Matchmakers: Summary and Review

Chapter 1: A Table for Four at Eight

Multi-sided platform businesses are difficult to get right, and many of them fail. This is no get-rich-quick scheme. Lots of people have tried using this business model, only to see their enterprise crash and burn. New businesses have a high rate of failure, whatever their philosophical underpinnings. There are no shortcuts for hard work and taking good advantage of circumstances.

Opentable is an online restaurant reservation service. The company is a good example to use for understanding multi-sided platforms. They faced an initial barrier because most restaurants didn’t use computers way back when they started around the turn of the millennium. They decided to lease equipment and software to restaurants rather than make the small businesses invest in pricey equipment.

Next, they faced a classic chicken-and-egg problem. They had trouble attracting restaurants and their customers to the website. Neither side saw any value in the site without the other side present. Restaurants didn’t want to use the service if they couldn’t reach any customers by doing so. Customers had no reason to use the site if there weren’t any restaurants participating. Opentable tackled this problem by narrowing their focus to four cities. They realized the only way they could achieve critical mass was by focusing in depth in a narrow geographical region. They created a standalone service that made the platform valuable to one side, the restaurants. It worked. They managed to attract enough restaurants in San Francisco that customers were drawn to the site, and things really took off. Then they were able to spread to more cities. They scaled the business by moving to new regions, one at a time.

Platform Competition in Two-Sided Markets by Rochet and Tirole (2000) was the paper that started a new way of looking at matchmaker businesses. These new insights have helped us understand that multi-sided platforms require different rules than traditional businesses use. For example, in the old way of doing things, it was bad business to give things away for free. But for platforms, freebies are good because the practice of giving things away can help grow a network.

The matchmaker business model is older than you’d think. In fact, platform businesses have been around for thousands of years, but they have often been misunderstood and mismanaged.

One-sided platforms include old fashioned stores. In this model, the store buys wholesale and the customer buys from the store. The customer and the wholesaler never interact. Two-sided platforms are venues where the two sides can interact and transact. Examples of multi-sided platforms include shopping malls, computer operating systems (which are intermediaries between app developers and users), magazines and quite a few websites.

Any channel that creates value by connecting different kinds of customers and clients is a matchmaker. Multi-sided businesses often have to use non-traditional, counterintuitive strategies to make money and thrive. One sign of a matchmaking business is when one side of a transaction gets a good deal, like free stuff or nice magazines for dirt cheap. The other side pays for participation. On Opentable, restaurants pay to be listed while customers can make reservations without incurring any additional cost.

Chapter 2: The “Grab All the Eyeballs” Fallacy

Fifteen years ago, the accepted wisdom was that internet businesses needed eyeballs most of all. High numbers of visitors should result in high numbers of sales. But this isn’t true. Network effects turn out to be way more complicated than all that.

The direct network effect is the key. The more people are on a network, the more that network is valuable for them; this is also known as positive direct network externality.

When videotape players were developed for the home market, two competing standards emerged: VHS and Betamax. At one time, it was difficult to predict which of these formats would capture market dominance. VHS, however, had the advantage of an early lead. Once one standard became more popular, more producers were creating content for it. This made it even more popular, eventually leaving poor Betamax in the dust. Sometimes superior technology loses out because of the network effect. There’s a big advantage in being first.

The reality isn’t this simple. Both VHS and Betamax worked to get content providers and consumers. VHS was more successful attracting them. They had a better two-sided strategy. The video player wars got economists to look at network effects which cast doubt on first mover advantage and winner takes all theories. These theories might be true sometimes, but they aren’t universal truths.

Network effects are more common than anyone realized. It turns out that most people use more than one platform. This is called multi-homing. For example, consumers might have more than one credit card, and vendors may accept more than one card. This obviously makes room for competition. User loyalty can take a hit if users are multi-homing.

