Platform Scale: How an emerging business model helps startups build large empires with minimum investment

progress bar

Platform Scale: Summary and Review

Keywords: Design, Infrastructure, Interaction, Monetization, Network effect, Platform, Startup, Scale, Transaction, Virality

Please Note: There are links to other reviews, summaries and resources at the end of this post.

Book Review

The information revolution has brought about the rise of a new business model: platforms. A platform business is rather like a base that has other things, like applications and products, layered on top. The platform itself doesn’t generate much value, the stuff on top does. Examples of platforms include YouTube, a platform with videos on it, and Amazon, a platform with merchandise on it.

Banking, education, healthcare — it’s hard to think of an industry that hasn’t been disrupted by platform businesses. As platforms emerge, they transform everything they touch. Today’s professional leaders must understand this new model or risk being left behind.

Platform Scale looks at the complexities of platform theory and explains how to use this knowledge to create and manage platform businesses. Instructions are provided for platform design, starting with the core value unit and build out from there. It’s loaded with examples taken from prominent new economy companies, such as Amazon, Reddit and many other businesses. There is a companion website: www.platformed.info. (Type carefully, platform.info will lead you elsewhere.)

This book is quite readable. Author Sangeet Paul Choudary moves easily between theory and practical guidance. Pragmatic in orientation, there is plenty of good advice for the designer. And although it was written for the platform designer, it avoids the pitfall of excessive jargon. For the most part, the ordinary laymen would have no trouble following this material. The glaring exception to this is perhaps so obvious that it escaped Choudary’s attention. Although the idea of the platform is well defined, that of scale is not. People with economic or business backgrounds will have no trouble understanding that scale relates to size and that to scale a company means to grow it. The layman doesn’t always know that which seems obvious to the professional, however. One omission with respect to the book’s organization: the lack of an index. Platform designers using this book as a reference will surely want to look up specific issues as they encounter them, but they will be unable to do so.

Within a year of publishing this work, Choudary went on to publish Platform Revolution. Curiously, Revolution is nearly identical to Platform Scale. Scalehas a slightly more practical orientation than Revolution. Most of the real life examples of platform companies that appear in Scale also show up in Revolution, although the latter includes additional examples that don’t appear in Scale. Revolution credits two additional authors besides Choudary: Geoffrey Parker and Marshall Van Alstyne.

The books are so nearly identical that all but the most devoted readers will probably not want to take the time to read them both. Perhaps the most germane factor to consider in deciding which of the two books to read is learning style. Those who can learn a lot by reading examples of how others built platform businesses will want to read Revolution; people who are bored by such narratives will find fewer of them in Scale.

Summary

Chapter 1: an Introduction to Interaction-first Businesses

The most popular business model used to be pipes, conduits through which companies produced goods and consumers consumed them. Value was created upstream and flowed in one direction: downstream. The model becoming popular in today’s business world is the platform. Platforms create little value themselves, instead providing space for producers and consumers to connect and interact in a manner that was previously impossible. The users of the platform create value.

Platforms represent a fundamental change in the design of businesses. Increasing connectedness, decentralized production and artificial intelligence is driving the rise of this new design.

Platforms cause three primary shifts in the marketplace:

Participants on a platform include customers and producers. For example, Amazon users include buyers and sellers; Uber users include drivers as well as passengers.
There is also a shift in competitive advantage from resources to ecosystems. Previously, big players were the ones who dominated resources; they had lots of employees and lots of physical assets. Platforms create ecosystems of consumers and producers, and successful platforms excel at creating and cultivating this ecosystem.
The third shift is in value creation, from processes to interactions. It’s the orchestration of interaction that creates value on a platform.

Platform scale is powered by a company’s ability to leverage and orchestrate a global ecosystem of producers and consumers, efficiently facilitating exchange and so creating value.

The old rules of pipeline business models no longer apply. The ecosystem is the new warehouse and the new supply chain. Network drives scale — the more people using a platform, the more valuable it is.

