Buzzing Bazaars
As part of our journey to gain a deeper understanding of business models, we’ve been attempting to catalogue our insights and research on a variety of them. We thought we’d turn one of these exercises into a blog post. Of late, we noticed that we’ve been recommending a number of direct and indirect investments into online marketplaces, and figured it would be a great model to start with.
A thought experiment
The earliest marketplaces are said to have originated 3,000+ years ago in Persia, where merchants and buyers would flock to local markets to trade goods. It’s believed that strict zoning policies in Persian cities about where trade could be done led to the creation of these marketplaces. However, what was probably a frustrating regulatory regime gave rise to one of the most impressive business models ever created.
To understand why, let’s fire up our imaginations and go through a thought experiment. You’re a merchant in Babylon who sells carpets that you’ve sourced from your hometown, a 3-day journey from the city. You’re new to the city and have just set up shop in a far corner of town, where rent is reasonably affordable. You don’t know anyone, have no regular customers, and don’t have much of a network in the city. But you’re an industrious type, and want to make it for yourself. You know you sell high-quality goods, and the feedback from the buyers you do have is very positive. However, it’s still tough to generate business. You try to capture as much foot traffic that comes by your store, but it’s not much. You pay a town crier to advertise for your store at the town square for a few hours each day, but your budget is limited.
After a few months of this, you’re just about to pack up and go back to your hometown, when a city representative comes through your door and tells you about a new concept the city is launching called a ‘bazaar’. This bazaar will be near the center of town and host the best merchants selling the best goods in the city. You’re intrigued by this concept, and even though there are increased taxes on goods sold there, you don’t have much to lose. So you give it a go, pack up the shop and move locations into the bazaar. Immediately, you start to see more foot traffic to your store, not only that but these are people who came to buy items, not just someone walking past your store. Because the bazaar is ‘the place’ in the city to shop, there is a constant crowd, morning to night. You do have to haggle with the administrator to get a prime position; do up your shop in a manner that attracts crowds, but on the whole, this is a much better setup. The only pesky point is that the bazaar administrator takes a 15% cut of everything you sell. You wonder if this is worth it, and break out your stone tablet to do the math.
It’s clear – you’re making a lot more money at the bazaar. You thank Marduk for your change of fortune, and write home to tell them you’re going to stay in the city a bit longer.
Online marketplaces – the modern-day bazaars, offer similar advantages as their ancient predecessors did. If you think about an independent merchant’s journey today, it plays out similarly to an independent merchant in Babylon. They have to set up a store (website) which is often in a far flung location (page 3 or 4 of google results), pay rent (hosting fees, payment fees, etc), enlist a town crier (Facebook ads) which can be expensive, and hope someone visits them (web traffic). While there are benefits to this model, especially if you have a big brand, it’s less likely that if today you’re a mom & pop store selling iPhone covers, you will want to pursue this method.
What is an online marketplace?
However, before we go any further, it’s important to define what a marketplace is in the modern context. A marketplace is an online property which attracts both buyers and sellers, and facilitates a transaction between the two. It’s important to differentiate marketplaces (third-party or “3P” model) from general e-commerce (frankly, every company is an e-commerce company these days) and more importantly, from online retailers (first party or “1P” model). Unlike online retailers (think Walmart’s online store or much of Amazon’s business), marketplaces do not take on any inventory. For a marketplace to succeed they need to do a few thing well
Build their supply base – No one is going to come to a market that has no stalls or where product discovery is weak. Not only does this require attracting high-quality merchants, but it also requires quality control and rooting out scam artists.
Bring customers to the sellers – Sellers will give up if they’re not making enough sales. Despite most marketplaces having minimal to no setup costs, responding to customers, inventory management, and online shop upkeep is time consuming. Attracting first-time customers requires heavy marketing, discounts, and potentially free shipping.
Facilitate the transaction – If the role of the seller is to display products of value and the role of the buyer is to purchase those products, the role of the marketplace has to be to facilitate the transaction via payment processing and to handle refunds/chargebacks.
Provide for a seamless user experience – The seller must be able to display their products easily and make changes quickly. The buyer must be able to search for products in an efficient manner, and have avenues to deal with refunds and returns.