Network effects are complex. They can be positive or negative, and the quantity of users is not the only important variable. There have to be enough desirable participants on both sides of an interaction. Indirect network effects are created when the value of a platform to one group is affected by how many of a different group participates.

Matchmaking firms have distinctive pricing strategies. Usually, traditional businesses don’t sell for less than cost, although there are some exceptions. This isn’t always the case with matchmakers. They have to balance (at least) two sides of transactions, and they have to keep the various parties happy. Demand is interdependent.

Free and negative prices are permanent features of such platforms and not gimmicks. This approach isn’t automatically the right strategy. It depends on the situation. The matchmaking model isn’t for some businesses, but it’s natural for others.

Chapter 3: Turbocharging

Technology drives the matchmaker revolution. It has allowed matchmakers to reduce costs, increase speed, improve quality and increase their reach. There are six technologies that have been the biggest drivers in platform growth:

Because of all this technology, matchmakers all over the world have access to huge populations and markets.

These technologies are continuing to evolve and change rapidly. They enable platforms that support other platforms, as Internet Service Providers (ISPs) and operating systems do. Operating systems are multi-sided platforms that provide a standardized environment where apps can be developed and users can frolic.

Mobile operating systems for phones and similar devices are platforms that host other platforms; they are known as foundational multi-sided platforms. Mobile operating systems and related devices are tearing down the wall between worlds.

There used to be two worlds: the virtual internet world and the real, brick-and-mortar world. These two worlds are starting to converge into one. Nowadays, you can be on the street and get run over by the Uber driver you just hailed on your mobile phone.

Revolution has an inherent destructive side. You’ve got to break a few eggs to make an omelet. Platform businesses have certainly disrupted many traditional industries. They are disruptive because they found a better way to do things. Evans describes the process as “creative destruction.”

Chapter 4: Friction Fighters

Platforms create value by reducing friction. The better they reduce friction, the more value they provide, the greater their success. But though reduction of friction is necessary for platform businesses, it’s no guarantee of success.

In order to effect transactions, every side must be satisfied. The customer will not buy if the price is too high, nor the seller sell if it’s too low. Everyone needs to get a slice of the value pie. Multi-sided businesses have to slice the pie up among more parties. It’s important to have a value pie that’s big enough for everyone to share. It must be worthwhile for all the parties to participate.

One way to increase the value pie is to reduce friction. Alibaba reduced the frictions people experienced when dealing with Chinese companies. Trust was a big problem. Not all companies were on the up-and-up. People had no way to know whether they should trust Chinese companies to honor their agreements. Alibaba developed International TrustPass which authenticated merchants. They ended up devising a number of other methods as well in order to improve the trust level between Chinese companies and their global customers. Alibaba succeeded because they addressed and reduced key frictions.

As the Chinese economy exploded in the aughts, traditional businesses struggled to keep up with demand. There were barriers to creating more brick-and-mortar stores. Online stores weren’t trusted for transactions. Then Alibaba invented Alipay. In this system, the consumer pays at the point of purchase. Alipay doesn’t release the funds to the producer until they either get feedback from the purchaser about the transaction, or until enough time has passed without complaint to assume the customer was satisfied. This, coupled with innovations in merchandise delivery greatly improved trust on both sides of the transaction. Friction was reduced.

This comes with a common-sense caveat. Just because it worked for Alibaba doesn’t mean that it is magic and it will work for everyone just by waving a wand. For example, lots of B2B platforms crashed and burned in 2001 dotcom bust.

The reason these businesses failed is that they weren’t fixing big problems for their constituents. The frictions that existed in China weren’t there in the West. In the U.S., there were many good networks to use offline. People could use faxes, go to trade shows, etc. The friction the businesses were trying to reduce weren’t causing big problems to begin with. In the U.S., frictions are more about bureaucracy: keeping records, invoicing, etc. Many platforms that reduced the right frictions survived the 2001 dotcom bust and went on to continued success.

