platform wars in cloud computing & social...
TRANSCRIPT
© 2011 MIT Center for Digital Business. All rights Reserved.
Platform Wars in Cloud
Computing & Social Media May 19, 2011
Proprietary & Confidential
Michael A. Cusumano MIT Sloan School of Management & Engineering Systems
2
My Six “Enduring” Principles Not original to me but “big ideas” underlying my work over 25-30 years, with lots
of support in empirical & theoretical research by others, in various disciplines
1. Platforms, Not Just Products
2. Services, Not Just Products (or Platforms)
3. Capabilities, Not Just Strategy
4. Pull, Don’t Just Push
5. Scope, Not Just Scale
6. Flexibility, Not Just Efficiency 3
4
Strategy
Push
Scale
Efficiency
Products
Narrow Way of Thinking About Focus and
Competitive Advantage at the Product Level
Capabilities,
Not Just Strategy
Platforms
& Services,
Not Just Products
Broader Way of Thinking About Agility and
Competitive Advantage at the Ecosystem Level
Examples:
• Toyota
• Microsoft
• Intel
• JVC in VHS Era
• Apple after mid-2000s
• Google, Adobe
• Cisco, Qualcomm, et al.
Examples:
• Ford in Model T Era
• GM in the 1920s
• Sony in Betamax era
• IBM before Open Source
• Apple before mid-2000s
Pull, Don’t Just Push
Scope, Not Just Scale
Flexibility, Not Just Efficiency
Platforms Examples
5
6
Industry Platform Definition
• A foundation technology or set of components (could also be a service) used beyond a single firm
• Brings multiple parties (“market sides”) together for a common purpose or to solve a common problem
• Value of the platform increases geometrically with (a) more complementary products & services, and (b) more users – ad infinitum (maybe).
– Various names: network or bandwagon effect, positive feedback, network externality
– Historical Examples: Railroad, Telegraph or Telephone, Electric Power Grid, Radio, TV, VCRs, computers, etc.
Platform Ecosystem: Platform +
Complements + Network Effects
7
Platform(E.g., VHS player,
Windows-Intel PC, Apple iPhone,
Barbie doll)complementary product
complementaryservice
Direct network effect
number of users
number of advertisers, content providers, channel partners, etc.
Indirect networkeffect
positive feedbackloop
Source: M. Cusumano, Staying Power (2010)
8
Ongoing Platform Battlegrounds • Web Search Google vs. Bing/Yahoo, foreign engines
• Smart PhoneOS Apple vs. RIM, Google Android, Microsoft/Nokia
• Digital Media Apple vs. Microsoft, Google?
• Social Media Facebook, Twitter, LinkedIn, et al.
• Video Games Sony, Nintendo, Microsoft
• Enterprise s/w SAP vs. Oracle/Sun, Microsoft, IBM
• Micropayments Sony Felica vs. PayPal, credit cards
• Displays E-Ink vs. LCD vs. Plasma (Sharp, Sony, Samsung)
• Batteries Sony vs. Panasonic, Sanyo, A123, others
• Power systems Traditional vs. hybrid vs. electric, fuel cell, etc.
And many more platforms, or platforms on top of
platforms, in smaller or emerging markets
App Ecosystem: Number of Applications Available
Apple is the leader in the application ecosystem, with he highest
number of available apps, though Android gaining fast
Platforms vs. Products:
Managerial Implications
• To succeed in a platform market requires
having the “best” platform & platform
strategy, not the “best” product or service
– Best product? Hard to define
– Best platform? We know some things…
(1) Open (but not too open) interfaces or access
(2) Modular architectures (easy to build on & extend =
outside complementary innovation)
(3) Compelling complements (generally the result of a
vibrant ecosystem + in-house capabilities) 10
The Platform Strategy “Toolkit”
Product or Platform (or Complement)? – Key
decisions, and how to think about them
The 4 Levers (Firm Scope, Platform Technology,
External Relations, Internal Organization) – broad
categories for implementing a platform leadership strategy
“Coring” & Tipping – Strategies for platform-leader
wannabes
WTAoM – Framework to analyze the dynamics of platform
markets & complements (Winner-Take-All-or-Most?), estimate
how much share is possible to win, and design initiatives to
influence outcomes
11
12
The 4 Levers – How Best Use?
