《供应链管理 Supply Chain Management》课程教学资源（研讨资料）How Companies Become Platform Leaders

WINTER 2008 VOL.49 NO.2 MTS oan Management Review Annabelle Gawer and Michael A.Cusumano How Companies Become Platform Leaders Please note that gray areas reflect artwork that has been intentionally removed.The substantive content of the article appears as originally published. REPRINT NUMBER 49201

How Companies Become Platform Leaders WINTER 2008 VOL.49 NO.2 REPRINT NUMBER 49201 Annabelle Gawer and Michael A. Cusumano Please note that gray areas reflect artwork that has been intentionally removed. The substantive content of the article appears as originally published

STRATEGY How Companies Become Platform Leaders Under the right circumstances, n recent years,many high-technology industries,ranging from "smart"cell companies of any phones to social networking Web sites such as Facebook Inc.and MySpace.com, have become platform battlegrounds.These markets require distinctive com- size can grow to petitive strategies because the products are parts of systems that combine core become platform components made by one company with complements usually made by a variety of compa- nies.If a platform leader emerges and works with the companies supplying complementary leaders.And particular products and services,they can together form an"ecosystem"of innovation that can greatly increase the value of their innovations as more users adopt the platform and its comple- business and technology ments.However,companies often fail to turn their products into industry platforms. decisions can help Our previous research focused on understanding the levers or strategic mechanisms that existing platform leaders use to maintain their positions.(See"About the Research,"p.31.) platform-leader This article focuses on the special problems of companies that want to become platform leaders-"platform-leader wannabes."Many companies do not succeed in becoming plat- wannabes achieve form leaders because their strategies fail to tackle adequately both the technology and their goals. business aspects of platform leadership.The technological challenges involve designing the right architecture,designing the right interfaces/connectors and disclosing intellectual property selectively,in order to facilitate third-parties'provision of complements.The busi- Annabelle Gawer and ness challenges include either making key complements or introducing incentives for Michael A.Cusumano third-party companies to create the complementary innovations necessary to build market momentum and defeat competing platforms. Our strategic recommendations consist of two basic approaches.(See"Strategic Options for Platform-Leader Wannabes,"p.32.)One strategy,"coring,"addresses the challenges of creating a new platform where one has not existed before.The second strategy,"tipping," tackles the problem of how to win platform wars by building market momentum. The Platform Vs.Product Strategy Choice There is an important difference between a product and an industry platform.Put simply, a product is largely proprietary and under one company's control,whereas an industry platform is a foundation technology or service that is essential for a broader,interdependent ecosystem of businesses.The platform requires complementary innovations to be useful, and vice versa.An industry platform,therefore,is no longer under the full control of the originator,even though it may contain certain proprietary elements. Managers sometimes underestimate the importance of deciding early on between pursu- ing a product or a platform strategy.This decision matters because the industry conditions Annabelle Gawer is lecturer in strategy and innovation at Tanaka Business School,Imperial College London.Michael A.Cusumano is the Sloan Management Review Distinguished Professor of Manage- ment and Engineering Systems at the MIT Sloan School of Management.Comment on this article or contact the authors at smrfeedback @mit.edu. 28 MIT SLOAN MANAGEMENT REVIEW WINTER 2008 SLOANREVIEW.MIT.EDU