Chapter 5: Ignite or Fizzle

Securing critical mass is a classic chicken-and-egg problem. A very hard problem. Both sides are needed in order for a platform to start, but users won’t want to join until there’s something there. No one, for example, would sign up on a dating site that didn’t have other people on it. There have to be enough users for users to want to join.

YouTube tried a number of different approaches to get beyond this phase, including creating their own videos. The thing that really helped YouTube take off was when they added the feature that allowed people to embed links to videos in other platforms. They continued adding all kinds of good features. It took a while before they decided to monetize YouTube. They worked hard at increasing users on both sides of the market first.

The critical mass frontier model demonstrates how critical mass can be achieved. Imagine, if you will, your standard Cartesian XY graph. In this case, the axes are labeled Type A Customers and Type B Customers. There is a curve showing how many of both types of customers are needed to reach critical mass. This curve is the Critical Mass Frontier. On the left of and beneath the curve is the Implosion Zone (too few users). On the right and above is the Growth Zone.

There are three main types of ignition strategies to combat the chicken-and-egg problem:

Tactics to complement these strategies:

It’s crucial to start with the right model, as it’s a decision with long-term repercussions. Platforms should stick with the pricing strategy they used when they were reaching critical mass.

Chapter 6: Long Haul

It’s a challenge to decide how to set prices for platform products and services. It’s a balancing act to determine the cost for both sides of a transaction. The pricing structure of a platform is determined by which sides are more sensitive to prices and which sides have more power over the interaction. The more valuable sides get a better deal. In the case of trucking fleets and gas stations, the trucking fleet has almost complete power over the transaction. The fleet can decide if the transaction is going to take place, so the pricing structure is largely in their favor.

When a transaction has exactly one buyer and one seller with no middlemen, single-sided pricing is used. The price of a product is the result of a combination of production cost and what the market will bear.

Multi-platform pricing is much more complex, as there is the question of what price each player in the transaction will pay. It’s a real balancing act. The price also has to be adjusted to maximize network effects. It becomes a matter of striking a relative price between the parties.

Many platforms have a subsidy side and a money side. The less valuable side incentivizes the more valuable side. One side pays for the other side to play. For example, on employment websites, usually prospective employees can post their resumes for free, while prospective employers must pay to list employment ads.

When a platform becomes vast and desirable, it’s time to monetize it. Firms may decide to do this by charging usage or access fees. Fees should be placed on the group that offers the least value to indirect network effects. Try to get them coming and going; require participants to pay to get into the system and to pay to do the deal.

Never, ever lose sight of the price sensitivity of each side. Which side needs what, and how much do they need it? Which group determines if an interaction will take place at all? Keep monitoring costs over time. Adjust and rebalance as appropriate.

In a competitive environment, you can just decide to copy what everyone else is doing when setting your rates. You are unlikely to get people to pay something extra when they can get it cheaper with your competitor. Save yourself the mental energy and copy successful platforms.

Chapter 7: Beyond the Castle Walls

Platforms require healthy ecosystems. To succeed, multi-sided platforms need to create a healthy external ecosystem. The value of a platform is directly tied to the ecosystem. An ecosystem consists of all the entities that positively or negatively affect platform value. Due to the importance of ecosystems, platforms have much less control over the customer experience than single-sided firms. With that loss of power comes benefits, however. With the proper incentives, platforms can control partners and have ecosystems provide a whole host of services that would normally exist inside the firm.

Some companies start as single sided for a while to get up and started. Not all of them can do this, however. Some companies find ways to mix elements of one-sided and two-sided businesses. Microsoft, for example, owns Windows as a platform, but it also develops its own apps that compete against other developers on the platform. Amazon is similar, but it acts as a hybrid reseller platform. One side connects buyers and sellers, while the other side does resale.

