FOR ARTICLE REPRINTS CALL 800-988-0886 OR 617-783-7500, OR ViSIT HBR.

ORG

Profess or, o va
Id Business Echo
oI

Profes.sir, Harvard
8uS fr›.OSS DC h OOH

.
I

H arv ard 8 u sines s

///usf ro I o ns by S HOUT
Review Jon u O ry— Feb rd
a ry a/g
This document is authorized for educator review use only by Fatima Habib, FAST School of Management until Mar 2023. Copying or posting is an intringement of copyright.
Permissions @ hbsp.har ard.edu or 617.783.7860
 S’ !"ry

Idea in ârief

It s easier for
digital platforms to
achieve scale than
to maintain it.

THE REASO?
Five basic network
properties shape their ”
scalability, profitability,
and ultimately their ” .› ”
sustai'nabiIity.

THE S OHT ”
Analysis of these properties wilt help
entrepreneurs and investors
understand platforms prospects for
long-termsuccess.

n2oi6, Didi became the world's largest ride-sharing

company, reaching 2 million trips a day in China and
surpassing the combined daily trips of all other ride-
sharing companies across the globe. It had arrived at
this milestone by merging in 2Oi5 with its domestic
rival, Kuaidi, and pushing Uber out of the Chinese
market after a fierce, expensive battle. With its
competition gutted, Didi gradually began to improve its
margins by reducing subsidies to drivers and
passengers.

4
Harvard Business Review
la nuory— February song

This document is authorized for educator re'viaw use only by Fatima Habib, FAST School of Management until Mar 2023. Copying or posting is an infringement of oopyright.
Permissions@Tbsp.harvard.edu or 617.783.7860
 Videogameconsotesexhibitweak
networkefects.Thetotalnumberofgame titles available isn't as important as
having a few of the right games. So an entrant with only a small technical
advantage can steal significant market share.

But just as the company began to reach profitability, operation. Lasting competiti•. e advantage hinges more
in early 2018, Meituan, a giant player in online-to-offline on the interplay between the platform and the network it
services such as food delivery, movie ticketing, and travel orchestrates and less on internal, firm-level factors. In
booking, launched its own ride-hailing business in other words, in the digitally connected economy the long-
Shanghai. Meituan didn't charge drivers to use its platform term success of a product •r service depends heavily on
for the first three months and aRerwaid took only 8% of the health, defersihiiity, end dominance of the ecosystem
their revenues, while Didi took 2O°ñ. Drivers and in which it operates.
passengers flocked to the new service. In April, Didi struck And as D'di is learning, it's ofien easier for a digital
back by entering the food delivery market in Wuxi, a city platform to achieve scale than to sustaiii it. After all, the
close to Shanghai. What followed was a costly price war, advantages that allow the platform to expand quickly work
Cth many meals being sold for itx competitors and anyone else who wants to get into
for next to nothing because of heavy subsidies from both the market. 3he reason that some platforms thrive while
companies. So much for Didi's profitability. others struggle really lies in their ability to manage five
Didi was taking other hits too. In March 2018, Alibaba's fundamen- tal properties of networks: network effects,
mapping unit—Gaode Map, the largest navigation service in clustering, risk of disintetrilediation, vulnerability to multi-
China—had started a carpooling business in Chengdu and homing, and bridging to multiple networks.
Wuhan. It didn't charge drivers at all, and in July it began
offering passengers the option of ordering from several ¥fs-en h
ride-hailing services. Meanwhile, Ctrip, China's larges.t
online travel service, had announced in Aprii +F.at it had
been panted a license to provide car-hailing sei vices Nef«• A!Jf’ects
across the country. The importance of network effects is well known. Econo-
Why hadn't Didi's immense scale shut down its mists have long understood that digital platforms like Face-
comp‹•- tition for ride services in China? Why wasn't this book enjoy same-side (“direct”) network effects: The more
a winner- take-all market, as many analysts had predict d? Facebook friends you have in your network, the more likely
Moreover, why do some platform businesses—such as you are to attract additional friends through your friends'
Alibaba, Face- book, and Airbnb—flourish, while User, connections. Facebook also leverages cross-side (“indirect”)
Didi, and Meituan, among others, hemorrhage cash? fat network effects, in which two different groups of partici-
enables digital pants—users and app developers—attract each other. Uber
platforms to fight off competition and grov' profits? can similarly ruine cross-side effects, because more drivers
To answer those questi‹iris, you need to understand attract more riders, and vice versa.
the networks a platform is embchñed in. The factors Less well acknowledged is the fact that the strength of net-
affecting the growth and sustainability ofolaUorm firms work effects can vary dramatically and can shape both
(and digital operating models generally! differ from those value creation and capture. When network effects are
of traditional firms. Let's dart w iii: th= far. that on many strong, the value provided by a platform continues to rise
digital networks the cost of serving an additional user is sharply with the number of participants. For example, as
negligible, which makes a business inherently easier to the number of users on Facebook increases, so does the
scale up. And because much of a network-based firm's amount and variety of interesting and relevant content.
operational complexity is outsoi'rced to the service Video game consoles, however, exhibit only weak network
providers on the platform or handled by soRn•aie, effects, as we discov- ered in a research study. This is
bottlenecks to value creation and growth usually at en't because video games are a hit-driven business, and a
tied to human or organizational factors--anuiher important platform needs relatively few hits to be successful. The total
departure from traditional models. Ultimately, in a digital number of game titles available isn't as important in console
network business, the employees don't deliver the product sales as having a few of the right games. Indeed, even an
or service—they entrant with only a small technical
just deign and oversee an automated, algorithm-driven

