My views on the market, tech, and everything else

A Dozen Things I’ve Learned about Multi-sided Markets (Platforms)

  1. Multi-sided markets bring together two or more interdependent groups who need each other in some way. Uber, eBay, and Airbnb are all multi-sided markets. A multi-sided market is sometimes called a “platform.” Hundreds of big and small firms fail trying to create multi-sided markets in different categories every year. Since the payoff from creating a significant multi-sided markets is so massive and the financial downside relatively small, founders and venture capital investors are willing to keep experimenting since it is magnitude of correctness and not frequency of correctness that determines financial success. Just one success in creating a multi-sided market can pay for many dozens of failed attempts. Apple, Microsoft, Google, Amazon and Facebook are all platforms. Why don’t firms create more multi-sided markets if they are so profitable? Because getting all of the elements just right at just the right time is very hard to do successfully. Creating a successful multi-sided market requires a lot of skill and luck as will be explained below. 
  1. A critical difference between single and multi-sided market is that the sides interact directly. A single-sided “market maker” buys x, puts x in inventory, and eventually sells x. As an example, a person employee who buys breakfast cereal from a grocery does not deal directly with the manufacturer of the product so that value chain represents a single-sided market. The authors of the book Platform Revolution refer to single sided markets as having a linear value chain. They describe a single-sided market as “a step-by step arrangement for creating and transferring value with producers at one end and consumers at the other.” In contrast, when a recruiter contacts a potential employee directly using the LinkedIn platform that interaction is part of a multi-sided market. Multi-sided markets look like this:


3. Multi-sided markets are not linear. A market that has two or more sides is vastly more financially attractive to create since it has the potential to scale and generate value in a non-linear manner. In other words, the whole of the value created by a multi-sided market can be more than the sum of the parts, if the multi-sided market is correctly structured. What can happen if the right conditions and structure is present is described well by Esko Kilpi as follows: “Network effect-based value can increase exponentially at the same time as costs grow linearly, if at all. If you follow the valuations of firms today, there is an ever-widening gap between the network-economy platforms and incumbents driven by traditional asset leverage models. Investors and markets have voted.” The nature of platforms has changed the value chain in many industries. Tom Goodwin of Havas Media famously said: “Uber owns no vehicles. Facebook creates no content. Alibaba has no inventory. And Airbnb owns no real estate.” When you do not need to own these assets less capital is required and the business scales vastly better both financially and operationally.

  1. A multi-sided market is not valuable if the sides can find each other easily. Solving a hard coordination problem is the key to sustainability for any market, including multi-sided markets. How do many firms and individuals involved in a complex division of labor successfully coordinate their actions so as to optimize the creation of value and profitability, when each possesses different and changing knowledge and expectations about a risky and uncertain future? Coordination is achieved through price discovery and a multi-sided market can provide just that for market participants if the conditions are right. For example, someone may need to find a freelance artist or software programmer and since there are so many potential suppliers with different skills and availability a platform like ProFinder or UpWork will therefore often be useful to them. But sometimes a significant coordination problem does not exist and a multi-sided market is unlikely to succeed. For example, if Boeing needs a supplier of rivets, once it finds that supplier it no longer needs a multi-sided marketplace. Airbnb and Uber are example of businesses that do solve hard coordination problems.
  1. “Demand side economies of scale” (also known as network effects) result when the value of a product or service changes in a positive way as more people use it. Network effects can be positive (for example the benefits of an increasing number of members of a social network) or negative (for example, network congestion; viruses, spam).  What a business seeks in any multi-sided market are network effects since they are increasingly the most important method that can be used by a technology business to create barriers to entry against competitors (a moat). Due to the falling costs of creating a new business, especially technology businesses, network effects are the most significant way to a company to create moat and generate a profit. The famous napkin sketch that illustrates Uber’s network effects is as follows:


Having a moat is essential for any business since otherwise competitors will introduce additional supply to a point where financial returns are equal to the opportunity cost of capital. Anyone who says a moat is not needed is essentially suspending the laws of supply and demand. If a business earns a financial return that exceeds its opportunity cost of capital for a significant period of time it has a moat. The only question open at that point is what the elements are that created that moat (network effects, brand, supply side scale economies, regulation, intellectual property, etc.).