The platform demands a change in mindset:

Data is the new dollar.
Supply and demand both must exist at levels that satisfy producers and consumers.
Curation and reputation are the new quality control.
User journeys are the new sales funnels.
Distribution is the new destination.
Behavior design is the new loyalty program.

Chapter 2: Designing the Interaction-first Platform

Technology alone doesn’t drive value. Some platforms with high value have simple technology. Value, rather, comes from the ecosystem and, perhaps more so, the interactions on the platform.

The platform itself doesn’t generate much value. It’s an underlying base on top of which other things — like applications — are layered. Instead of thinking of the platform as the infrastructure, think of it as a network. Information is shuttled between users (producers and consumers). The platform needs data to make recommendations and facilitate transactions.

Units of value are created in these interactions between producers and consumers. For a platform to be successful, these units need to be created constantly. There needs to be awareness of what kind of supply should exist so that there can be interactions.

In The Nature of Order, Christopher Alexander talks about architecture and building outward from the center. This is also true with platforms. Identifying the center and building out reduces the likelihood of clunky features that don’t fit well. The core value unit is the minimum standalone unit of value that is created on the platform. eBay would be nothing without sales listings; Instagram would be meaningless without pictures. These features constitute core value units.

The core value must be considered in the context of the platform stack, the components of determine how the platform functions. There are three primary patterns:

Network dominated — If there’s a marketplace, the advertisements listing the goods for sale are the core value unit. If it’s a marketplace for standard services, the listings for the service are the value unit. If it’s a non-standard service (in other words, something that can’t be bought off the rack), the listings for the provider make up the value unit.
Infrastructure dominated — Applications form the core value unit on a development platform. The smallest unit of content are value units on content platforms — a video, for example, on YouTube. Sometimes these units can be created by more than one producer, as happens on Wikipedia. In such cases, the creators’ experiences must be considered.
Data dominated — Whether a Nest thermostat or a retail loyalty program, the value unit for data dominated platforms is data — the information that they traffic in.

Value units are needed to create interactions. Platform design should be all about creating value: the more value units or the higher the value of the units, the greater the platform value. Value differs, though, by user type. To a producer, a platform is a place to create or store value and a marketplace to find customers. To a consumer, a platform filters out the most relevant value units for them.

To build a platform business, start with the core value unit and build the core interaction around the core value unit. Then design the process — for example, a process might include producers posting ads, consumers viewing ads, making purchases and PayPal processing the money. In this example, the advertisement is the core value unit. While a platform might have more than one kind of value unit or interaction, there is one that’s core to the business.

Design the platform around this core interaction, and evaluate all added features and functions through this lens. If there are different kinds of interactions on a platform they must each be designed one at a time, working out from the center.

The three things you need for the interaction are producer, consumer and platform. Interactions always involve an exchange of information, goods and/or services and currency, and filters are central to the transaction. Abundance leads to too much choice, and people could abandon the platform if it’s too hard to find what they’re looking for.

A good platform is stretchy enough so users can take it in unexpected directions. There is randomness. There is luck. Platforms need to allow for emergent behavior. It can sometimes lead the platform into a new direction.

Chapter 3: Building Interaction-first Platforms

There are a variety of different ways to drive interactions on a platform: pre-existing relationships, content creation and consumption, coordination, competition, culture, etc. Some users drive interaction through clout. Platforms usually have more than one driver, and different drivers can have very different results.

To design a platform, it’s critical to understand what motivates producers. Access to the market is very important, so tools that simplify production are useful features. There should also be good curation to bring the best to the top, with clear rules so everyone has access to top content. Non-monetary incentives are important — things like reputation and recognition. It’s good to have mechanisms to convert consumers to producers and for consumers to provide feedback to producers.

Cumulative value makes a platform more valuable the more you use it. There are four kinds:

Reputation.
Influence.
Collections. Producers can have portfolios of their work.
Learning filters. Makes the platform more customized to the individual.