H arvar d B u siness R evie›v

I a n u a ry— Feb ru any 2 O g
 FOR ARTICLE REPRINTS CALL 800-988-0886 OR 617-783-7500, OR ViSIT HBR.ORG

This document is authorized for educator review use only by Fatima Habib, FAST School of Management until Mar 2023. Copying or posting is an intringement of copyright.
Permissions @ hbsp.har ard.edu or 617.783.7860
 advantage (and a good business development team) can
steal significant market share from incumbents. That
explains why in 2001 Microsofi's new Xbox posed such a In a research project with Xinxin Li of the University of
threat to Sony's then-dominant PlayStation 2, and why each Connecticut and Ehsan Valavi, a doctoral student at Harvard
console has gone up and down in market share, alternately Business School, we round that the structure of a network
taking the lead, over the years. influences a*!stfoim business’s ability to sustain its scale.
Even more critically, the strength of network effects can The more a network 1s fragmented into local clusters—
change over time. Windows is a classic example. During and the more isolated tllose clusters are from one another
the heyday of personal computers in the l99os, most PC —the more vuliiersble a business is to challenges.
applications were “client based,” meaning they actually Consider Uber. Drivers in Boston care mostly about the
lived on the computers. Back then, the software's network number of riders in Soston, and riders in Boston care
effects were strong: The value of Windows increased mostly about drivers in Beeston. Except for frequent
dramatically as the number of developers writing apps for it travelers, no one in Boston cares much about the number
climbed, topping 6 million at the peak of its popularity. By of drivers and riders in, say, an Francisco. This makes it
the late 199Os Windows seemed entrenched as the leadirig easy for another ride-sharing service to reach critical
platform. However, as internet-based apps, which worked mass in a local market and take
across different operating systems, took off, the network off through a differentiated offer such as a lower price.
effects of Windows diminished and barriers to entry fell, Indeed, in addition to its rival Lyfi at the national level, Uber
allowing Android, Chrome, and iOS opera.t'ng systems to confronts a number of local threats. For example, in New
gain strength on PCs and tablets. Mac shipments had also York City, Juno and Via, as well as local taxi companies, are
begun to rise in the mid-2ooos, incrc.asing more thaI: five- giving it competition. Didi likewise faces a number of strong
fold by the end of the decade. This turn of events i1tu:.tiates contenders in multiple cities.
that when an incumbent’s network elects weaken, so does Now let's compare Uber's market with Aiibnb's.
its market position. Travelers don't care much about the number of Airbnb
It is possible for firms to design features that strengthen hosts in their home cities; instead, they care about how
network effects, however. Amazon, for example, has built many there are in the cities they plan to visit. Hence, the
multiple types of effects into its business model over the network more or less is one large cluster. Any real
years. In the beginning, Jw.azon’s revie'z systems gener- challenger to Airbnb would have to enter the market on a
ated same-side effects: A‹. the number of product reviews global scale—building brand aware- ness around the world
on the site increased, users became more likely to visit to attract critical masses of travelers and hosts. So breaking
Amazon to read the rex jews as wcil as write them. Later, iiito Aiibnb's market becomes much more costly.
Amazon's marketplace, which allows third parties to sell It's possible to strengthen a network by building global
products to Amazon users, generated cross-side network clusters on top of local clusters. While Craigslist, a classified
effects, in which buyers and third-party sellers attracted ad site, primarily connects users and providers of goods and
each other. Meanwhile, Amazon's recommendation system, services in local markets, its housing and job listings attract
which suggests products on the basis of past purchase users from other markets. Facebook's social games (like
behavior, amplified the impact of the company's scale FarmVille) established new connections among players who
by continually learning about consumers' preferences. were strangers, creating a denser, more global, more inte-
The more con•.umers used the site, the more accurate the grated network, which is easier to defend flour competition.
recommendations Amazon could provide them. While Both Facebook and WeChat, a popular social-networking
not usually recognized as a network effect per se, learning app in China, have been enhancing their networks by
effects operate a lot like same-side effects and can increase getting popular brands and celebrities—those with
barriers to entry. national and ofien international appeal—to create public
accounts and post and interact with users.