1. Firm Scope: Making or controlling key complements
yourself helps ensure success of current & future platforms
2. Technology: Modular architectures, cheap or accessible IP,
and “open” interfaces facilitate complements; but still need to
protect your technology (“open, but not open” vs. “closed, but not
closed” vs. “free, but not free”)
3. External: Build an ecosystem (win friends and influence
people) through cooperation and assistance, rather than coercion
4. Internal: To build trust with ecosystem partners, need to
separate groups helping complementors or partners use the
platform from those making competing complements or offering
competing products & services (if you are doing both in-house,
conflicts of interest are likely to arise)
Leader-Wannabe Strategies
13
“Winner Take All” (or Most) if…
1) Strong network effects between the platform
and complements (direct or indirect – crossing
market sides – or both)
2) Little differentiation among competing
platforms (i.e., few niche opportunities or ways
to be distinctive among competitors!)
3) Multi-homing rare difficult or costly for users,
app developers, or other ecosystem players to
use more than one platform as their “home”:
MAKE THEM CHOOSE!
Ref: Eisenmann, Parker, and van Alstyne, Harvard Business Review (2006); Cusumano, Staying Power (2010)
14
Platform WTAoM Analysis?
• VCRs
• Desktop OS
• Videogame Consoles
• Internet Search
• Smartphones
• Cloud Computing
• Social Media/Networks
• Others?
15
Cloud Computing: 3 Service Types
16
Apps
Platform
Infrastructure
• “Cloud computing is a
model that enables
convenient, on-demand
network access to a shared
pool of configurable
computing resources.”
• “resources can be rapidly
provisioned and released
with minimal management
effort or service provider
interaction.”
• Source: NIST Cloud Computing Definition. http://csrc.nist.gov/groups/SNS/cloud-computing/
• Compiled by MIT Sloan 15.358 Student Team, 2010
Definition* Service categorization
Infrastructure as a Service (IaaS)
17
Client Services
End user apps
App Services
Middleware
OS (s)
Storage/DB
Processing
Networking
Facilities
Services offered Description: • Offers “data-center” services (e.g. compute power, storage)
• Various operating systems, but typically simple OS or virtualization
• Key advantages:
- Flexibility for developer to custom-build applications and
services using preferred development environment
- Ability to move application in-house since typically use open
standards and relatively fewer proprietary services
• Disadvantage:
- More difficult to use than PaaS or SaaS
What is it used for? • Developing new web-based applications (e.g., by start-ups) with-out
incurring fixed costs
• Often used for testing new systems before deploying in production
• Can be used for peak capacity handling, instead of provisioning for
peak-capacity handling in-house
Key vendors: • Amazon.com
• Rackspace Services offered
Platform as a Service (PaaS)
18
Client Services
End user apps
App Services
Middleware
OS (s)
Storage/DB
Processing
Networking
Facilities
Services offered Description: • Combines traditional development platform + pre-built services
• Platform customized by each vendor (OS, dev environment and
services specific to that environment
• Key advantages:
- Built-in services reduce development time & time-to-market
- Potentially familiar development environment
- Potentially reuse others’ services = strong network effects
• Key disadvantage:
- Lock-in to platform vendor
What is it used for? • Developing new web-based applications (e.g., by start-ups) with-out
incurring fixed costs
• Developing services for other platform users
Key vendors: • Microsoft Azure
• Google Apps
• Salesforce.com – Force.com
Services offered
Product/App Layer – SaaS
19 19
Client Services
End user apps
App Services
Middleware
OS (s)
Storage/DB
Processing
Networking
Facilities
Services offered Description: • Software product for end users – not a service for developers
• Each software vendor develops an application on an in-house or
(PaaS/IaaS) “platform”
• Software not installed on customer’s hardware, but accessed through
a web browser
• Key advantages:
- Low monthly charge by usage, not license + maintenance
- Easier to adopt & install than conventional product or building
your own system
- No need for users to constantly upgrade or apply bug fixes – all
maintenance handled by SaaS provider
What is it used for? • Developing new web-based apps on top of existing product
• Delivering existing applications in a pay-as-you-go model
Key vendors (there are many): • Salesforce
• NetSuite
• HubSpot and many other web-based startups Services offered
Transition: SaaS to IaaS & PaaS (e.g., Salesforce.com)
• Software product company delivers its product(s) as a
service from a remote site, accessed via browser, pay by
usage, no maintenance = SaaS
• Software product company offers access to its general
infrastructure resources (e.g., computing power, data
storage) for product hosting = IaaS
• Software product company opens its product APIs as
“services” or functions, and provides infrastructure &
development tools to 3rd-party app developers = PaaS 20
Will There Be One Winner in the
Global IaaS Market? Network effects? Differentiation? Multihoming?
