Marketplaces have created some of the world’s highest-profile consumer businesses in the last decade. Think DoorDash, Instacart, Airbnb, Spotify, Zillow, Uber, and Lyft. Although much has been written about marketplaces, there are still a lot of misconceptions on best practices to start and build enduring marketplace startups. Inovia’s Venture Partner, Andre Charoo, sat down with Benchmark’s Sarah Tavel, NFX’s Pete Flint and Version One’s Angela Tran for a conversation about the misconceptions in building marketplaces businesses.
PART ONE: Solving the chicken and egg problem
Andre Charoo: Angela, which comes first, demand or supply? Which side is more important? On the supply side you’ve written about identifying unique inventory. Can you explain further and also how you convince them to join?
Angela: The misconception that we see most commonly is having this mindset that we need to collect all of supply first or all of demand first, when it’s really a stepwise approach. Get some supply, then some demand, then some more supply, then some more demand; this where you really get the flywheel going. The MVP isn’t about aggregating all supply or aggregating all demand, but showing that you’re able to efficiently connect very specific supply with very specific demand in the early days. It’s very common to see the marketplace of supply first because I think it’s easier to find people wanting to make money and make themselves known much more than people who are looking to buy.
Your question about unique inventory, we should caveat by saying not all marketplaces have or can have unique inventory, especially if you’re aiming to build in a space where there might already be an incumbent or an adjacent competitor. In any case, you can find unique inventory and markets where supply is not online already. This brings to mind a lot of old school antiquated industries or B2B marketplaces, I think that’s how they’ll form, or in categories where they just weren’t as developed. A classic example is Airbnb where shared accommodations had existed on Craigslist, but it just didn’t take off until Airbnb came. Everyone knows the cool growth hack of Airbnb hiring professional photographers to increase their unique supply.
We just led an investment in a climate company that’s mapped every tree in the US with satellite data. We see this as data driven creation of supply because the company now can go to landowners and ask them not to harvest or cut down their trees, which is how they normally make money, but instead encourage them to sell carbon to Fortune 500 companies that are looking to offset their emissions. That’s a new idea in terms of data-driven creation of supply, something that we kind of luckily stumbled across.
To your last point, how do you convince suppliers to join your platform? I think the most common tactic is really providing ways in which supply can manage their business. Whether it’s SaaS based or whatever tools that you can offer.
Andre Charoo: Reflecting on the early days of Uber during the limo Uber Black product, we paid drivers to secure supply. Then we weaned them off of hourly pay as demand came, to bring them on to commission.
Andre Charoo: Pete, you’ve talked about a concept of creating network independent value. Can you share more on that?
Pete Flint: The fundamental thing we’re trying to figure out here is that it’s incredibly hard, as an early stage marketplace startup, to get that liquidity on one or both sides of the marketplace without spending a lot of money. You have a small amount of capital, you need to get to that scale, ideally on both sides, but certainly on one side, the most important side early on. How do you do that?
Until you get to that scale, how can you provide value to either the infrequent or the casual user or an individual? In the case of Trulia, the way that we approached that was, how do we give information about the home that the consumer was interested in? Zillow did a similar thing where we provide comps, neighborhood information, schools, they put the valuation, stuff you are interested in, but we didn’t have all the listings. People come for that information and then once you’ve got a portion of the demand, the supply will start to fill in.
This notion of network independent value is critical to provide value before you get the full marketplace up and running. In 2020, there’s quite a lot of competition around some of these big categories, and what I look for is what’s the unique insight or distribution channel or product, What’s that growth? What’s that tactic? What’s that platform that you’re trying to explore? I think it’s often misunderstood by founders, how hard that is, and also how important that is to figure out.
Andre Charoo: When is your business a SaaS-enabled marketplace? Or is it simply a SaaS business and then separate to that, you can launch a marketplace once the participants are on-board.
Pete Flint: That’s certainly one of the most increasingly common tactics around building a SaaS tool, which provides engagement that can often lock in supply that can provide utility. The classic example is probably OpenTable providing reservation information and then building the marketplace on top. It’s very interesting in a couple of dimensions. One is that you deeply engage with a number of customers, you understand their business, you understand their workflow, you get into payments, which is very valuable. Then you can start to build sort of marketplace dynamics on top. It’s very defensible in the sense that SaaS tools are kind of limited defensibility other than features and brand, etc, but when you add a marketplace on top of that, it becomes incredibly valuable.