Friction impacts interactions. It’s bad when it interferes with the smooth functioning of interactions. It can be good when it curtails undesirable behavior. Friction can be a source of quality. It can be a source of signaling, (for example, LinkedIn has a lengthy process to set up a profile, but then it results in very detailed information.) Friction can also be a barrier; for example, sometimes people don’t want to or can’t pay membership fees. As ever, it’s all about the interactions. Evaluate how friction impacts interactions.

Curating provides evidence for new consumers to base their decisions on and allows the best to rise to the top. Social curation can be accomplished through upvoting, reviewing, etc. It’s good to let people try something and then express their opinion, it gets them in the door. Once opinions are expressed, social curation begins.

Platforms should have some form of quality ranking and allow for feedback for the producer. The more expensive it is to sample some form of content or service, the more time it will take to achieve a scale where social curation works well. That’s why social curation tends to be more inefficient on platforms with higher sampling costs.

Interactions require trust: Knowing someone’s identity increases trust. Involvement by a moderator builds trust. So does community feedback, codified behavior, culture, completeness of information and insurance.

Interaction failure means that, for whatever reason, an interaction didn’t result in a transaction. A grand example of interaction failure came about through the competition between Uber and Lyft. Uber played dirty. They hired people to use disposable phones, call Lyft and order rides and then later cancel the rides. All these cancellations mucked up the system and users left the platform. Interaction failures should be avoided, but when they do happen they need to be tracked.

It’s important to keep the interaction on the platform. Some services need to be discussed in person. The risk is that the users will circumnavigate the platform and do their transaction without the platform getting its cut. There are ways to provide incentives to keep transactions on the platform. The platform can charge a fee to place an ad, or require a subscription. Adding features, (like invoicing services, for example,) can make it worth users’ time to stay on the platform. Workflow management tools are particularly called out, but really any administrative task could potentially add value.

Chapter 4: Solving Chicken-and-egg Problems

Platforms present a chicken-and-egg problem: how to get enough users on both sides of the platform to achieve the critical mass necessary to make the platform operate. One way of approaching this problem is to find something that will attract one side of the market even when the other side has yet to be developed. Once one side is engaged, however, it is important to ensure there is no friction in the feedback loop — no barriers to user participation. The platform needs to reach critical mass as quickly as possible. Appreciate that and tailor incentives to them.

In some situations, it’s possible to launch in standalone mode, without users. But there has to be a way users can enjoy it even when others are not yet on it, a hook. Users will sign on because they like the features of the hook, not because they are promised a community with other users somewhere down the road. OpenTable did this by offering restaurants a booking management system that worked as a standalone application. After enough restaurants joined the platform, OpenTable added a feature allowing the public to make reservations. Delicious began with a tool that allowed people to store browser bookmarks in the cloud. After lots of people joined, social features were added so people could share bookmarks. With a community of interacting users, network effects began to take hold.

Another approach is to fake it. Counterfeit profiles can provide the illusion that a community exists long enough to get users onboard so that a real community develops. Demand can be seeded by doing things like making real purchases from sellers when there aren’t enough transactions on the system. Tactics like faking and seeding must be approached cautiously, however, as they can backfire.

The reality is that some users are harder to capture than others. You can get the more difficult side to join by offering incentives — the easier side will follow. This strategy is used by nightclubs that offer cheap or free drinks to women on “ladies’ night.” Women are drawn to the bar by the promise of cheap drinks, and men are drawn because women are present.

Platforms are all unique, and there is never a one-size-fits-all solution to acquiring users. The cases presented in this chapter, however, offer a good starting point for developing solutions.

Chapter 5: Virality: Scale in a Networked World

Platforms are propelled by growth, and growth is created by increasing users and interactions.
The tendency for viral growth is called virality, a word that is guaranteed to trigger spell check. Virality is different from word-of-mouth growth, when people hear about the new thing and want to try it. With virality, people discover the new thing by using it. For example, someone sends their friend a link for a YouTube video. The recipient clicks the link and discovers YouTube by watching the video, by participating as a user on the platform.