1. Strong network effects? – Not obvious for IaaS
vendors compared to PaaS or SaaS.
2. Little differentiation? – Probably. Future should
mostly see competition based on price of basic service.
3. High cost of multihoming? – No? Developers --
inconvenient but can use multiple IaaS & definitely switch over
time. Users -- should be agnostic to the IaaS provider, and can
therefore use multiple vendors, if compatible.
21
Will There Be One Winner in the
Global PaaS Market? Network effects? Differentiation? Multihoming?
1. Strong network effects? – Yes, direct – if vendors such
as Microsoft can tie platform services to use of SaaS versions of
Microsoft products as well as development tools
2. Little differentiation? – No, because the “servitized”
products, dev tools & dev environments differ. Yes, in that
standardization going on for cloud development.
3. High cost of multihoming? – Not so clear. Web
services make it easier to mix & match SaaS or hosted products.
But customers will still be tied to particular applications (e.g.
SAP, Oracle, Msft, Salesforce), databases, etc.
22
Conclusions Re Cloud Platforms • IaaS already being “commoditized”
– Even utility companies, like AT&T and Verizon, getting into
this space
• Enterprise software product competition likely to be
duplicated in PaaS competition
– Intense price competition (toward “free”), consolidation,
dominance by a few vendors,
– But no “winner take all or most” scenario, as in VCRs, OS, or
even search
• Product/tool companies like Microsoft have an
advantage: Can combine SaaS and PaaS strategies
23
Cloud Conclusions cont’d • As in other platform markets, software product
companies need to think about alternatives such as
in the Platform Strategy Toolkit:
– How use the 4 levers (what to offer in-house versus from the outside;
how open to make the platform, ecosystem promotion measures, how
organize internally to be perceived as neutral)
– How try to “tip” the market (e.g., where to subsidize or bundle
capabilities; use of coring techniques such as FB Connect, or “Azure
Connect”?)
– How generate strong network effects (e.g. through APIs to
products, promotion campaigns)
– How minimize multihoming (e.g., through “compelling” PaaS
offerings tying products to services, “open but not open” technologies,
strategic subsidies, etc.) 24
Bet How Much on SaaS?
Another long discussion, but…
• SaaS clearly important for web-based startups and
enterprise software for SMEs; less interest among
bigger customers
• SaaS (on-demand) no more than 3-4% of revenues for
major product companies like Oracle or SAP; relatively
stable over past several years
• Customer surveys indicate strong future interest in
SaaS, though ongoing issues of security & response
time slowing adoption
25
Social Media Platform Wars
• Social media networks represent a technology and business not
possible without the internet. But, as businesses, they need to
leverage and sustain the viral phenomenon.
• Faddish element or diminishing returns – growth rates in usage of
FB, Twitter, et al. slowing. Privacy concerns. Also, “90-10” rule in
usage, not even “80-20!”
• Facebook -- in strong position due to size & potential to increase
scope. But what of Google OpenSocial coalition challenge?
• LinkedIn and Twitter -- strong niche positions. But vulnerable to
“envelopment” by FB if cannot protect their differentiation!
