What to Read Next
Editor’s note: This article was originally published Sept. 15, 2017, with the title “All Platforms Are Not Equal.” This version reflect updates that were made for the winter 2018 print edition of the magazine.
The dramatic influence of the internet on how businesses operate and the emergence of a handful of gigantic, digitally enabled corporations have led to breathless pronouncements regarding the importance of a new class of monopolies built on digital platforms. Such platforms, it is said, can fuel network effects that lead to winner-take-all marketplaces. This perspective is often coupled with infectious optimism and investment euphoria regarding the extraordinary scale and strength of network-effects businesses.
In theory, the key attribute of a network-effects business is its momentum-driven flywheel. Every new participant increases the value of the network to existing participants, attracts more new users, and makes the prospect of a successful competitive attack ever more remote — thereby bolstering the relative attractiveness of the business. The imagined innate indomitability of network effects stems at least in part from the breathtaking strength of notable platform businesses such as Facebook Inc.’s social network or Microsoft Corp.’s Windows operating system.
The problem is that not all platform businesses exhibit network effects that reinforce a market’s winner-take-all tendency. For every Facebook and Microsoft, there are numerous network-effects businesses operating in crowded sectors where it is not always clear that anyone will turn a profit.
Nor are digital platforms necessarily better businesses than the analog versions that they displace. Analog malls had the benefit of their shoppers being miles away from competing malls, as well as the benefit of their retail tenants being committed to long-term leases. On the internet, platform competitors are only a click away, and companies regularly and dynamically optimize their customer reach across competing platforms and directly via their own sites.
It is not that marketplace businesses built on e-commerce platforms do not have advantages or cannot thrive. Rather, it is that the mere existence of network effects tells entrepreneurs and investors relatively little about the attractiveness of a particular business. For example, Uber Technologies Inc. and Airbnb Inc., the global leaders in the ride-hailing and short-term lodging marketplaces, respectively, both benefit from network effects. However, other characteristics of those industries make it likely that Airbnb will enjoy dramatically stronger results than Uber will ever achieve.
Email Updates on Managing Tech
Get periodic email updates on how to incorporate new tech into your company’s strategy and operations.
Please enter a valid email address
Thank you for signing up
Why Airbnb Is Better Than Uber
Three key structural attributes drive the value of network effects in the digital domain. The first is the minimum market share at which the network can achieve financial breakeven. The second is the nature and durability of the customer relationships spawned by the network. And the third is the extent to which the data generated by the network facilitates product and pricing optimization.
These should sound familiar. They are updated versions of the same core competitive advantages that have long underpinned the best business franchises: economies of scale, customer captivity, and learning. And they are as relevant to today’s digital platforms as they were — and continue to be — to analog ones. Comparing Uber and Airbnb along these dimensions highlights both their profound relevance and Airbnb’s inherent advantages.
Minimum Viable Market Share
Two attributes determine the minimum viable scale of a network: product/service complexity and fixed-cost requirements. With regard to these two attributes, Uber and Airbnb could not be more different.
In any given city, the financial viability of both companies is a function of local density — of drivers on the platform on the one hand and available property inventory on the other. A key distinction between Uber’s and Airbnb’s respective marketplaces, however, is the level of intrinsic product complexity and the resulting marketplace liquidity required to establish a competitive service. In ride-hailing, the ability to deliver a car within three to five minutes is the top customer consideration (other than price). Having so many drivers that cars arrive sooner than that is not useful. In lodging, there are multiple customer considerations that ensure that the value of higher incremental density in local short-term lodging listings does not top out in the same way. Indeed, more listings attract more travelers and drive higher occupancy rates.
The second difference is the role of fixed costs. Although both Uber and Airbnb are predominantly variable-cost businesses, Airbnb’s relative fixed-cost requirements are far greater than Uber’s. Both companies have similar technology and overhead costs, but users of ride-hailing services primarily use those services in a single city. In contrast, customers of companies providing short-term lodging services use those services in many different locales. That creates an incentive for those companies to incur the higher fixed costs associated with operating in multiple popular locations.
It may be that in some small markets, the fixed operating costs won’t sustain more than one or two ride-hailing services. But in larger metropolitan areas, multiple robust offerings are available, with viability achievable at market shares of less than 20%. This effectively translates to a permanent pool of five or more Uber competitors, severely limiting achievable returns.
Conversely, the greater fixed-cost needs in short-term lodging mean that Airbnb competitors can break even only at far higher market shares. It is not a coincidence that Airbnb has far fewer direct competitors of size in any given market than Uber does.
The nature and durability of customer relationships determine the speed at which market shares can shift among competitors in a particular marketplace. When combined with minimum viable market share, the level of customer captivity enables a potential new entrant to quickly calculate how long it can expect to lose money before achieving a break-even market share. So, for instance, in an industry where customer loyalty limits annual share movement to a couple of points and break-even market share is 20%, an insurgent can expect at least a decade of losses before establishing viability.