28 MIT SLOAN MANAGEMENT REVIEW WINTER 2008 STRATEGY How Companies Become Platform Leaders n recent years, many high-technology industries, ranging from “smart” cell phones to social networking Web sites such as Facebook Inc. and MySpace.com, have become platform battlegrounds. These markets require distinctive com￾petitive strategies because the products are parts of systems that combine core components made by one company with complements usually made by a variety of compa￾nies. If a platform leader emerges and works with the companies supplying complementary products and services, they can together form an “ecosystem” of innovation that can greatly increase the value of their innovations as more users adopt the platform and its comple￾ments. However, companies often fail to turn their products into industry platforms. Our previous research focused on understanding the levers or strategic mechanisms that existing platform leaders use to maintain their positions. (See “About the Research,” p. 31.) This article focuses on the special problems of companies that want to become platform leaders — “platform-leader wannabes.” Many companies do not succeed in becoming plat￾form leaders because their strategies fail to tackle adequately both the technology and business aspects of platform leadership. The technological challenges involve designing the right architecture, designing the right interfaces/connectors and disclosing intellectual property selectively, in order to facilitate third-parties’ provision of complements. The busi￾ness challenges include either making key complements or introducing incentives for third-party companies to create the complementary innovations necessary to build market momentum and defeat competing platforms. Our strategic recommendations consist of two basic approaches. (See “Strategic Options for Platform-Leader Wannabes,” p. 32.) One strategy, “coring,” addresses the challenges of creating a new platform where one has not existed before. The second strategy, “tipping,” tackles the problem of how to win platform wars by building market momentum.1 The Platform Vs. Product Strategy Choice There is an important difference between a product and an industry platform. Put simply, a product is largely proprietary and under one company’s control, whereas an industry platform is a foundation technology or service that is essential for a broader, interdependent ecosystem of businesses. The platform requires complementary innovations to be useful, and vice versa. An industry platform, therefore, is no longer under the full control of the originator, even though it may contain certain proprietary elements. Managers sometimes underestimate the importance of deciding early on between pursu￾ing a product or a platform strategy. This decision matters because the industry conditions Annabelle Gawer is lecturer in strategy and innovation at Tanaka Business School, Imperial College London. Michael A. Cusumano is the Sloan Management Review Distinguished Professor of Manage￾ment and Engineering Systems at the MIT Sloan School of Management. Comment on this article or contact the authors at smrfeedback@mit.edu. I Under the right circumstances, companies of any size can grow to become platform leaders. And particular business and technology decisions can help platform-leader wannabes achieve their goals. Annabelle Gawer and Michael A. Cusumano SLOANREVIEW.MIT.EDU

prices on the operating system they will load onto their PCs. Failure to decide early on between a product or platform strategy can result in dangerous strategic confusion.Achieving platform status requires specific decisions that govern technology evolution,product and system design and business relation- ships within the ecosystem-and they are different decisions than those made when pursuing a product strategy.Another com- mon mistake is that managers can simply overlook the platform potential of their products.For example,Apple Inc.'s Macin- tosh personal computer was the leading product when it was introduced but didn't become the dominant personal computing platform,primarily because Apple did not open the Mac's architecture and software to third-party complementors and licensees. While the benefits of becoming a plat- form seem clear,not every market has to have a platform leader.In some large mar- kets,such as video game consoles or Web portals,several platform companies can persist without one clear winner.For that scenario to occur,it seems important that the market contain enough room for dif- ferentiation in user needs so that multiple companies can persist in specific niches or segments,particularly if it is not too diffi- cult for users to switch among more than one platform.2 Nor can every product become a plat- form.3 To have platform potential,however, research suggests that a product(or a tech- nology or service)must satisfy two and business choices that favor a platform can differ from those prerequisite conditions:(1)It should perform at least one essential that favor a product-creating differing incentives for owners of function within what can be described as a"system of use"or solve industry platforms than for companies that assemble proprietary an essential technological problem within an industry,and(2)it products.In particular,owners of industry platforms benefit from should be easy to connect to or to build upon to expand the system lots of innovation in complementary products as well as from of use as well as to allow new and even unintended end-uses. competition at the overall system level that would bring its price It is possible to test for these conditions.For the first,one can down.Thus,Microsoft Corp.benefits from competition among evaluate whether the overall system could function without the personal computer manufacturers that use its operating system, particular product or technology.If the system cannot operate, but they,in contrast,benefit when customers perceive their prod- then the product does indeed perform an essential function.For ucts as unique and therefore do not want cutthroat competition at example,Microsoft's Windows operating system and Intel's mi- the product or system level at which they compete.PC makers croprocessor were both essential platform components of the would probably rather see Microsoft face tough competition in original IBM and IBM-compatible personal computers.For the computer operating systems so that they could bargain for better second condition,the challenge is to test whether a product or a SLOANREVIEW.MIT.EDU WINTER 2008 MIT SLOAN MANAGEMENT REVIEW 29