These tasks, while not exhaustive, make the backbone of a great marketplace. The first point about building a merchant base is, in our opinion, the most crucial. Marketplaces are nothing without high-quality merchants in it, and while starting a marketplace is easy enough, scaling it is very difficult, and takes time, money, and sharp management.
Makings of a great marketplace:
Looking at the marketplaces we’ve invested in (or wanted to invest in until the price ran too high) a few key characteristics jump out at us when considering a great versus average marketplace:
Fragmented vs consolidated environments: When a marketplace is consolidating a number of small merchants, it builds an ecosystem of choice, fairness, and overall trust in the system. However if one merchant has too much power, it can lead to a breakdown of the marketplace. Let’s take two examples. The first is a marketplace that consolidates a number of small merchants and brings to them demand they would not have otherwise seen. None of the merchants on their own have too much bargaining power as each individual makes up a small percentage of sales in the marketplace (and the overall industry). In industries such as fast fashion, even large brands that might join the marketplace don’t have any sort of monopolistic power over consumer choice. This allows for a great consumer experience as buyers are exposed to brands they may never have heard of at great prices, and merchants get access to numerous customers who would have never come to their digital or physical storefronts. Let’s compare this to a marketplace for airline tickets. On the surface, the proposition seems the same, as online travel agents try to consolidate a number of airline prices and options to make it simple for the consumer. However in any given market/route, there are only a few major airline players that control most of the business. Thus, the benefit of a marketplace to them is quite limited, as consumer awareness of them is high, trust is already established (consumers don’t have a problem transacting with their local carrier), and market share is consolidated. We saw this take place in real-time when over the last several years we’ve seen airlines (and some major hotels) offer the best pricing and selection on their own websites rather than on marketplaces.
Local vs Global/Regional supply: Ride-hailing is a fascinating form of marketplaces, in which they match drivers and riders in a seamless process. However ride-hailing apps are not one marketplace, rather they are a collection of several marketplaces across geographies. This is because a driver in Singapore is of no use to a rider in Malaysia. While ride-hailing companies have overcome this by aggressively scaling in several countries – the unit economics of one sub-market does not help the unit-economics of the other. Essentially, each time say Uber or Grab enter a new region, they have to build the business from scratch. On the other hand, marketplaces focused on goods and online-services are very scalable. Fiverr is an excellent services business which links freelance service providers with those who require those services (for example designing business cards). For most services, a freelancer sitting in Tel Aviv can just as easily help a customer living a few blocks from her as she can a customer living in Tokyo. This allows for rapid scale as the supply side does not have to be re-built in every single jurisdiction the company enters.
Low AOV/High Repeat versus High AOV/Low Repeat – All things being equal we tend to view marketplaces which have low average order values and high repeat purchases versus those that do not. Again we say all things being equal, because there are plenty of marketplaces selling large ticket items (cars, houses, etc) which have done really well for their customers. However, given two marketplaces in the same vertical, we found that companies focused on lower AOV and high repeat purchases tend to get an edge. For example, Shopee (owned by Sea Limited) was launched in Southeast Asia three years after its biggest rival Lazada (owned by Alibaba) began operations. Both are marketplaces that sell a variety of items from clothes to household goods. Lazada in its earlier years tended to focus on large ticket items such as electronics (in fact when we met Shopee management, they admitted Lazada was still better than them in electronics). On the surface, electronics seems like a great category as ticket items are high, thus commission for the marketplace will also be high. However, the problem is that electronics are an infrequent purchase (you’ll only buy a TV once every 3-5 years or so), and if the customer is not coming back frequently to the marketplace to make purchases, it becomes harder to recoup the cost of customer acquisition. Shopee, on the other hand, focused on fast-fashion, makeup, etc, which are categories which have low ticket size but result in frequent purchases. This allowed Shopee to make up ground and overtake Lazada even though it launched 3 years later.
