Full-Stack Marketplaces — Emerging Markets, VC disagreements and the future of Marketplaces
Full-Stack Marketplaces are one of the most interesting and misunderstood developments in the world of marketplaces. I try to understand the levers that go into why a full-stack (as opposed to ‘lite’) marketplace evolves, the current discourse within the VC community and its relevance to emerging markets.
- What is the Job to Be Done for a Full-Stack Marketplace?
- Understanding the next trend of Marketplaces: Full-Stack Marketplaces
- Definition Disagreements and its importance
- Are Full-Stack Marketplaces better?
- Boundary Conditions for Full-Stack Marketplaces
- Relevance to Emerging Markets
- Further questions to ask
In the last part of this article, I will also talk about how I believe such full-stack models are more appealing in emerging markets due to issues of trust, fragmented supply chains and binary capital fundraising outcomes (the best founders can disproportionately raise a significant amount of capital going to a market as opposed to places like the US)
High capital requirements and complex transactions made real estate (or proptech) a perfect breeding ground for MINOs — other examples include WeWork (office space), Compass (estate agents), and Nested. However, proptech is by no means the only place to find them. MINOs are also common in the used car space — Vroom, Carvana, Fair.com, Cazoo, Beepi, etc. Scooter rental startups like Lime also fall squarely into this category. As do fashion startups like Rebag.
What is the Job to Be Done for a Full-Stack Marketplace?
Full-Stack Marketplaces tend to expand vertically and become full-stack typically because they:
- Aim to capture a greater wallet share/take rate to make economics work
- Aim to provide more consistent supply/inventory/quality control
- Aim to fundamentally disrupt the value chain by owning multiple parts of the supply/value-chain simultaneously
Full-Stack Marketplaces solve for the following structural limitations of existing industry players/dynamics:
- Opaqueness of Quality
- Complexity of service delivery
- Variability of Service Quality
Understanding the next trend of Marketplaces: Full-Stack Marketplaces
The first thing to make clear is the scope of what a full-stack marketplace is. There is a lot of linguistic confusion over what is the definition of a managed marketplace vs a full-stack marketplace
A full-stack marketplace should encompass ownership or liability over significant portions of the value-chain it is involved in, which can typically be observed by looking at what liabilities (in terms of contract liabilities as well as potentially asset ownership) sits on the marketplace.
A managed marketplace should have some level of filtering, service/product standardization and even tools that help to improve the existing process flows. Managed marketplaces can be ‘lite’ like Amazon where the barrier to being a merchant is relatively low or Uber, where the operating standards set is relatively low, or it can be a ‘heavily managed’ like GOAT, which requires stringent criterias to be set before products can be transacted on the platform.
The difference between ‘lite’ and ‘heavy’ is, for now, subjective but encompasses the following criteria:
- Filtering the onboarding
- Filtering the interaction
- Severity/strictness of standard-setting of the services provided
- Visibility of supply-side branding
To quote Boris from Version One, who recognizing the linguistic confusion in the industry, notes that a managed marketplace ‘takes on additional parts of the value chain to deliver a better overall experience’.
For example, Beepi, a peer-to-peer marketplace for buying and selling cars, inspects every car listed on the platform — giving people a more trustworthy, stress-free way to buy a car. The platform also adds value to its sellers by guaranteeing a sale: Beepi buys the car if you don’t sell it within 30 days.
The managed approach isn’t limited to auto marketplaces. Threadflip and Real Real handle the logistics/cataloging for its sellers in women’s clothes and luxury products respectively.
In contrast, Full-Stack marketplaces typically own the end-to-end process within that value-chain.
Breadcrumb.VC has by far the best definition — using the ‘Marketplace In Name Only’ to describe Full-Stack Marketplaces.
In some cases, companies get so involved in the transaction that buyers and sellers no longer interact with each other. Instead, the company takes on the role of the supply side itself to control the customer experience. When this happens, buyers are attracted to the inventory, assets, or resources owned by these companies — not very different from a retailer like Tesco or asset rental company like Hertz, albeit enabled by technology.
