Machine, Platform, Crowd: Harnessing Our Digital Future

by Andrew McAfee

ClassPass’s big data, broad reach, and many members allowed it to bring the benefits of revenue management to individual exercise studios, and its user interface and user experience around experimentation convinced many studios that their interests were well served by adopting it.

Rent the Runway thus practices a type of revenue management that differs from ClassPass’s approach—one aimed at preserving the value of durable inventory as long as possible by presenting it to the people around the country with the highest willingness to pay, as determined by the company’s algorithms.

These include UserTesting, Survata, dscout, and Google Consumer Surveys.

eliminated the time and uncertainty associated with haggling.

Edaixi, a service that uses a digital platform to make it easier for people to have a large bag of laundry collected, cleaned, ironed, and returned within seventy-two hours for $15.

For less than $5, Guagua Xiche will send a professional technician to wash a car wherever it is parked in a city. Users specify the vehicle’s location and license plate number, and they don’t need to be present during the cleaning. The firm raised $58 million in 2015 and has expanded to twelve cities.

time. It also means that the demand-side economies of scale—the network effects, in other words—can eventually grow much faster than costs.

The huge amount of data that O2O businesses generate makes them fertile territory for machine learning, the information-heavy approaches to artificial intelligence that are now dominant, as we discussed in Chapter 3. User interface and user experience design, too, are experiencing a heyday, in large part because of the popularity of platforms. It’s extremely hard to make websites flexible, powerful, and intuitive all at the same time, and even harder to do the same for apps (because they have to run on phones’ small screens). All the platform builders we’ve talked to have stressed how hard they’ve worked on their user interfaces, and how much they continually iterate and experiment to refine them over time.

It’s also clear how hard they work on their broader user experience. Troubleshooting, customer support, and problem resolution are critical activities, not least because bad word of mouth spreads quickly.

Because they can add new members quickly, control the customer experience, leverage preexisting capital and labor resources, and use data and algorithms to improve matches, O2O platforms can scale rapidly and compete aggressively. Investors see the potential of O2O and have been willing to fund aggressive expansion plans.

Medallion-holding incumbents found it difficult to reverse their losses, because Uber’s two-sided network effects, smooth user interface and user experience, and ample capital were formidable advantages. Attempts to build competing platforms such as Lyft in the United States and Hailo in Europe did not slow down the fast-growing startup. The only thing that could, it sometimes seemed, was regulation.

Modern smartphones are incredibly complicated devices that must be designed well and built reliably. The engineering expertise and global supply chain capabilities required to do this are so formidable that only a handful of companies in the world ever try.

widely. The network effects they can tap into are powerful. They can also become self-reinforcing, especially if they’re supported with a steady stream of refinements to both the visible user interface and the algorithms running behind the scenes. And such refinements themselves become easier to generate as platforms grow, since growth provides ever-larger test beds for iteration and experimentation.

Larger networks also generate ever more data, which smart platform operators exploit to great advantage. They use it to better understand their members, to predict and shape their members’ behavior, and to give their members sophisticated tools for revenue management, pricing, and other critical activities.

Regardless of their size, platforms control the user experience and user interface for their members. They determine what information is seen by users and how processes and transactions are executed. When platform owners exercise control well, two important things happen: First, they can reduce or remove long-standing barriers that have kept people from transacting with each other. Second, they can influence the flow of transactions so that more benefits go to the platform owner.

It’s bad for business when one party knows a lot more about a proposed transaction than the other one does. It’s bad for both parties because the less informed party will often recognize being that they are at an information disadvantage and therefore unable to properly evaluate the proposal.

Instead, these companies used their platforms’ user interfaces to overcome the information asymmetries that plagued their markets. In particular, they asked all parties to rate each other after each transaction, and they prominently displayed everyone’s cumulative ratings.

Dual reputation systems can support a remarkable amount of exchange even in the absence of law or regulation.”