Pioneers — those starting platforms so unique they can’t just copy what the last guy did — have to decide right from the get-go how many sides they want. They should figure out the order in which they’ll roll out their platform to each side. They don’t have to nail down every detail, but they should have some idea of the general framework of their platform structure.

The choice of platform sides is complicated. Adding sides can make for more network effects. Adding sides can also make for a messy business model with unneeded complexity and noise. There can be too many interests to appease them all.

It’s important to nurture a healthy infrastructure.

Apple’s iOS and Google’s Android mobile technology both illustrate the choice between multi-sided platforms very well. Because they wanted to have strict control over quality, Apple made the iPhone so it would only run proprietary apps. They kept absolutely everything in-house that they could, essentially turning it into a single-sided platform. Google went the other way. They had many partners when they developed Android. They used a certification program to keep Android and its apps standardized. Users wanted third-party apps. Apple saw the way the wind blew and eventually opened up to app developers. Both companies ended up creating thriving ecosystems around app development, although Apple ultimately saw greater success due to different internal structure.

Chapter 8: Interior Design

Southdale Center was the first modern indoor shopping mall. Located in Edina, Minnesota, it opened in 1956. Since then, mall developers have continued to learn how to maximize value to a mall’s stores, customers and owners.

In an auction, you want lots of buyers to drive up the price. This is an example of a thick market. Good platforms have thick markets. eBay has benefitted greatly from a thick market created through network effects. Stock markets can’t always achieve a thick market for each and every equity, so they rely on market makers to adjust the market on either side as needed.

Platforms don’t need to be as large as eBay or the stock exchange to be successful. Platform size can be large or small. Many platforms use screening devices to connect to a small market, making their platform more or less exclusive. This creates a thick market for a small group. Creating a specialized ecosystem is a good way to target specific markets. Some platforms also want to control size to limit competition and congestion.

Platforms should also be designed in a way that facilitate good matches. One way they do this is by designing standards that help relevant parties interact. Everyone should know the rules and have the same expectations. Examples include signaling devices like VISA signs at points of sale, search algorithms and search tools.

Platforms should also be designed to make matches with people that don’t want to be matched, but it still manages to make them happy. For example, media businesses give users good content at or below cost. In return for this cheap content, users must deal with a barrage of ads, but it is usually still worth it.

The shopping mall layout is a good example of platform design. Malls have anchor tenants that each attract their own customers. This increases foot traffic and creates a thick market. They have a bunch of non-anchor stores, too. Customers have a wide selection of shops to peruse. There are shopping malls geared toward bargain shoppers and those aiming for the high end of the market. The types of stores signal to the desired audience: come shop. The physical layout of the mall also encourages shopping. Stores are spread out across the mall and escalators are placed in odd locations to increase walking and create more interactions.

Chapter 9: Fakesters and Fraudsters

Facebook regulates user behavior. They require users to identify themselves by their real names. Pornography and bullying are not tolerated. Diligence in pursuing these policies has resulted in a civil social environment. People feel safe interacting with each other on the site. Facebook has likely prevented the site from being overrun by trolls by maintaining a certain level of standards.

If people aren’t rewarded for providing a benefit or if they aren’t punished for causing harm, it results in what economists call an externality. An externality is a consequence that affects third parties, whether for good or for bad. An example of an externality is the pollution from a car’s exhaust. The pollution is an undesired side product, in this case one that might negatively affect the health of the people in the neighborhood the car drives through. Negative externalities are magnified by network effects. Negative externalities can overwhelm a platform and discourage users, starting a death spiral that will doom the platform.

People can be bad. People can be opportunistic. People can be very naughty indeed. When they act like this and engage in antisocial behavior, it reduces the value of the platform. There is strong financial incentive to police bad behavior. Negative behavioral externalities reduce the platform’s value and they push good participants away.