By enhancing the ability to easily search out, compare, and switch between sellers, the internet has set the bar far higher for businesses to articulate truly compelling reasons for customers to stay put. What’s more, customers and business partners operating in an environment characterized by swift technological change are generally wary of long-term commitments. Nonetheless, robust captivity is still achievable when the service quality, breadth of offering, verification of nuanced counterparty credentials, and/or seamless integration of critical data into buying processes are central to the ultimate decision to transact.
Unfortunately, even the best-run ride-hailing company will struggle to encourage loyalty among drivers and riders. Already, about two-thirds of drivers work with two or more services. Moreover, although a minority of riders currently use multiple ride-hailing apps, that percentage has been growing. Among my MBA students in New York City, it is greater than 90%.
There is a difference between an individual’s willingness to entrust short-term rentals of his or her home to multiple companies and a professional driver’s willingness to drive for multiple ride services. And, as we’ve discussed, for the customer, price and speed are the overwhelming factors influencing the decision about which ride service to use for a short trip, but a variety of factors shape the decision to a stay in a stranger’s house. Homeowners want to know who will be staying in their homes, and guests want to know the experiences of others who have used those homes. A single good experience will make customers much less likely to take a chance with an alternative platform that seems to offer a comparable or even slightly better proposition.
When Airbnb establishes a leadership position in a market, competitors are at a disadvantage in terms of inventory availability. However, the same is not true for Uber’s competitors in the ride-hailing market, because the majority of drivers use more than one app. The observance of ride-hailing market-share shifts of greater than 5% over a matter of months nationally (and of even more in some localities) suggests that Uber can expect a steady stream of new competitors. Such competition is less likely for Airbnb, where the time required to recruit and sign up new lodging units significantly slows the potential rate of market-share shifting and the resulting time it would take a new entrant to break even.
Finally, the fortunes of network-effects businesses depend on the value of the data that they can elicit in their respective markets. Zillow Group Inc.’s continued dominance in the online real estate marketplace, for instance, is in part a function of its ability to use its unique access to data to continually improve its automated valuation models and its home search and recommendation engines. In contrast, peer-to-peer lending platforms discovered that, for most borrowers, their proprietary data yielded little more insight than was readily available elsewhere from sources such as credit scores.
Before I book a stay in a stranger’s apartment, I pore over the reviews of previous visitors before taking the plunge, no matter how nice the pictures. By contrast, riders typically don’t use Uber driver reviews to select cars. (Mostly, the reviews are used by the company to manage fleet quality.) And while feedback on drivers helps Uber cull out those who undermine the service and facilitates training, the nuanced picture that emerges from travelers around the world allows Airbnb to direct regular users to the most appropriate venues and helps those listing their homes deliver a satisfying experience.
Beware the Network-Effects Fetish
Uber has built a remarkable business. However, the structural attributes noted above suggest that ride-hailing will continue to be an intensely competitive business in large, local markets. Moreover, in many international markets, Uber is the insurgent, and the network effects it enjoys in the United States provide limited advantage. More broadly, the resilience of Uber’s position hinges on a relentless aggressiveness rather than a structural tendency toward a global winner-take-all (or even winner-take-most) equilibrium.
Airbnb’s network effects, on the other hand, are paired with significant customer captivity. Given the advantages afforded by its global fixed-cost base, the competition it faces is less intense than the competition Uber faces. The lesson for investors and entrepreneurs is to be wary of the fetishization of network effects as an inherently superior form of competitive advantage. Excessive optimism regarding the winning power of digital platforms’ network effects is not justified by either a close study of their structural impact on entry barriers or any empirical evidence of generally increasing market dominance. In fact, many signs suggest quite the opposite — that in the absence of the same characteristics (most notably fixed-cost scale and customer captivity) that have long supported the strongest analog platforms, digital platforms are likely to be significantly harder to build and maintain. It is not a coincidence that two of the largest and most enduring purely digital platforms — Google LLC and Amazon.com Inc. — are companies that benefit from leveraging multiple, complementary sources of competitive advantage.
With the help of other sources of competitive advantage, network-effects businesses can deliver remarkable value to users and riches to entrepreneurs and investors. On their own, however, network effects in a digital context are a peculiarly fragile barrier to entry. Seen in this light, entrepreneurs and investors should treat the identification of network effects as the beginning, not the end, of their analysis. Meanwhile, platform operators should curb any complacent confidence that they may have in their destinies as the conquerors of global markets. Instead, they should redouble their efforts to establish complementary barriers before being displaced by one of what are likely to be many competing platforms.