Furthermore, trying to sell a SaaS tool is quite hard. This sort of CAC LTV on those things is, particularly on any kind of SMB, is really hard, but the retention is often very high. So if you combine what is a hard sales process with a high retention product, and add a marketplace, which is often the flip of those, you can sometimes have the killer combination
. These sort of SaaS tools with marketplaces have, particularly over the last couple of years, been extremely valuable for startups to explore.
Andre Charoo: There seems to be an evolution on the various types of marketplaces. I think Sarah, your partner, Bill Gurley, tweeted a while ago, “what’s the difference between a managed marketplace and a true marketplace?” Why are we seeing this evolution?
Sarah Tavel: One of the crazy things that Uber did, was this experience of the ‘on-demand’. You don’t have to choose who the supply side is, Uber will make that match for you. The amount of friction that it removes from that matching process and marketplaces are constant, it’s a never ending battle to remove friction from the transaction.
That introduces this idea of traditional marketplaces that have always been about more supply, more supply, and surfacing the things that are good. Then you see this kind of counter reaction to that, which is make it easy for people to find the right supplier. Don’t rely on people having a bad experience to create a bad reputation for someone and then get rid of that supplier; get more involved, vet the supply. This is where it becomes a spectrum of how involved you want to be in the curation of the supply side. One of the things that’s really fantastic about marketplaces is when they’re humming, you just don’t have a lot of costs because the system should be self-reinforcing so this is a different angle of attack.
Angela Tran: The more managed it is, it starts to look very e-commerce, and then you start to wonder if there are even network effects anymore. Then the question around defensibility, the moat that all comes up.
I did write about an “API-as-a-marketplace”. Just like a traditional marketplace, an API-as-a marketplace has two sides, suppliers and buyers. It’s just that the interface is different. In a traditional marketplace, discovery and purchasing is done on the application layer so the transactional process is visible to the end user and it’s essentially part of the platform or their marketplaces identity and brand. In an API-asa-marketplace, the API is the UX for the transaction, or in other words, you can think of it as the transaction occurring on the infrastructure layer and is abstracted away from the end user.
Andre Charoo: I do want Sarah to explain in the early days why we shouldn’t focus on growing transactions. You have a concept called “Minimum Viable Happiness”. Can you explain the trade off?
Sarah Tavel: I wrote a piece on my Medium called Hierarchy of Marketplaces that I’m going to be synthesizing here. The insight came from studying Postmates and in hindsight, what they had done wrong. I felt like there were so many companies, so many founders I’d meet with and they were chasing GMV. Everybody wants to come in to meet with a VC and have that GMV number going up into the right. You want to hit these milestones and a million of annualized GMV for your Series A or whatever it is, but if you are optimizing for GMV, I actually believe it leads you in the wrong direction.
In order to be able to tip a market, you’ve got to focus on a really constrained problem. DoorDash versus Postmates, Postmates started a year and a half before DoorDash. They both went after the San Francisco Bay area, but Postmates went chasing GMV. They went after San Francisco. They went after multiple verticals; it was cafes, restaurants, retailers — you could get an iPhone, a bicycle, your Chinese food, a coffee. That’s how you drive big short-term GMV numbers. What DoorDash did is that they went after the suburbs. There was no competition because it was a terrible place to focus but what happens when you do that is you’re actually able to get your flywheel spinning in a more efficient way; you’re able to really have your network effect come alive. Then you’re able to expand from that position of strength, that kind of white hot center that you’ve discovered. I think DoorDash was just incredibly effective and it was just a lesson to me of focus. If you get that right then growth will be a positive externality, but you’ve got to focus on getting that value proposition really strong.
You asked about this idea of what I call “minimum viable happiness”. If you know you want to focus on a white hot center, the way to tip a market when your network effect comes alive (think of it as your flywheel spinning really fast) is with the value that you provide; I kind of think of that as how happy you make your buyers and sellers when they come and you become the obvious place to go to. That’s when you start to see organic traffic come to you from the buyers and sellers. It’s because your value proposition, relative to any other substitute, is just getting stronger and stronger as you’re getting more and more of the market. That value proposition is what I think of as happiness.
How do you make sure that you’re making your buyers and sellers so much happier than any substitute? What I’m saying is you want to create an experience where people will buy or sell and return. They’re voting with their feet and you’ll see it in the net revenue retention over time of your cohorts, which is that you’ve found the white hot center where you’ve got enough of the things, the alchemy, that’s the kind of threshold that I think of as minimum viable happiness. There are ways to operationalize this. I could go on and on, but that’s basically the concept.
Andre Charoo: There’s a misconception around target market size. Are there markets too small to support a marketplace?