Airbnb CEO and cofounder Joe Gebbia refers to this system of peer reviews and ratings as “design for trust” and highlights another of its benefits: it can help us overcome our personal biases. We’re sure that most Airbnb hosts don’t consider themselves racist, but the company’s data show that, on average, hosts are slightly less willing to rent out to minority guests than to white ones.¶ However, this effect reverses itself for minority guests with a good overall rating and more than ten reviews. Prospective hosts are actually more likely to rent to such guests than to white ones without many ratings. “High reputation beats high similarity,” as Gebbia puts it; his company has found that its platform’s user interface and user experience “can actually help us overcome one of our most deeply rooted biases—the ‘stranger-danger’ bias.”

But as platforms grow, they want to keep more of the consumer’s share of both mind and wallet. Toward this end, the best weapon that platforms have is control of the user interface and of the digital user experience. Excellence here is often beyond the reach or outside the expertise of a single fitness company, but it is in the sweet spot of a platform builder like ClassPass.

But increasingly, instead of charging the users, they give them away for free to maximize demand. That way, they make more money on merchant fees on the other half of the two-sided market. By lowering the annual fees and other user charges on their cards, credit card issuers can not only increase the market share of the cards, but also increase the attractiveness of their networks to merchants, as well as the associated processing fees. If giving away

The wonderful virtuous cycle of two-sided networks then becomes a vicious cycle of downward-spiraling market share. Furthermore, platform owners can’t just focus on pricing. They have a variety of other levers to manage, including the user interface and user experience, reputation systems, marketing budgets, and core network technology. The most successful platform owners carefully curate the value that each side of the market gets from participating and aren’t too greedy.

As it works to build its platform, Uber has two huge advantages. The first is a set of deep-pocketed and patient investors, who are willing to cover Uber’s costs while it scales. These initial costs—for technology development, marketing, driver recruiting, staffing, and so on—are substantial, and Uber was estimated to have raised more than $15 billion in loans and investments by July 2016 to fund its global growth. Investors have been willing to supply

Of course, even platform economics doesn’t provide immunity to new competition, management mistakes, or technological shifts. None of today’s platform heroes can grow complacent.

platforms are particularly good at displacing incumbents when there’s not a lot of perceived difference among their offerings.

This contrast highlights the fact that while urban rides might be close to a single product market within each city, urban stays are clearly not. With its platform, Airbnb essentially introduced a second product in this market, one aimed at people who wanted something different, and often cheaper, than what hotels traditionally provided.

participants (only one buyer and very few sellers). In addition, the transaction is incredibly complex and requires enormous amounts of communication. Markets in which players are few and offerings are complicated will probably be some of the least amenable to platforms. So, activities like designing and building a power plant, providing tax advice on a large merger, and coordinating all the details of bringing together art from around the world for a major museum exhibit will likely continue to be carried out as they have in the past, and will not be taken over by digital platforms.

Being a big strapping male or a nubile female won’t affect the amount of deference you get. . . . This does lead to a freer, truly disembodied mingling of minds.”

The core is characterized by government bodies, approval loops, and people and groups with the formal power to say no.

As the mathematician and author John Allen Paulos observed in the early days of the web, “The Internet is the world’s largest library. It’s just that all the books are on the floor.”

“Act in good faith, and assume good faith on the part of others.” Instead

We are confident that effective solutions will be possible by bringing together minds and machines. One promising approach here is to allow people to flag fake or offensive content while training machine learning systems to spot it automatically.

Hayek argued that something like Polanyi’s Paradox applies throughout the economy: we can’t tell all of what we know, what we have, what we want, or what we value.

Markets are therefore called “emergent” systems: prices emerge from all members’ interactions and can’t be observed just by looking at a few.

events with final share prices of about $0.70 tend to actually happen about 70% of the time, making these prices pretty accurate probability estimates.

Because information is often widely dispersed among economic actors, it is highly desirable to find a mechanism to collect and aggregate that information. Free markets usually manage this process well because almost anyone can participate, and the potential for profit (and loss) creates strong incentives to search for better information.”