26
27
Top 10 Social Media Websites, February 2010
Rank Website U.S. Share of Visits
1 Facebook 49.62%
2 MySpace 15.47%
3 YouTube 15.20%
4 Tagged 1.19%
5 Twitter 1.12%
6 Yahoo! Answers 1.05%
7 Yahoo! Profiles 0.80%
8 myYearbook 0.60%
9 Windows Live Home 0.54%
10 Meebo 0.54%
Source: Hitwise
28
Weekly Share of U.S. Visits to Facebook + Google Source: Experian Hitwise US
29
http://www.businessinsider.com/chart-of-the-day-how-many-users-does-twitter-really-have-2011-3?op=1
http://www.businessinsider.com/chart-of-the-day-how-many-users-does-twitter-really-have-2011-3?op=1
http://www.businessinsider.com/chart-of-the-day-how-many-users-does-twitter-really-have-2011-3?op=1
Will There Be One Winner (FB) in
the Social Networking Market? Network effects? Differentiation? Multihoming?
1. Strong network effects? – Yes. Friends, colleagues,
etc. Specific applications and some services, though weakening
with cross-platform APIs.
2. Little differentiation? – No. Many small niche
platforms or more focused, also many overseas. Yes –Facebook
copying specialized features of others, and Google is pushing
more cross-platform openness & applications.
3. High cost of multihoming? – Yes – individual users,
and application developers for “closed” platforms. No – many
people use multiple platforms, for different purposes.
30
31
Problem Focus for Users
Platform elements
Business Model WTAoM Potential?
-- strong network effects?
-- minimal differentiation?
-- little multihoming?
Envelopment Potential?
Main Strategic Challenges
Facebook How interact with friends, and meet friends of friends, share information and photos, with privacy controls
Open, but not open. APIs, query language & markup language – for app developers. FB Connect for easy access to other sites, bringing community with you
$2B sales for 2010, maybe $4B for 2011? Was $275m in 2008! Profitable. Sources: 55% - ads via FB Pages & Social Ads (but 50 times weaker results than Google), 30% - search (Msft), 15% - virtual goods. 1000 employees in 2010
WTM, yes. Already most visited web site, ahead of Google. Especially strong in US though 70% of users outside
Yes, for users + app developers, advertisers, affiliates Now biggest internet site. 600m users in 2011, from 175m in mid-2009, gaining 25+m/mo. By end of 2010, 550K apps., up from 20,000 in 4/2008. But growth slowing, and active users less than registered users .
Maybe. Facebook has not adapted Google OpenSocial, so may remain distinctive and semi-closed – not WTA but WTM. Also copying others (professional, messaging)
No. Nothing prevents FB users from using other SN sites, such as MySpace (30%) or LinkedIn or Twitter, or others
Already into professional and company pages. Already added real-time updates and search. Will launch many more apps. New email, video platform?
1. How much to take on prof or LinkedIn and Twitter-type functionality? 2. Remain closed or not? 3. How help companies better engage with users (adds) and make more money? 4. How help dev new apps using F. Connect?
LinkedIn How create a professional network and leverage contacts of your contacts to search for jobs discreetly or search for people to hire.
Closed, but not closed? Opened iAPIs, but strictly controls apps. Adopted Google OpenSocial APIs, enabling sharing of apps from different sites.
$220m est. rev for 2010; was $100m in 2008. Profitable. Revenue: Job placement fees (41%), ads (32%), subscriptions (27%)
Maybe, in category. But prof market has niches and regions. Also, soc networking sites could move into professional
Yes, for users + app developers, recruiters, advertisers, experts. 100m users in 2011/3, up from 23m users in mid-2008, gaining 1.5m/mo. Half of all members now non-US.
No. Other similar sites, but PNS can be regional, professional – lots of vertical specializations possible. Also different networks via Facebook, etc.
No. Users also use Facebook (41%, MySpace (28%) – Ex. 4
Can add soc net features. Will add recruiting, knowledge mgt, workplace collaboration, other tools.