WINTER 2008 MIT SLOAN MANAGEMENT REVIEW 29 and business choices that favor a platform can differ from those that favor a product — creating differing incentives for owners of industry platforms than for companies that assemble proprietary products. In particular, owners of industry platforms benefit from lots of innovation in complementary products as well as from competition at the overall system level that would bring its price down. Thus, Microsoft Corp. benefits from competition among personal computer manufacturers that use its operating system, but they, in contrast, benefit when customers perceive their prod￾ucts as unique and therefore do not want cutthroat competition at the product or system level at which they compete. PC makers would probably rather see Microsoft face tough competition in computer operating systems so that they could bargain for better prices on the operating system they will load onto their PCs. Failure to decide early on between a product or platform strategy can result in dangerous strategic confusion. Achieving platform status requires specific decisions that govern technology evolution, product and system design and business relation￾ships within the ecosystem — and they are different decisions than those made when pursuing a product strategy. Another com￾mon mistake is that managers can simply overlook the platform potential of their products. For example, Apple Inc.’s Macin￾tosh personal computer was the leading product when it was introduced but didn’t become the dominant personal computing platform, primarily because Apple did not open the Mac’s architecture and software to third-party complementors and licensees. While the benefits of becoming a plat￾form seem clear, not every market has to have a platform leader. In some large mar￾kets, such as video game consoles or Web portals, several platform companies can persist without one clear winner. For that scenario to occur, it seems important that the market contain enough room for dif￾ferentiation in user needs so that multiple companies can persist in specific niches or segments, particularly if it is not too diffi￾cult for users to switch among more than one platform.2 Nor can every product become a plat￾form.3 To have platform potential, however, research suggests that a product (or a tech￾nology or service) must satisfy two prerequisite conditions: (1) It should perform at least one essential function within what can be described as a “system of use” or solve an essential technological problem within an industry, and (2) it should be easy to connect to or to build upon to expand the system of use as well as to allow new and even unintended end-uses. It is possible to test for these conditions. For the first, one can evaluate whether the overall system could function without the particular product or technology. If the system cannot operate, then the product does indeed perform an essential function. For example, Microsoft’s Windows operating system and Intel’s mi￾croprocessor were both essential platform components of the original IBM and IBM-compatible personal computers. For the second condition, the challenge is to test whether a product or a SLOANREVIEW.MIT.EDU

STRATEGY technology is easy to connect to or to build upon.One way to do accomplish this,but looking at successful and unsuccessful com- this is to see whether external companies have succeeded in de- panies can provide ideas on what to do and what not to do veloping complementary and interoperable products,or at least have started to do so.Unless these two conditions are fulfilled,the Google:Coring in Internet Search Google Inc.is a particularly well- strategic game of platforms cannot begin.But they are far from known and clear example of successful coring in Internet search sufficient to win the platform game. technology.The company,founded in 1998,started off as a sim- Our research explores the issue of platform leadership inin- ple search engine company and went on to establish its proprietary formation technology industries such as computing and search technology as a foundation for navigating the Internet. telecommunications because these industries not only have visi- First,Google improved upon existing solutions to an essential ble demarcations between platforms and complements but also technical problem:how to find anything in the maze of the Inter- have strong"network effects"between the two,leading to clear net,with millions of Web sites,documents and other online interdependencies.However,companies can pursue platform content.Google's improved search function became an essential strategies in many different industries.For example,new energy technology for fully using the Internet.Second,Google distrib- sources,such as hydrogen fuel cells or hybrid gasoline-electric uted its technology to Web site developers and users as an systems,may become platforms for powering a variety of devices embedded toolbar,making it easy to connect to and to develop made by different companies.Banks,credit card companies and upon.It also allowed different uses,such as combining a search Internet services companies all are competing to develop a plat- with different kinds of information or graphics. form for micropayments and other specialized financial services. But where Google really won the platform leadership battle In biology,the human genome database has become a platform for Internet search was on the business side.Google solved a for many companies and research laboratories.Pharmaceutical fundamental problem,which was that it was not initially clear and chemical manufacturers develop certain compounds that how companies could make money from using the Internet. can become the basis for a variety of drugs or other products Google found a way to link focused advertising to user searches. made by themselves and many partner companies. Ads appear only along with specific searches,meaning that users should have some interest in the advertisers.In effect,Google Coring:How to Create a New Industry Platform revolutionized the advertising business by rearchitecting the rela- "Coring"is the set of activities a company can use to identify or tionships between advertisers and Internet users.Today,Google's design an element (a technology,a product or a service)and market value is over $200 billion,many times that of the largest make this element fundamental to a technological system as well advertising agencies. as to a market.An element or component of a system is "core" Of course,Google had competition.In the mid-1990s,Digi- when it resolves technical problems affecting a large proportion tal Equipment Corp.created a powerful search engine tool for of other parts of the system.Coming up with platform-like tech- the Internet,AltaVista;several other companies created power- nologies may well be easier than coming up with business ful search engines,such as Yahoo!and Inktomi.But Google strategies that encourage partners and customers to adopt a par- proved to be much more effective than its competitors at the ticular technology. business aspect of market coring,even though Internet search Platforms open the overall system in which they operate to and Web portals are a broad enough market that more than one new usage possibilities.These different uses are essential to the company is likely to persist.As of April 2007,Google accounted growth of an installed base,but one question arises:Who will for about 55%of Web searches,compared to about 22%for develop these new uses?How can platform-leader wannabes suc- Yahoo!and 9%for MSN/Windows Live Search,according to a cessfully encourage other companies to join their ecosystems and Netratings Inc.survey. develop essential complementary applications?Answering that Google continues to extend and promote its platform.In June question is one of the two essential business aspects of coring. 2007,Google held its first developers'conference,with 1,000 The platform leader must create economic incentives for ecosys- programmers in attendance and another 5,000 at 10 other loca- tem members to invest in creating complementary innovations tions around the world.The agenda included presentations on and to keep doing so over time.In addition,platform-leader Google's application programming interfaces to enable develop- wannabes need to protect their ability to profit financially from ers to embed Google applications such as search,maps and their innovations,just as any innovator company should.The calendars on Web sites or to develop custom search engines. balancing act-protecting one's sources of profit while enabling Google also presented APIs for the Web 2.0 social networking site complementors to make an adequate profit and protect their own YouTube Inc.,which it purchased in 2006.Google has increased proprietary knowledge-is perhaps the greatest challenge to the amount of free online software it provides,ranging from platform leadership.There is no simple framework for how to e-mail to word processing. 30 MIT SLOAN MANAGEMENT REVIEW WINTER 2008 SLOANREVIEW.MIT.EDU

