaper it is to sell through CastleGate versus drop shipping or versus FBA?

When I was at Wayfair, a lot of the tools we were developing aimed to optimize total profit for the supplier. We ensured that the items we recommended for specific warehouses were profit optimal for the supplier. These inventory decisions were based on a profit optimization formula, which considered the cost of positioning items in particular warehouses. So, when Wayfair recommended a supplier to position a certain amount of inventory across specific warehouses, it factored in the cost of positioning in those warehouses.

So, for items going into CastleGate, are you suggesting that it's cheaper for the supplier to sell through CastleGate than to handle it on their own or sell through FBA?

Yes, after considering the demand lift.

I see. We've discussed a lot about this demand.

I believe that many suppliers are not sophisticated enough to understand the dynamics of the market.

I concur with your observation. That's the impression I've gathered from numerous conversations with suppliers over the years.

Wayfair has been making significant efforts to educate suppliers on market dynamics. They engage extensively with key suppliers, but a broader transformation is still required across the supplier base. The shift towards e-commerce savvy suppliers gaining market share from brick and mortar stores hasn't fully materialized.

Currently, brick and mortar stores hold the majority of the market share, but e-commerce savvy suppliers are growing rapidly. Interestingly, brick and mortar suppliers, due to their larger asset balances, have the capacity to invest more in new product development.

Therefore, it's crucial to maintain a balanced engagement with brick and mortar suppliers. E-commerce savvy suppliers are generally reactive, they pick up on trends and respond effectively, thereby gaining market share. However, it's the brick and mortar suppliers who are innovating and designing new products.

The challenge lies in educating these brick and mortar suppliers to comprehend the market dynamics and understand the quality of recommendations that Wayfair provides, and why these recommendations are beneficial for them as well.

We've discussed demand forecasting at the SKU or supplier level. Now, let's shift our focus to a higher level. Wayfair has outlined their company-wide growth algorithm for normal times. It's been quite some time since we've had a normal year in this category.

In a typical year, the category might see low to mid-single-digit growth, with e-commerce gaining share on top of that. Wayfair is a significant share taker within e-commerce. They also have other growth initiatives such as their commercial business, professional business, Perigold, and now physical retail.

If we sum all these factors, Wayfair suggests that they could achieve growth of 15%, mid-teens, high teens, or even better in a normal year. This is a significant departure from their current position and their performance over the last few years.

Considering the company-wide level, do you have an opinion on how Wayfair might perform over time?

I won't speculate, but I will say that the current macroeconomic signals indicate that inflation remains high and home buying activity is not as robust. Wayfair’s category is discretionary, and in times like these, demand tends to remain soft until people have more disposable income. This situation won't last forever, as various macroeconomic projections suggest improvement, but it will take time.

From a generational perspective, we are seeing a mix of generations who prefer omnichannel retail. The generation with disposable income still prefers brick and mortar retailing, especially for home goods, because they can touch, feel, and enjoy the experience. The upcoming generation, and perhaps even my own, prefer online shopping due to its convenience and time-saving aspect.

This shift requires companies like Wayfair to have both a physical and online presence. In my opinion, Wayfair has been a bit late in establishing their physical presence in this category. However, I believe that physical stores could be a significant turnaround for Wayfair. I don't think the company is moving fast enough on this front, but they are trying.

A combination of tailwinds from Gen Z earning more over time and driving the e-commerce business forward, along with physical stores that cater to the generation with higher disposable income, will allow the company to grow.

I tend to agree with you that physical retail could be a significant factor for them.

I believe the rate of year-over-year growth will largely depend on the rate at which physical retail can be expanded and improved. How can Wayfair create a differentiating experience compared to Ikea or Pottery Barn, where people still enjoy shopping? These companies are still doing well. Wayfair has always aimed to grow faster than them due to their e-commerce business, but the market is soft.

It's crucial to understand which part of your customer segment has the disposable income to spend and how to capture their attention to convert that disposable income into sales. As you know, the market is highly fragmented and there's significant income inequality across the market. Despite high interest rates, high-value purchases are still happening.