H arv ard 8 u sines s R evie v

la nu a ry - Fe bru a r0y 2 9

This document is authorized for educator review use only by Fatima Habib, FAST School of Management until Mar 2023. Copying or posting is an intringement of copyright.
Permissions @ hbsp.har ard.edu or 617.783.7860
 FOR ARTICLE REPRINTS CALL BOO-98B-0886 OR 617-7B3-7500, OR VISIT MBR.ORG

Jome digirof n•fwori‹s

cfustr:re a/users. /›a Uber's

network riders and
drivers
Uber :nreroct with network
mi,•mbers c•ntside their home

requfc‹rfy connect with hosts

around the world.
Platjârrns on
globa( networks are
much fess vufnerobfe to
cha(lenges, because it's
di cuJt/or
new rivofs fa enter a market
on a global scale.

ofEtisiatecniediatioa saw this effect in a study of an online freelance

marketplace. As the platformimprovedits reputation-rating
Disintermediation, wherein network members bypass a system, trust between clients and fieelancers grew
hub and connect directly, can be a big problem for any stronger, and disin- termediation became more frequent,
platform that captures vaiue directly from matching or by offsetting the revenue gains from better matching.
facilitating transactions. Imagine that you hire a house Sume platforms address disintermedi tion risks by
cleaner from a platform like Homejoy and are satisfied with introducing different strategies for capturing value—with
th-• service Would you really go back toHomejoy to hire varying results. Thumbtack, a marketplace connecting
the same person again? If a user has found the right match, consumers with local service providers such as
there's little incen- tiveto return to the platform. electricians and guitar teachers, charges for lead
Addition£ly, aRer obtaining enough clients from a generation: Customers post requests on the site, and
platform to fillhis or hDT schedule, the house cleaner won't service providers send them quotes and pay Thumbtack
need that platform anymore. This was exactly the problem fees if those customers respond. That model captures
that doomed Homejoy, which shut down in zois, five years value before the two sides even agree to work together
after it was founded. and has helped save the company from withering like
Platforms have used various mechanisms to deter disin- Homejoy. Thumbtack today is handling over
termediation, such as creating tennis of see.ice that prohibit $i billion worth of transactions annually. The downside of
users from conducting transactions elf the platform, and its revenue model is that it doesn't prevent the two sides
blocking users froin exchanging contact information. Airbnb, from building a long-term relationship outside the platform
for example, withholds hostss••xact locations and phone after a match.
numbers untii payments are made. Such strategies aren't Alibaba took a different approach with its Taobao e-
always effective, ti:ougi:. .L.ything that makes a platform com- rnerce platform. When Taobao entered the market,
more cumbersome to use can make it vulnerable to a com- in zoo3, eBay's EachNet had more than as•a of the
petitor offering a soeanilined experience. Chinese consumer-to-consumer market. However,
Some platforms try to avoid disintermediation by Taobao didn't charge listing or Oansaction fees and even
enhancinethe value of conducting business on them. They set up an instant- messaging service, Wangwang, that
may facilitate transactions by providing insurance, payment allowed buyers to ask questions directly of sellers and
escrow, or communication tools; resolve disputes; or haggle with them in real time. In contrast, EachNet
monitor activities. Fsut those services become less valuable charged sellers transaction fees and, because it was
once trust develops among platform users—and the concerned about disintermedi tion,
strategies can back- fire as tire need for the platform didn't allow direct interactions between buyers and
decreases. One of us, Feng, and Grace Gu, a doctoral sellers until a sale had been confirmed. Not surprisingly,
student at Harvard Business School, Taobao quickly took over leadership of the market, and at
the end
 Harvard Bu s in ess Review January— Fe bru ary Haig