However, Breadcrumb.VC makes the claim that-
Since the demand and supply sides do not interact, network effects no longer exist. In other words, they no longer exhibit the defining feature of a marketplace. For example, in Opendoor, buyers are attracted to Opendoor’s inventory of homes and not the number or variety of active sellers.
I disagree with this. Just like any broker, Opendoor’s value proposition is enhanced by a greater amount of sellers interacting with Opendoor. Opendoor’s inventory is tied directly and comes from the number of high quality sellers interacting with Opendoor. This marketplace function and network effect is thus an internal marketplace rather than a traditionally understood external-facing marketplace, but the levers for success remains the same (i.e. greater amount of sellers and buyers).
Furthermore, there is also a slight contradiction later in the article whereupon Breadcrumb notes that MINOs can subsequently build network effects to be defensible, which contradicts the earlier statement about network effects.
Nevertheless, Breadcrumb’s acknowledgement of the distinction that Full-Stack Marketplaces are very distinct from Managed marketplaces is fundamentally astute.
Definition Disagreements: Some other examples of definitions I differ from:
A16Z includes many great elements of the industry progression to a managed marketplace. However, I believe it inaccurately classifies models like Opendoor as a managed marketplace rather than a full-stack marketplace.
In this great article by Point 9 Capital, they classify Airbnb as a full-stack marketplace, which I believe is a mistake — given that while Airbnb certainly manages and filters inputs into the marketplace, there are plenty of other operational aspects of the real estate inventory that are not controlled by Airbnb. As an example, a platform like Blue Ground which leases and operates the real estate inventory is far more operationally involved.
And its importance: Why does this distinction matter?
As Breadcrumb VC notes, Full-Stack Marketplace tend to be more capital hungry, have weaker moats, and are more expensive to grow than true marketplaces (“managed” or not). These companies need to find other avenues to create defensibility beyond capital — layering network effects on top of this model is one such avenue. — Breadcrumb.VC
The margin profiles of Full-Stack Marketplaces can also fundamentally change what can be possible with the business model.
Larger margin profiles (backed by venture capital) can help introduce a 6-star customer experience in Full-Stack Marketaplces that may not previously have been possible.
Are Full-Stack Marketplaces better?
Full-Stack Marketplaces will likely result in higher operational complexity and expenses, high capital requirements (thus dilution or a lower rate of capital return), and scaling issues.
Full-Stack marketplaces are therefore not better or worse, but rather fundamentally dependent on the industry being tackled.
Boundary Conditions for Full-Stack Marketplaces
A. Require situations where:
- Quality control or quality expectations are unreliable (and the corollary — that Branding effect have significant weight in the industry — is typically true), or where a new customer behavior/experiences is being created where existing supply is inefficient in delivering said experiences, and;
B. May be better where:
- Capital requirements are typically high
- The value chain is fragmented or complex
- In Traditional industries/emerging markets where traditional relationship-based supply-chains or network effects are hard to crack
The way I personally think about full-stack marketplaces is that it is a simple equation of:
[Cost of Capital + Opportunity Cost (Focus + Hiring) + Marginal Revenue] < [Opportunity cost of reliance on downstream/upstream suppliers]
In traditional industries, in particular, the marginal revenue is high given the existence of abnormal middlemen margins, and the opportunity cost of downstream/upstream suppliers are high due to poor reliability.
Upon securing the supply side by solving for either complexity of service delivery or variability of service quality, scalability is a function of brand value. In other words, once you know you have a better backend process, capturing the advantage by building a standout brand perception is the next necessary step (through high Brand NPS).
A function of this is that Organic Channel Customer Acquisition should increase and further compounds over time. Premium pricing may also be a function of branding power, but is not a necessity.
Examples of Full-Stack Marketplaces:
Examples of Full-stack marketplaces, especially in emerging markets include Zillingo, Zetwerk and Shein.