  1.  “Supply side economies of scale” exist when there are reductions in the average cost per unit associated with increasing the scale of production. Sometimes you hear people say that supply-side scale economies of scale no longer matter in an age where network effects are so powerful. This is not true since supply-side scale economies can in some cases still enhance the moat of a business that also enjoys network effect benefits. A moat can never be too strong since it is always under attack by competitors.  As an example, the leading providers of cloud infrastructure  have supply side economies of scale in building web scale services that smaller companies will not be able to match.
  1. Creating a successful multi-sided market requires that the business overcome the “chicken and egg problem.”  The problem can be stated simply: How do you get one side to be interested in a platform until that other side exists, and vice versa. Part of the challenge is to get enough customers on both sides so there is critical mass.  Critical mass is tricky to obtain particularly if the two sides need to show up simultaneously.
  1. The chicken and egg problem is best overcome if one side is clearly made the loss leader. Experience has shown that it does not matter which side gets the free or subsidized offering- what is important is that one or more sides be chosen as the loss leader and one or more sides as the profit pool. The subsidy side tends to: (1) have more elastic demand; (2) be an offering that is harder to get; and (3) is needed more by the other side. The profit generating side tends to be the reverse, obviously. The price of the subsidy side is often below marginal cost and in some cases less than $ zero. For Uber, the subsidy side is drivers since they are relatively scarce. In the real estate business in the United States, buyers are the subsidy side.
  1. The sides of the market should complement each other – if the sides are complements, it not only reduces the customer acquisition cost (CAC) but assembles sides that want to to enter into exchanges.  A “complement” is any product or service that increases the value of another product or service.  Multi-Sided Markets work best when the offering on one side of a market complements an offering on the other side of the market. Some complements are “demand-side complements” like cars and oil, hotdogs and mustard or chips and salsa. Some complements are supply side complements, like beef and leather. Search and advertising are complements as are social networks and advertising.
  1.  Subsidizing the side of the multi-sided market with lower marginal cost/COGS is optimal. Giving away goods that have a high marginal costs is often deadly. The objective is to have zero or very low marginal cost. As an example, it is particularly financially attractive to give away software as your incentive since it has zero marginal cost after fixed costs are recovered. Giving away subsidized hardware or storage is far less attractive, since there are always additional marginal costs no matter the scale of the business.
  1. Businesses that are slow to get to critical mass can run out of cash and momentum. This problem is hard to solve for the same reason the opportunity exists. Getting the balance right and achieving critical mass is an art and not a science. People creating multi-sided markets is relatively is rare at scale. In technology business today there are few areas that do not have existing competition. If you go too slow you may fail just as you can fail if you go too fast. And as I wrote in my recent post on Bill Gurley, you may need to deal with competitors who are not acting rationally or at least are acting extremely aggressively. You need to play the game that is on the field.
  1. Clear profit pools should exist.  Products and services at a given business cannot all be loss leaders. The adoption of a multi-sided market strategy must be considered holistically. If something is a loss leader then there must be a leader that is profitable. Providing one side of a multi-sided maker at a loss cost is a tactic intended to optimize the customer acquisition process and lifetime value for a service that is profitable overall, rather than a standalone strategy.


Invisible Engines: How Software Platforms Drive Innovation and Transform Industries. By David S. Evans, Andrei Hagiu, Richard Schmalensee.   https://www.25iqbooks.com/books/314-invisible-engines-how-software-platforms-drive-innovation-and-transform-industries-mit-press

Platform Revolution: How Networked Markets Are Transforming the Economy–And How to Make Them Work for You.  By Sangeet Paul Choudary and Marshall W. Van Alstyne. https://www.25iqbooks.com/books/315-platform-revolution-how-networked-markets-are-transforming-the-economy-and-how-to-make-them-work-for-you

Matchmakers: The New Economics of Multi-sided Platforms Hardcover – May 24, 2016 by David S. Evans (Author), Richard Schmalensee   https://www.25iqbooks.com/books/133-matchmakers-the-new-economics-of-multisided-platforms

Sangeet Paul Choudary http://platformed.info/twitter-whatsapp-uber-airbnb-network-effects/

David Evans papers: http://www.marketplatforms.com/wp-content/uploads/Downloads/Platform-Economics-Essays-on-Multi-Sided-Businesses.pdf

Van Alstyne: http://ebusiness.mit.edu/research/papers/232_VanAlstyne_NW_as_Platform.pdf

Haigu paper: http://www.hbs.edu/faculty/Publication%20Files/15-037_cb5afe51-6150-4be9-ace2-39c6a8ace6d4.pdf

Hagui interview:  http://hbswk.hbs.edu/item/new-research-explores-multi-sided-markets

Booz: http://www.strategy-business.com/article/03301?gko=16442

Esko Kilpi:  https://workfutures.io/one-theory-to-rule-them-all-c942486e4b30#.gj4wcbxtz

Sam Ghosh: https://www.linkedin.com/pulse/understanding-multi-sided-platforms-social-networks-more-sam-ghosh

Categories: Uncategorized