This document is authorized for educator review use only by Fatima Habib, FAST School of Management until Mar 2D23. Copying or posting is an infringement of copyright.
Permissions @hbsp.harvard.edu or 617.783.786D
 FOR ARTICLE REPRINTS CALL BOO-98B-0886 OR 617-7B3-7500, OR VISIT MBR.ORG

of 2Oo6, eBay shut down its Chinese site. Taobao today to n u a ry - Fe bru a ry 2Org

continues to offer its C2C marketplace services flee of

charge and captures value through advertising revenues
and sales of storefront software that helps merchants
manage their online businesses.
ARer estimating that it could lose as much as 90% of its
business to disintermediation, the Chinese outsourcing
marketplace ZBJ, which launched in 2006 with a model of
charging a 20% commission, began looking for new reve-
nue sources. In 2014 it discovered that many new business
owners used its site to get help with logo design. Typically,
the next job those clients would need done was business
and trademark registration, which the platform started to
offer.
Today ZBJ is the largest provider of trademark registration
in China—a service that generates more than $7o million
âi annual revenue for the firm. The company has also
signifi- cantly reduced its transaction fees and focused its
resources on grong its user base instead of fighting
disinteiniedia- tion. As the experience of ZBJ, which is now
•.'alued a* more than $i.s billion, shows, when
disinterinetiiaiion is a *iireat.
providing complementary services can work a lot better than
charging transaction fees.

Multi-homing happens widen users or service

providers tnetwork “nodes”) form ties with multiple
platforms too “hubs”) at the same time 3his generally
occurs when the cost of adopting an additiona!
platform is low. In the
ride-hailing industry, many rlrivers and riders use both, say,
LyR and ñ'bei—ri*ers to compare prices and wait times,
and drivers to reduce their idle time. Similarly, merchants
oRen work Cth mu1tipl.• group-buying sites, and
restaurants with multiple food-deliver y platforms. And
even app developers, whose costs are not trivial, still find it
makes sense to develop piorlucts for both iOS and Android
systems.
¥t’lien mule-homing is pervasive on each side of a
platform, as it is in ride hailing, it becomes very difficult for
a platform to generate a profit flour its core business. Uber
and Lyh are constantly undercutting each other as they
compete for rihers and drivers.

H arvar d B u sines s R evie›v

 Incumbent platform or.mers can reduce multi- LivingSocial started
homing by locking in one side of the n*arâ.•t (or even to source more exclusive deals. While Groupon was able
both sides). To encourage exclusivity, both Uber and LyR to reduce merchant-side multi-homing, the research
gave bonuses in many markets ‹o people who completed found, consumers became more likely to visit both sites,
a certain number of trips in a row withuiit rejecting or because there were fewer overlapping deals on them,
cancellng any or going offline during head hours. And and it cost little to multi-home. That finding points to a
while rides are in progress, both platforms provide key challenge platform firms face: Reducing multi-homing
drivers new requests for pickups very close to curreiit on one side of the market may increase multi-homing on
passengers' drop-off locations, reduc- ing the drivers' idle the opposite side.
time and hence the temptation to use other platforms. Other approaches seem to work better. Let's look again at
Yet because of the inherently low cost of adopting the video game industry: Console makers oRen sign exclu-
multiple platforms, multi-homing is still rampant in ride sive contracts with game publishers. On the platforms' user
sharing. side, the high prices of consoles and subscription services,
Attempts to prevent multi-homing can also have such as Xbox Live and PlayStation Plus, reduce players'
unintended side effects. In one research project, Feng incentives to multi-home. Lowering multi-homing on both
and Hub Ei of Carnegie Mellon University examined sides of the market decreased competitive intensity and
what allowed the console makers to be profitable. Amazon, which
happened in ooh when Groupon retooled its deal counter— provides fulfillment services to third-party sellers, charges
which tracks the amount of people who have signed up them higher fees when their orders are not from Amazon's
for a specific offer on its site—to show ambiguous marketplace, incentivizing them to sell exclusively on it.
ranges, rather than precise numbers. It then became Amazon Prime, which gives subscribers flee two-day ship-
more difficult for LivingSocial to identify and poach the ping on many products, helps the company reduce online
popular merchants on Groupon. As a result, shoppers' tendency to multi-home.