30 MIT SLOAN MANAGEMENT REVIEW WINTER 2008 SLOANREVIEW.MIT.EDU technology is easy to connect to or to build upon. One way to do this is to see whether external companies have succeeded in de￾veloping complementary and interoperable products, or at least have started to do so. Unless these two conditions are fulfilled, the strategic game of platforms cannot begin. But they are far from sufficient to win the platform game. Our research explores the issue of platform leadership in in￾formation technology industries such as computing and telecommunications because these industries not only have visi￾ble demarcations between platforms and complements but also have strong “network effects” between the two, leading to clear interdependencies. However, companies can pursue platform strategies in many different industries. For example, new energy sources, such as hydrogen fuel cells or hybrid gasoline-electric systems, may become platforms for powering a variety of devices made by different companies. Banks, credit card companies and Internet services companies all are competing to develop a plat￾form for micropayments and other specialized financial services. In biology, the human genome database has become a platform for many companies and research laboratories. Pharmaceutical and chemical manufacturers develop certain compounds that can become the basis for a variety of drugs or other products made by themselves and many partner companies. Coring: How to Create a New Industry Platform “Coring” is the set of activities a company can use to identify or design an element (a technology, a product or a service) and make this element fundamental to a technological system as well as to a market. An element or component of a system is “core” when it resolves technical problems affecting a large proportion of other parts of the system. Coming up with platform-like tech￾nologies may well be easier than coming up with business strategies that encourage partners and customers to adopt a par￾ticular technology. Platforms open the overall system in which they operate to new usage possibilities. These different uses are essential to the growth of an installed base, but one question arises: Who will develop these new uses? How can platform-leader wannabes suc￾cessfully encourage other companies to join their ecosystems and develop essential complementary applications? Answering that question is one of the two essential business aspects of coring. The platform leader must create economic incentives for ecosys￾tem members to invest in creating complementary innovations and to keep doing so over time. In addition, platform-leader wannabes need to protect their ability to profit financially from their innovations, just as any innovator company should. The balancing act — protecting one’s sources of profit while enabling complementors to make an adequate profit and protect their own proprietary knowledge — is perhaps the greatest challenge to platform leadership. There is no simple framework for how to accomplish this, but looking at successful and unsuccessful com￾panies can provide ideas on what to do and what not to do. Google: Coring in Internet Search Google Inc. is a particularly well￾known and clear example of successful coring in Internet search technology. The company, founded in 1998, started off as a sim￾ple search engine company and went on to establish its proprietary search technology as a foundation for navigating the Internet. First, Google improved upon existing solutions to an essential technical problem: how to find anything in the maze of the Inter￾net, with millions of Web sites, documents and other online content. Google’s improved search function became an essential technology for fully using the Internet. Second, Google distrib￾uted its technology to Web site developers and users as an embedded toolbar, making it easy to connect to and to develop upon. It also allowed different uses, such as combining a search with different kinds of information or graphics. But where Google really won the platform leadership battle for Internet search was on the business side. Google solved a fundamental problem, which was that it was not initially clear how companies could make money from using the Internet. Google found a way to link focused advertising to user searches. Ads appear only along with specific searches, meaning that users should have some interest in the advertisers. In effect, Google revolutionized the advertising business by rearchitecting the rela￾tionships between advertisers and Internet users. Today, Google’s market value is over $200 billion, many times that of the largest advertising agencies. Of course, Google had competition. In the mid-1990s, Digi￾tal Equipment Corp. created a powerful search engine tool for the Internet, AltaVista; several other companies created power￾ful search engines, such as Yahoo! and Inktomi. But Google proved to be much more effective than its competitors at the business aspect of market coring, even though Internet search and Web portals are a broad enough market that more than one company is likely to persist. As of April 2007, Google accounted for about 55% of Web searches, compared to about 22% for Yahoo! and 9% for MSN/Windows Live Search, according to a Netratings Inc. survey.4 Google continues to extend and promote its platform. In June 2007, Google held its first developers’ conference, with 1,000 programmers in attendance and another 5,000 at 10 other loca￾tions around the world. The agenda included presentations on Google’s application programming interfaces to enable develop￾ers to embed Google applications such as search, maps and calendars on Web sites or to develop custom search engines. Google also presented APIs for the Web 2.0 social networking site YouTube Inc., which it purchased in 2006. Google has increased the amount of free online software it provides, ranging from e-mail to word processing. STRATEGY