Studying Linux’s history reveals several principles that appear to be important, perhaps even essential, for bringing the crowd together to accomplish something significant. These include openness, noncredentialism, verifiable and reversible contributions, clear outcomes, self-organization, and geeky leadership.

writing an operating system is very different from endeavors to generate other creative products, like novels or symphonies. It’s not at all clear or externally verifiable whether someone’s proposed contribution of a new chapter or character to a novel improves the work.

Torvalds once said, “I am not a visionary. I do not have a five-year plan. I’m an engineer. . . . I’m looking at the ground, and I want to fix the pothole that’s right in front of me before I fall in”—but it does need to be clear, and it needs to be able to motivate people to devote their time and effort to achieving it.

Openness and noncredentialism made the work available to as many people as possible. Self-assignment meant that they worked on what they wanted, which typically turned out to be what Linux most needed. Verifiability ensured that only helpful contributions survived in the software, and clear outcomes kept people from feeling that they could be duped or their efforts hijacked. And dedicated geeky leadership from Torvalds and others maintained the ideals, culture, and momentum of Linux.

Wikipedia, unlike Nupedia, was able to activate the crowd because it adopted the principles of openness, noncredentialism, and self-organization. It abandoned the notion of standardized, multistep work flows and the requirement that editors be experts or have PhDs. Instead, it threw open the work of building an encyclopedia to any and all to come together and work together in any ways they saw fit. To keep these collaborations from descending into chaos, Wikipedia soon adopted the principle of verifiability, which meant that “other people using the encyclopedia can check that the information comes from a reliable source. Wikipedia does not publish original research.”**

Overcentralization fails because of Hayek’s insights and Polanyi’s Paradox: people can’t always articulate what they have, what they know, what they want, and what they can do.

as people become more prominent and senior in their fields, these blind spots—such as the well-documented biases toward overconfidence and confirmation (that is, only really considering information that supports what you were already thinking)—become stronger and thus lead to worse outcomes.

It could even be that many “experts” actually aren’t expert at all—that they’ve been kidding themselves and the rest of us about their abilities and the quality of their work. In today’s complex, fast-changing, technologically sophisticated world, it can be quite hard to distinguish who actually knows what they’re talking about.

It’s that many problems, opportunities, and projects, if not most, benefit from being exposed to different perspectives—to people and teams, in other words, with multiple dissimilar backgrounds, educations, problem-solving approaches, intellectual and technical tool kits, genders, and so on. This is absolutely the definition of the crowd, and it’s very hard, probably even impossible, to replicate within the core.

Crowdfunding is a way to preassemble financial capital against something with social capital.

It’s like a totem; it’s a psychological anchor into something that you care about.

When Ornua, an Irish agricultural food company, shipped $100,000 worth of cheese to the Seychelles Trading Company in September 2016, it was the first international trade transaction that used the blockchain to record all details of its trade financing. Trade across borders typically doesn’t happen until two conditions are met. First, the parties involved have ironed out all details about the trade financing: insuring the goods while they’re in transit, identifying exactly when ownership is transferred, and so on. Second, all involved parties have satisfied themselves that they have received identical sets of properly signed legal documents related to this financing. Posting all documentation for the Ornua–Seychelles Trading transaction on the blockchain reduced a seven-day process to four hours.

We don’t even need to trust that the courts in our area will be competent, impartial, and expedient, since the smart contract doesn’t rely on them to enforce its terms or verify its legitimacy.

What will the world that they create look like? Here’s what I think: Your technology will work perfectly within the silo and with an individual stacks’s [sic] (temporary) allies. But it will be perfectly broken at the interfaces between itself and its competitors. That moment where you are trying to do something that has no reason not to work, but it just doesn’t and there is no way around it without changing some piece of your software to fit more neatly within the silo? That’s gonna happen a lot.