Sarah Tavel: It’s a good debate. I actually think that if you’re worried you’re focused on too small a market, you’re probably on the right path. What you have to make sure is that it’s not a dead end. There are plenty of scenarios where you can focus on a small market and you get stuck there. If you focus on a small market, if you can get that white hot center really right and not spread yourself too thin, really get it intensely right, then if you pull on the thread of the buyer need, does that pull you into a broader and broader supply side use case that lets you expand the market over time?
Pete Flint: I would agree a hundred percent with Sarah. While the market size is relevant, the market structure is potentially more relevant in the sense that what you’re trying to look for here is fragmentation. Yo
u’re looking for some sort of promiscuity in the marketplace; like restaurants, you want to look at different restaurants, you want to look at different houses, you want to look at different hotels. There’s sort of an element of promiscuity in the marketplace, either the supply side or demand side or both. Then there’s an element of repeat usage. You’re trying to come back and back again. If you start to see those components alongside, as Sarah said, this sort of attractive, underserved niche that can grow into a bigger market, then there’s real potential.
Angela Tran: We also like to look at the marketplace as just solving an efficiency problem, which is very valuable, or is it offering a new opportunity to create new transactions altogether? If we were to think about Uber, we have all these people who have never thought about driving their own cars or driving other people around before. They created a new supply, which creates new demand and a whole bunch of new transactions versus let’s say a substitute teacher marketplace or a nursing marketplace where you create this efficiency, but you’re not necessarily creating more nurses or more hospitals that might need nurses, etc. That’s another way in which we kind of evaluate market size.
PART TWO: Creating flywheels of success
Andre Charoo: Sarah, you did talk a little bit about loops. Can you explain your concept of tipping and happiness loops?
Sara Tavel: Angela and Pete talked a bunch about the hard chicken and egg problem, kick-starting a marketplace. Basically in the beginning of a marketplace, you’re creating this playbook of these things that don’t scale to get your flywheel spinning and to start to get transactions going in your marketplace. Then you make this transition where you start to develop a second playbook. I think of it as identifying these loops that hopefully naturally occur already, but are easy for you to add to the surface area of your product in some way that helps you grow systematically.
I think there are two different categories of loops that work together symbiotically. The first is growth loops; if you don’t know growth loops there’s a fountain of knowledge waiting for you on Google and it’s kind of classic supplier to supplier referrals, buyer to buyer referrals, there’s SEO; there’s so many different types of growth loops that you can take advantage of.
The second thing that I think about is what I call happiness loops, and it’s basically creating systems in place, in your product, so that as you grow, particularly as you grow your supply side, you don’t actually suffer from a degradation in the happiness of your buyers. That’s because you’re still figuring out a way to get rid of the bad suppliers and reward the good suppliers. Sometimes I say it’s like the kidneys of a marketplace. How do you figure out a way to have these naturally occurring systems that reward that good and get rid of the bad? Reputation is the classic example of that. When you have people review someone and that helps the buyers know with whom to work. Search ranking is another example of that; you’re creating these loops and it could be implicit or explicit data to help keep your marketplace really healthy as you grow.
Andre Charoo: Angela, in your Guide to Marketplaces, you mention the key to building defensible marketplaces is to aggregate demand. Can you talk more about this?
Angela Tran: We talked about starting with supply and as your marketplace becomes successful there is this risk that supply becomes a commodity; because as your marketplace scales, there’s naturally going to be copycats. The goal of supply is just to maximize distribution so there’s obviously loyalty, but the more channels in which suppliers can sell through — you can’t blame them for being on as many platforms as possible. With that said, we think the moat is really around the mindshare of demand and also that’s where money is situated right in demand’s pockets.
Andre Charoo: Pete, it would be remiss if we didn’t talk about network effects with you. What are some signs that give founders confidence that their marketplace is actually truly experiencing network effects?
Pete Flint: I think part of this is part art and part science, in the sense that it is kind of like the science piece is really accelerating trends and underlying engagement of the product; increasing metrics, increasing organic demand, increasing repeat usage. There’s the top scale metrics and then there’s the underlayer metrics, which is harder to engineer specifically, but if you’re engineering the product correctly and you have critical mass and a focus on density they will start to tip up.
Then there’s the element of art: you get people randomly Tweeting about stuff on social media or getting a fan-based kind of virality. You’re starting to see people love the product and start to share the product on supply sales on the demand side, which is hard to quantify, but it does happen. It doesn’t happen overnight, but when this is happening, something very interesting is happening.
PART THREE: How do you win?
Andre Charoo: Pete, in your opinion, how do you win? In your journey of building Trulia, you clearly had a big competitor with Zillow. You also talk about a concept of a “market network” in a path to winning and to building further defensibility. Can you share more?