1. Add soc net features or stay focused on professionals? 2. Open up the platform more or stay a “walled garden”?
Twitter How share messages or information in real-time with friends and followers
Open. APIs introduced in 2006. Apps and connections rising
Maybe $10m/year, from $25m search deal w Google, Msft, Yahoo. Not yet profitable. Future: Charge ffor special services, or search, or sponsorships (Suggested Users list – super direct marketing). Lots of rich information from users
Yes, for real-time update niche. Leader in category today. But features too easy to copy, such as by Facebook
Yes, for users + app developers, advertisers, media outlets, celebrities, gov’t, firms (PR, mktg) . 200m users in 2011/3, up from 10m users 2009/3, growing 8m/mo. Wide coverage – media, etc. 2000 apps by 2009/3 and maybe 300k in 2011. But active users not so many
No,. Twitter has attained a special status, already. But functionality being copied by Facebook.
No. Twitter users also heavy users of other sites. Twitter functionality very limited.
Can add soc net functions. Can also compete with email, IM.
1. How make more money from this viral phenomenon (e.g. Sponsored Suggested Users) without killing the mystique and credibility?
Google OpenSocial a Threat?
• Coalition strategy to mobilize non-FB sites (so far joined by
MySpace, LinkedIn, Plaxo, Salesforce and a few dozen others) to
promote standard app development languages (HTML, JavaScript)
and some data sharing
– The enemy of my enemy is my friend?
– But coalitions very useful in other markets (anti-Microsoft, anti-EMC, anti-
Qualcomm; pro-Linux, pro-Symbian, pro-Java, etc.)
• Not much allowance so far for moving rich individual or group data
sharing from site to site. Not clear how this will work
• Doesn’t have the same kind of viral features as Facebook – e.g.
news feed, invitations, notifications
• Lowest-common denominator APIs, so not so rich applications?
32
WTAoM Strategy for FB? Increase network effects!
– Scope: Develop more compelling in-house applications, with FB Platform
– Technology: Offer more compelling APIs through FB Platform/Connect
– External: Try to make FB Connect “ubiquitous”; subsidize more outside
development; acquire competing or related sites; offer services to tailor
company pages and ads
– Internal: In-house R&D group for ecosystem partners
Decrease differentiation of competitors! – Copy everything useful from Twitter, LinkedIn, et al.!
– Offer specialized FB “channels” (e.g., professional, gaming, hobbies)
Make multi-homing more difficult! – Do not join Google’s OpenSocial!
– Try to market FB Connect through exclusive partnerships?
– Do not allow transfer of rich user content through FB Connect!
33
WTAoM for LinkedIn & Twitter?
• Also focus on:
– increasing network effects (e.g., through more compelling
platform-specific content or apps, or FB Connect-type links);
– making multi-homing more costly for users, app developers,
advertisers, channel partners (e.g., “closed, but not closed”
strategy should lead to more unique content and applications)
• But must retain differentiation relative to Facebook,
while reducing differentiation attempts by their niche
competitors (e.g., Monster, Plaxo, and Xing for
professional networking; and Facebook for tweets)
34
Monetization Challenges • All SN sites need a strategy that (a) fits with the social
platform mission (the problem it is solving for its users
and ecosystem partners) and (b) does not destroy the
spontaneous or viral nature of the social network.
– Facebook: Brings together friends and friends of friends, etc., a social
networking function. Not so useful for selling ads or corporate pages
unless these can be integrated into social (not professional) functions.
– LinkedIn: Very direct monetization strategies that leverage core
platform functions and are not disruptive : 41% Job placement fees from
employers; 32% Ads; 27% Premium Subscriptions
– Twitter: Rich information on user preferences for products, global
trends, etc. But need to worry about usage rates, disrupting spontaneity.
35
Social Media Future?
• P2P technologies & social networking & media here to
stay, despite faddish element and 90-10 usage patterns.
• Entrepreneurs and established firms will come up with
ever more creative ways to use the technology for
personal, professional & other uses (economics, art,
problem-solving such as with crowd-sourcing, etc.)
• Particularly interesting are ways to merge social
networking with mobile platforms and location-
specific applications and enterprise applications 36