Communities need to have rules, and someone needs to enforce them. Platforms must use governance to nurture positive behavioral externalities and minimize the negative ones.
There are many ways to defend against negative externalities. Platforms are able to prevent low quality entrants from joining. This was the reason Steve Jobs wanted to keep app development closed to outsiders. Platforms can be structurally designed for high standards of social behavior. When problems do arise, they can be addressed swiftly. Unlike governments, who must go through bureaucratic walls, platforms can act very quickly when dealing with issues.

It is crucial to stay on top of this. Sometimes it can take a lot of effort. There are tools that help. Negative consequences for bad behavior reduces bad behavior. Rating systems are useful; other users might not want to deal with a user who has a low score. Ratings can encourage people to behave and follow the rules and act civil. Exile is another useful tool; tossing miscreants out of an environment and banning them from the platform. Platforms should not tolerate anti-social behavior.

Chapter 10: Fizzle or Sizzle

Pioneers can’t follow someone else’s blueprint to success. They have no precedents to follow. There is, however, a checklist:

Knowing how to predict whether a platform will succeed or fail is useful for pioneers of new platforms and for investors.

Platform startups exist on borrowed time. If a platform grows slowly and can’t reach critical mass in a few years, it is likely doomed. Generally, platforms should succeed within the first few years of their existence, if they’re ever going to be a success. When a platform can’t get any traction with anyone at all, that’s a bad sign. It’s time to take action.

What if growth is not fast enough? This must be monitored closely. At some point, there should be a steep acceleration in growth. Other signs a platform is in trouble is if it can’t get the even most adventurous early adopters, or if it fails to get any marquee customers.

Platforms die either when most of the participants leave or when investment funding dries up.
The history of the mobile payment system Apple Pay is used to illustrate the ideas introduced in this book. Apple Pay has yet to take off, no doubt in part because it’s only available to owners of the iPhone 6. When they launched, few vendors had the chip reading technology required to use the best features of Apple Pay. This was a classic chicken-and-egg problem. Apple Pay tried to go broad and shallow and cover the entire company. This resulted in many thin and fragmented markets.

Conversely, Starbucks has a phone app. In 2011, they implemented a mobile app to use in Starbucks stores. The stores were equipped with the new readers in plenty of time for launch. The app was intuitive and easy to use. Starbucks has demonstrated that under the right conditions, mobile payment apps can succeed.

Apple Pay could still work out its problems. They have unusually ample resources to weather the difficult early years as the environment comes together.

Chapter 11: Moving Money

Mobile phones have revolutionized money movement in Kenya, making it much easier for most people to send cash. They have helped to relieve some amount of stress for people who faced barriers to banking services. The creation of M-PESA is a good example of how platforms can succeed despite facing many challenges.

Many Kenyans had been moving away from their families for work. They sent money back home through a money transfer system that moved slowly and was filled with trust problems. This was the source of a lot of friction. Vodafone aimed to address this friction by introducing a system of money transfer that involved SMS messaging. They noticed that many Kenyans already had cellphones, so they could develop something with little additional technological requirements. Consumers already possessed the technology that they needed to participate in the service.

The most popular cell phone service provider was a mobile operator named Safaricom. Vodafone worked with Safaricom to develop a money platform for phones. They called it M-PESA. This solved big problems for people who couldn’t afford bank accounts. They could pay their bills. They could get birthday money from Uncle Charlie who lives in a distant village. They could finally engage in normal modern commerce.

It was hard to get M-PESA started because of chicken-and-egg problems. Vodafone had to get enough users on both sides of the transactions to make the system work. They had to recruit stores to be their agents and sign people up for the service.

They made it cheaper to send money to others who were already signed up for the service. This encouraged money senders to convince the money receivers to join the service. They leveraged Safaricom’s existing network of retail dealers. When Vodafone launched their new platform, they had people on all sides of the platform already on board. They also did some traditional advertising. They were successful. Over time, they added more services. All these strategies helped M-PESA expand into a multi-sided business (including financial services and merchant platforms) and eliminated huge frictions in an impoverished country. Studies show that M-PESA increases household liquidity and helps Kenyan families become more economically resilient.