Chart lifted from Hayden Capital’s Presentation
High Trust – One of the key roles of a marketplace is to establish trust between suppliers and customers. Even though we docked them a few points in a previous point, trust is one thing that ride-hailing apps do very well. Could you imagine getting into a stranger’s car a decade ago? This trust also extends to quality control, and if a marketplace is filled with sub-par products or service providers, trust in the entire system erodes. A key example of this is the Indian hotel marketplace, Oyo. Oyo is actually an excellent idea, founded by (at the time) a teenager who was fed-up with the terrible quality of lower-tier hotels he had to stay at when traveling. So he worked with those hotels to refurbish certain rooms under the Oyo brand which would represent cleanliness, good service, and consistent quality. In turn, Oyo would help hoteliers sell those rooms online via Oyo’s website/app. The problem was that Oyo scaled so quickly that they could not maintain that quality, which led to a significant backlash from customers who posted horrific photos of rats, sewage, and dirt staying with them in their Oyo rooms. This quickly eroded trust in the offering, and led to Oyo pulling out of a number of countries (covid didn’t help either).
The Flywheel and Unit Economics
Now we can’t have a conversation about marketplaces which does not mention the biggest business model benefit of a 3P model versus a 1P model – the flywheel. We won’t go into too much detail here as this is a well-covered subject, as most of you would have seen this famous Amazon flywheel image (I know we said above that Amazon was a 1P model but it’s actually now a blend of 1P and 3P). Any good marketplace will eventually start to see this flywheel take place and spin faster and faster as management executes on building the supply side, effective marketing, quality control, and customer satisfaction. The reason why 1P models typically do not see the same kind of flywheel is because it’s costly to bring on ‘supply’ as the platform itself is the seller and has to take inventory risk (and tie up capital in the process).
Now generally the unit economics of a marketplace are very attractive. They have gross margins of 80% with the cost of goods sold made up of payment processing fees, hosting fees etc., all of which the marketplace covers on behalf of the seller. However, you rarely see marketplaces EBITDA-positive in their first several years of operations as they spend most of their margin into marketing and promos in order to attract buyers and incentives/training/shipping subsidies in order to attract sellers. Eventually though, the unit economics should work out if customers repeat purchases continue to increase. As can be seen from the most recent Redbubble (a marketplace for independent art) earnings report, the marketplace experienced an explosion in repeat purchase growth. This has led to the company to put up impressive EBITDA margins compared to historical numbers. This is also a representation of the flywheel working, as more customers come to the site, the more artists they attract, the more artists the more selection a customer has, and the more selection, the more customers that come to the site, and so on and so forth. When speaking to an ex-employee of Redbubble, we asked him why we thought Redbubble attracted artist in much higher numbers than their competitors and he quickly replied “because that’s where all the customers are!”.
Redbubble saw a strong increase in repeat customer purchases, making a very profitable half year for the company.
Does the market agree with marketplaces?
We’ve noticed an interesting trend of late where traditional 1P businesses have been shifting towards 3P models as the unit economics become quite clear. This has particularly been evident in European e-commerce players where marketplace penetration is very low. Below you’ll see excerpts from the investor presentation from CDON (Sweden), OZON (Russia), and Cnova (owns Cdiscount in France) all of who are shifting from 1P models to hybrid models. These excerpts show how each company is purposely highlighting/stripping out their marketplace returns from their overall returns. The YTD returns for the companies show that the market is appreciating this shift, as they have far outperformed the S&P 500 (Ozon is up 40% YTD, Cnova up 150% and CDON up 258%!) .
CDON was spun-off from its parent company in late 2020, and is building its marketplace business to be its key value proposition.
OZON’s transformation into a hybrid 1P/3P model has been well receive by customers and was a key part of its pitch during its recent IPO.
Cnova credits its marketplace business as its key growth drive. They also credited a product focus shift from electronics to household goods/food/beauty as a reason for increased profitability via repeat purchases.
Conclusions
We believe marketplaces will dominate online commerce models going forward due to their low capital intensity, impressive flywheel models, and benefit they provide to both merchants and customers. That said, we believe that not all marketplaces are created equal as some models tend to work better than others, and as with any company management, execution, and valuations play a big role in long-term success.
We hope you found this write-up informative and got you thinking more about marketplace business models. As usual thanks for reading and happy investing!