This document is authorized for educator review use only by Fatima Habib, FAST School of Management until Mar 2023. Copying or posting is an intringement of copyright.
Permissions @ hbsp.har ard.edu or 617.783.7860
 FOR ARTICLE REPRINTS CALL 800-988-0886 OR 617-783-7500, OR ViSIT HBR.ORG

“Managing Our Hub Economy” ' “Alibaba and the

Marco lansiti and ruture of
Business”
Karim R. Lakhani ring Zeng
HBR, September-0ctober zo1$ , HRR, Septen br•r-October 2o18

together industries. Just as the Alibaba Group moved

flour commerce to financial services, .mazon has moved
In many situations the best growth strategy for a platform beyond retail to entertainment and consumer electronics.
may be to connect different networks to one another. In Platforms are thus becoming crucial hubs in the global
any platform business, success hinges on acquiring a high economy.
number of users and amassing data on their interactions.
Such assets can almost invariably be valuable in multiple ¥¥HEH EVALUATMfi Ah opport anity involving a platform,
scenarios and markets. By leveraging them, firms that have entrepreneur s (and investors) should analyze the basic
succeeded in one industry vertical ofien diversify into differ- properties cf the networks it will use and consider ways to
ent lines of business and improve their economies. This is a strengthen network effects. It's also critical to evaluate the
fundamental reason why Amazon and Alibaba have moved feasibility of minimizing multi-homing, building global net-
into so many markets. work structures, and using network bridging to increase
When platform owners connect Cth multiple networks, scale i:while mitigating the risk of disintermediation. That
they can build important synergies. Alibaba successfully exercise win L!uniiriate the key challenges of growing and
bridged its payment platform, Alipay, with its e-commerce sustaining
platforms Taobao and Tmall, providing a much-needed *he platform and help businesspeople develop more-realistic
service to both buyers and sellers and fostering trust bed' assessments of the platform's potential to capture value.
e.n them. Alibaba has also taken advantage of transaction As for Didi and Uber, our analysis doesn't hold out
and user data from Taobao and Tmall to launch new.• much herpe. Their networks consist of many highly local
offer.rigs through its financial services arm, Ant Finarciai— clusters. they both face rampant multi-homing, which may
including worsen as more rivals enter the markets. Network-bridging
a credit-rating system for merchants anrl consumers. And opportuni- ties—their best hope—so far have had only
information from that rating system allcwed Ant Financial limited success.
to issue short-term consumer and merchant loans with They've been able to establish bridges just with other highly
•/eiy low default rates. With those loans, consumers can competitive businesses, like food delivery and snack
purchase more products on Alibaba's e-cc.hmm.free vending. (In 2018 Uber struck a deal to place Cargo's snack
platforms, and Ali- baba's merchants can fund more vending machines iii its vehicles, for instance.) And the
inventory. These networks mutually reinforce one another's inevitable rise of self-driving taxis will probably make it
market positions, helping each network sustain its scale challenging for Didi and Uber to sustain their market
*ndeed, ever afier the rival platform Tencent offered a capitalization. Network properties are trumping platform
competing tligital wallet service, scale. %
WeChat Pay, through its app WeChat. Aiipay remained HBR Reprint Rl90lJ
attrac- tiveto consumers and merrhaais because of its tight
bridging with Alibaba md Ant Financial'sother services. FENG ZHU is the Piromal Associate Professor of Business
As the most s‹ccessf•a! pi*tforms connect across more and Administration at Harvard Business Schoo/. viARCO IANSITI /S
more markets, ti‹ey'ie becoming increasingly effective at tying the oovid Sntnojf PtoJessot of Business Administtotion or Homotd
Business Scñco!. He hos ndv/sed many companies in the technology
sector, including Nicrcso/, Facebook, and Amazon.

Harvard Busi ness Review

This document is authorized for educator review use only by Fatima Habib, FAST School of Management until Mar 2023. Copying or posting is an intringement of copyright.
Permissions @ hbsp.har ard.edu or 617.783.7860