Mobile money platforms have helped in other African countries, and there are still more people out there in other places who could be helped with this service.

Chapter 12: Gone Missing

Platforms are transforming the world of retail. The internet and smartphones have had big impacts. Despite a slight increase in spending at brick and mortar stores, many shops are shutting down or greatly cutting back on space. People do still leave the house to shop sometimes, but when they do they don’t walk around as much as they used to and they visit fewer shops. Shoppers are doing research and browsing on online platforms, leading to much less time spent in-store. Traditional retail depends on customers spending time moseying around and making impulse purchases.

Platforms disrupted physical stores in three waves:

Traditional businesses can learn the new way or prepare for failure. Change is inevitable; the choice is to move with it or be swept away.

It’s a steep curve. Everything is changing at once: a new way of doing things, a completely different pricing model, unfamiliar and non-replicable network effects and new competitors who aren’t limited by geographic region. If a company can’t compete in the new environment, they must give serious thought to shutting down shop, liquidating their assets and going out of business.

Brick and mortar isn’t going away. People still like to try on clothes and look at things before they buy them. Even millennials have a strong preference for retails stores when it comes to checking out the merchandise. Successful retailers have brick-and-mortar as one of the sides of their platform.

In order to survive in the face of this disruption, existing firms have three choices:

Blockbuster failed to do anything about this transition in retail. They had the opportunity to partner with Netflix before it launched, but Blockbuster rejected their idea. They had the capacity and connections necessary to succeed. They could have done well as a video streaming service, but instead they denied the revolution and the rest is history. Blockbuster tanked.

Change goes both ways. Just as some stores have their own websites and stores online, so have some online retailers moved to brick-and-mortar stores.

These changes are still happening in the real world. It will take a long time to see the ultimate results platform businesses will have in the world.

Chapter 13: Slower and Faster Than You Think

The sharing economy as typified by Airbnb is all about multi-sided platforms.

Matchmaking — making markets — has been going on forever. Everything old is new again. What’s different now is that there is fancy technology. IT has “turbocharged” matchmaking businesses, making them more efficient than ever before.

Have matchmakers reached their final form or will they evolve further? History suggests nothing really ends, and new technology does not bring apocalypse. Rapid change, certainly, but not apocalypse. The matchmakers that exist today won’t be the last ones to be significant.

Turbocharged matchmakers will transform industries in waves as matchmakers rapidly expand and replace traditional firms. Matchmakers are becoming increasingly central to global economies. They will continue to bring big change.

To reiterate, many of today’s matchmakers are just combining old ideas with new technology. The telegraph was a messaging service, village matchmakers used to make dates, and the sharing economy has been around forever. Now we have WhatsApp, eHarmony and Airbnb instead.

Indirect network effects drive multi-sided platforms. Technological advances magnify those network effects, which is why turbocharged matchmakers have the power to disrupt so many industries. Because we’re in the thick of it, our view of these big changes in industry and in the economy is distorted. Actually, it’s not as big of a deal as you might think. The development of the telegraph, or that of the ancient agora, were probably more significant in the long run than the significance of the internet.

This is not the end of history. Don’t for a moment presume that all the cool stuff has already been invented.

Even though change sometimes comes at breathtaking speed, there is precedence for everything going on right now. Technological breakthroughs of the past have had similar impacts on their societies. The patterns of change and disruption are roughly similar.

Despite all these matchmakers, there is still a lot of room for new platforms. Anyone looking to be a part of this brave new world should understand how platforms grow and function.

Multi-sided platforms are important players in today’s economy, and their influence is only going to continue to grow. They have become vital in the developed world. They help drive progress in underdeveloped economies. Like other innovations before them, matchmaking multi-sided platforms are causing vast amounts of creative destruction. Everything is changing, and will continue to change into the foreseeable future.