To conclude, let's integrate everything we've discussed. We've touched upon logistics, customer experience, merchandising, and selection. How would you rank Wayfair in its category against competitors like Amazon and Walmart? Do you think Wayfair's business model has advantages that these other players will struggle to replicate? Or do you think these larger, more sophisticated companies will eventually be able to emulate what Wayfair does? Essentially, I'm asking about the competitive dynamics between Wayfair and its competition over the next five to ten years.

That's an excellent question. From my personal perspective, having worked at both Amazon and Walmart, I believe that Wayfair will continue to lead these competitors in the category it focuses on. This is the advantage of verticalization, where you can concentrate on creating delightful experiences for your customers that are specific to a particular product segment or category.

Amazon and Walmart, on the other hand, focus on a wide range of product categories, and the home category is not their top priority. It might take them a while to make it their top priority, if they ever do. In some ways, the home category is more about customers wanting to shop under one umbrella.

Amazon and Walmart may not offer innovative products or items that inspire customers. They provide the basics, but if a customer wants something special and unique, they won't find it at Amazon or Walmart. They will find it at Wayfair.

The challenge for Wayfair is to continue educating customers to come to them for these unique products. This need also evolves over time. For example, a Gen Z individual may not want a unique product when they first furnish their home or apartment. But as they start to want to live better or more eco-friendly, their tastes may change.

These will be interesting market trends to watch. It will depend on who can capture these trends. Historically, furniture design has become less intricate over time. But it's about capturing when these trends change and ensuring that even in a market dominated by generic products, you can still carve out a place for yourself due to the loyalty your customers have for you.

It's also a matter of experience. If you consider the returns experience that Amazon or Walmart offers for home goods, it's not as good as Wayfair's. These are capabilities that Wayfair will likely continue to improve on, maintaining their first-mover advantage. Can other companies catch up? Yes, they can. However, will they catch up in the next five to ten years? Perhaps not, as they have other larger priorities.

So, would it be accurate to say that, in your view, Wayfair's sole focus on this category is the reason for their success? This singular focus results in numerous small advantages, such as differentiated logistics, merchandising, imagery, customer support, and returns experience. Rather than one major advantage, it's a multitude of smaller ones that keep Wayfair ahead in all these areas. Is that a correct summary of your statement?

Yes, it's about cost structure and experiences. By ensuring that you can provide these experiences at a lower cost structure than your competition, you essentially build a competitive moat around yourself. This gives you a lead time, making it more difficult for others to catch up. They may even reconsider whether they want to make that investment or not. This is a classic play between vertical and horizontal marketplaces. Amazon and Walmart are horizontal, while Wayfair, Etsy, and Chewy are vertical. This allows you to delve deeper into the customer experience and build capabilities faster. For instance, building the demand forecasting capability that I developed at Wayfair would take longer at Amazon or Walmart. The dynamics are different.

To conclude, I'll share my thoughts regarding the physical retail conversation. I believe that Wayfair is going to unlock a truly differentiated customer experience with their new store in Chicago. I think they're going to use CastleGate to drive this unique experience. For example, in a Pottery Barn or a West Elm, some items can be shipped quickly if they're in stock. However, I suspect that Wayfair will require suppliers to have their items in CastleGate if they want them displayed in the store. This way, when a customer likes a couch, instead of waiting two or three months for delivery like at West Elm, Wayfair can deliver it for free on Tuesday. If they can provide this kind of experience through CastleGate, it will be a significant differentiator compared to other retailers.

Indeed, consider the fact that the inventory in CastleGate is under Wayfair's control. The worst scenario is when a customer likes a particular product in your store, and you have to inform them that it's out of stock and will be available in four months. That's a sure way to lose a customer. It's not just about incentivizing the supplier, but also about enhancing the customer experience.

I've experienced the "pay $200 for delivery and wait four months" scenario before, and it's not something I wish to repeat. This conversation has been very informative and helpful. I've learned a lot.