Zilingo and Shein have very relatively similar models within the fashion ecosystem, incorporating a full-stack model to serve B2B and B2C customer segments respectively.
Zillingo, a SEA B2B marketplace, focused on the fact that inefficiencies within the B2B wholesale fashion industry were due to bottlenecks that cropped up from time to time within various aspects of the supply chain, be it working capital, materials, design delays, etc. Within the complicated supply chain that was fast-fashion, Zillingo realized that a full-stack marketplace model made the most sense, and worked with manufacturers, designers and procurement teams to use their all-in-Zillingo tech platform. In doing so, Zillingo became one of Southeast Asia’s fastest growing Unicorn.
In a similar manner, Shein, a B2C US/China marketplace, focused on the fact that those same bottlenecks had consequences for B2C fast fashion plays.
Shein also took on roles and responsibilities traditionally handled by factories such as prototyping, an expensive but necessary step in producing clothes. Shein brought this work in-house, simplifying the relationship while reducing risk and cost for its partner factories. Before long an entire industry was falling over themselves to gain Shein’s business. In return, there was one requirement: everyone had to use Shein’s supply chain management (SCM) software.
Today, Shein has been described as the fastest growing e-commerce company in the world, with $10B in revenue in 2020.
Last but not least, Zetwerk showcased similar principles within the Indian B2B industrials custom manufacturing space.
Custom parts manufacturing was extremely complicated due to the need to communicate and find suitable manufacturers. As both procurement and manufacturing were always unpredictable and in small batches, the industry was extremely undigitized and inefficient.
More than a simple marketplace, Zetwerk also provides software for manufacturing, design, project planning, auditing etc. More than just a SaaS enabled Marketplace, Zetwerk has also been able to leverage the project-data flowing through the platform to understand the custom needs that were left unfulfilled, and built their own captive manufacturing hub to serve the missing gap of needs themselves.
As a result, Zetwerk, a 3 year old company, has raised close to $200M from Sequoia, Accel and others to tap into this huge market and build bridges between Indian manufacturing and the rest of the world.
As marketplace opportunities become increasingly crowded, we are likely to see more innovations around managed and full-stack marketplaces to demonstrate superior customer value propositions or superior unit economics.
Relevance to Emerging Markets
I believe that we will see full-stack marketplaces much more commonly in emerging markets due to:
- Trust deficit higher in emerging markets than developed markets
- Capital deficit in emerging markets covered by VC capital willing to allow startups to blitz-scale
- More unreliable & patchy infrastructure/actors within parts of the value-chain represent opportunities or necessitates deeper vertical integration
- Relationship-based supply-chains may require building parallel supply-chains to ensure competitiveness
- Shallower markets (in terms of wallet size/GDP per capita) require deeper horizontal integration to achieve venture-scale valuations.
Some examples of Full-Stack Marketplaces that have grown significantly in scale in emerging markets include Udaan, Zillingo, and Zetwerk.
As a quick aside, many emerging market founders I have spoken to have found embedded finance, specifically embedded lending, as a useful way to patch up capital and cash flow deficits among their supply-chain. In my opinion, this is a half-way point to being a full-stack marketplace, and points directionally to the depth of this specific job to be done.
Further questions to ask:
- Does increasing supply-chain complexity within an industry lead to a ceiling for the full-stack marketplace model?
- Is it necessarily true that all verticals are moving towards heavily managed or full-stack marketplaces?
- Across other verticals, can we find complicated service and product delivery categories that can be moved into a full-stack marketplace model?
Any areas I missed out? Feel free to reach out.
If you are creating more efficient ways to connect supply and demand, discussing next possible features for your Marketplace or moving up the stack, we are more than happy to hear from you.
Chia Jeng Yang, Principal at Saison Capital, dives into consumer, SaaS, and fintech investment trends across the U.S. and Asia, builds projects in the venture capital and public policy space, works closely with early-stage (Pre-Seed) founders and can be contacted at firstname.lastname@example.org. Previous work here: http://chiajy.com