Virality isn’t the same thing as network effect. Network effect comes from growth within the walls of the platform. For example, the more people that are on LinkedIn, the more people you can contact through their messaging system. You can’t, however, use that system to email someone who isn’t on LinkedIn. Virality means that a platform is pulling from other platforms. Gmail has virality because you can use that platform to send an email to a different platform, like Hotmail. Hopefully, the exposure the platform gains from contact with other platforms will draw new users in. Keep in mind, however, that virality is good, but it can’t be your only growth strategy.

Virality should be well integrated into a platform without any feeling of awkwardness. People shouldn’t be encouraged to needlessly spam their friends for the sake of virality. Look at it as a design problem. Find a way to incentivize its spread.

Make core value units that can be spread to other networks. For example, it’s easy for users to share JPEGs just about anywhere, but not so much with holograms. Don’t make holograms the core value unit of your platform. People won’t be able to circulate them very easily. The units that get spread around act like advertisements for the platform.

Be careful in selecting the networks on which to spread your units. They should be relevant to your platform. It makes good sense that LinkedIn chose to partner with Microsoft Outlook; the two platforms are appropriate for each other. Add value to users on those eternal networks. The users will see how great the platform is and gravitate towards it.

Build viral engines that are part of the normal functioning of the platform. Have lots of units moving to external platforms. Give people on those external platforms a way to interact with the units, perhaps by clicking on them or taking some other simple action. Build cycles of viral growth; keep those cycles short so everything happens quickly.

Chapter 6: Reverse Network Effects

Usually network effects are good for platforms, but there are some that can create undesirable consequences. To avoid negative network effects, a platform should always strive to keep old users engaged and maintain high quality at scale.

There are specific things you can do to scale different dimensions of the platform. There are many suggested tactics, such as how to increase scale of content creation, consumption, customization, curation, corner cases and interaction risk.

To increase interactions on the platform, reinforce the core interaction by focusing on creation, curation, customization and consumption. To scale creation, focus on building the producer base. Make sure producers are sufficiently rewarded for participating, and have a strong feedback system to help producers maintain a high level of quality. Scaling creation can help to scale consumption. Better and more content can draw more consumers. This is a positive feedback loop: the more consumers there are, the more producers will be interested in the platform.

With all that content, the platform needs good filters. There needs to be good data capture in order to understand users so that the filter can make the best recommendations. Data collection on users and their preferences should be continuous. This will allow customization to tailor the platform to the needs and desires of each individual user.

As a platform grows, curation ensures that quality remains high. There are several different approaches to this. Content can be curated editorially, with human editors who act as gatekeepers for contributions to the platform. Algorithmic curation allows algorithms to identify undesirable content and removes it from the platform. Social curation is usually achieved by letting users rate content, identifying both the good and the bad for the benefit of other users. In practice, most platforms use some combination of these tactics.

It’s important to mitigate risk as platforms scale. The platform should be designed to diminish the opportunity for misuse. Interaction risk needs to be minimized or else people won’t want to use the platform.

Beyond a certain scale, a platform might become less useful for its users. Maybe new users ruin it, or curation sucks and it’s hard to find anything good among the junk, or else it gets spammy. Mindlessly pursuing scale for scale’s sake can harm the platform. Keep the negative network effects at bay by scaling up quality as the platform grows in scale.

There are numerous real life examples illustrating various ways to fail at scale. These cases reflect and reinforce the concepts introduced in the chapter. Like the descriptions of different ways to fail, these descriptions of barriers also reflect and reinforce the concepts introduced in the book.

A brief epilogue offers some pointers to transform an existing pipe business into a platform. For businesses to survive into the future, they will find that the platform model gives them the advantages they need to achieve success. The platform world is here to stay.