Pete Flint: In reality, we merged with Zillow, so we all won. I think there’s a lot written about the kind of economic distribution in marketplaces that the number one will capture, say, 60% of the profits and number two 30%, and this is sort of a common theme across most geographic marketplaces. What is often interesting is that the smaller the market the less capacity there is to create competitors. So geographically, you see this across the world, that there are smaller regions. Google, the market share of Google in the US is surprisingly low relative to smaller countries, because they’re just not enough economic value just to create substitutes. So in some ways, these kind of niche markets, you can have outsized market share, which enables you to actually have outside profit potential which enables you to be successful.
Then, as you become more valued than you think, the UK version of Trulia is worth billions — who would have thought that that would be the case? In terms of winning, I think one thing to realize is that there is a significant premium for category leadership, but when there’s a large degree of supply side promiscuity or multi-tenanting, then you do start to see the ability for multiple competitors to do well.
Andre Charoo: Sarah, in your framework, you share how it’s actually not just about being number one. It’s about being number one by a lot. How should founders think about this?
Sarah Tavel: Early on in my career, I was an observer for this company called OLX, which was one of the early online classified sites.
They sourced this analysis that was super interesting to me where they’d actually done a two-by-two and looked at the profitability of the classified site that they had on one axis and on the other axis, if they were number one, how much bigger were they than the number two. It was this very strong correlation, which was that when one of their sites, in a given category or geography, was number one, 8x, it was basically like 80%, 90% contribution margin or profit margin on the classified side. If it was 1.5x and it was not profitable, it was kind of this dog fight where you’re constantly fighting for position.
The insight there is very clear. It’s back to the bigger you are, the more you penetrate the market relative to any other substitute, the more value you’re creating for your buyers and sellers, and therefore the more you can capture through your buyers and sellers. You don’t want to be in a place of Uber versus Lyft, or the dog fight that’s happened in the food delivery world, which has just taken an incredible amount of capital because there’s been no clear leaders. There’s no point of being a marketplace unless you are number one by a big margin.
Andre Charoo: I love that bold statement. I hope the founders are listening. Angela, we’ll end with you on the concept of identifying inventory and as that becomes a bit of a commodity, or less and less unique, as the marketplace gains popularity, are there strategies that position you to win right now more than others?
Angela Tran: I really like Sarah’s framework of it being the pursuit of happiness and constantly looking for ways to reduce friction in the transaction. If we think about protecting our supply again, that goes back to recognizing it and acknowledging great supply and perhaps having your reputation tied to the platform so that people don’t leave. I always think that fostering trust and credibility with buyers is a big driver of their mind share.
Andre Charoo: Are there any misconceptions around achieving dominance. According to Crunchbase, the average marketplace had to raise seven rounds of funding before going public. Is the only way to really win to raise boatloads of venture capital?
Sara Tavel: I think about subsidies as a crutch for happiness. It’s almost like you’re artificially raising the floor of where the experience actually is by using venture dollars to subsidize where things actually are. The problem is that when there’s a lot of competition you have to do that. It’s about that race to become number one. It’s part of the reason why I try not to announce my rounds, my investments, and I love my companies to be underestimated from the outside because I’d rather people think that they’re small crappy markets for as long as possible, then all see the market and think this is a great fit, but let’s build something here too, but for another geography.
Angela Tran: I also think it depends on the capital intensity of the marketplace itself. On one end of the spectrum, you have capital-light marketplaces where it’s much more product-based, whether it’s physical or digital, and has a global audience.
On the other end of the spectrum, there are these more capital intensive ones where it might be more of a managed service in a local geography where you have to invest in building network effects on a city by city basis. I think it’s somewhere along the spectrum. I think you can build an efficient marketplace, but again, intensity and also competition is something to always consider.
Pete Flint: Trulia, we raised $33 million before we went public, which is kind of like a small Series A round. We would’ve raised more.
The capital matches the opportunity and the opportunity, when the network effects kick in, is significant. The other thing is that capital can force inefficiency in an organization, which can be destructive. This year in particular, I think many companies have kind of revisited the business plan and focused on efficiency or pivoted their product to focus on profitability in the sort of depths of deep COVID in Q1 / Q2. It’s been remarkable. There’s a number of companies that have found product-market fit where they didn’t have it before and they’ve driven efficiency. Airbnb is a case example of that, where they pulled back on marketing and usage went up. I think that yes, capital does match the opportunity, but also the best marketplaces are incredibly capital efficient and figure out ways to grow whether that’s from hacks or utility, or just as we said, happiness.