About the Research Over the past decade,we have investi- The focus of our initial work was on ecosystem.The fourth lever was internal gated dozens of companies that have how Intel.Microsoft,Cisco and other organization:how and to what extent attempted to formulate and implement companies had been able to drive indus- platform leaders should use their organ- platform strategies.These companies try innovation and sustain positions of izational structure and internal operated in a variety of industries includ- platform leadership.We identified four processes to give assurances to external ing computing,telecommunications, "levers"or mechanisms through which complementors that they are genuinely electronic appliances,semiconductors, successful platform leaders were able to working for the overall good of the eco- enterprise software,data storage,auto "architect"or influence external innova- system.Taken together,the four levers mobiles,Web portals and electronic tion.The first lever was company scope: offer a template for sustaining a posi- payment systems.The major companies the choice of what activities to perform tion of platform leadership. we studied in the first phase of our re- in-house versus what to leave to other This article presents findings from search included Intel,Microsoft,Cisco, companies-in particular,whether the the second stage of our research,which Palm,and NTT DoCoMo,the Tokyo- platform leader should make at least draws heavily on public information.It based mobile communications some of its own complements in-house. has been inspired primarily by several company.We interviewed hundreds of The second lever was technology design consulting engagements(such as with managers and engineers and comple- and intellectual property:what func- Nokia,EMC,Tokyo-based information mented the interviews with analysis of tionality or features to include in the technology company NTT Data and e companies'archival records and com- platform,whether the platform should frontier,the 3D computer graphics de- pany and industry data.This first be modular and to what degree the plat- veloper based in Santa Cruz,California), research stage aimed at uncovering form interfaces should be open to contacts with managers at organiza- the drivers of success at established outside complementors and at what tions using our original framework(such platform leaders.The results of that price.The third lever covered external as enterprise resource planning soft- work were published in MIT Sloan relationships with complementors:the ware provider SAP,the Internet Home Management Review in 2002,as well process by which the platform leader Alliance and Siemens Automation)and as in our book Platform Leadership manages complementors and encour- numerous MIT master's theses and Ph.D. (HBS Press,2002). ages them to contribute to a vibrant dissertations,as well as class projects. Qualcomm:Coring in Wireless Technology Another company that has upon Qualcomm's technology-the second prerequisite condi- done very well in the technological aspects of coring is Qual- tion for platform potential.To facilitate third parties'adoption of comm Inc.in the wireless technology industry.It has been its technology,Qualcomm invested in chipset designs embedding extraordinarily successful in terms of profitability,although the its technology and made CDMA widely available for licensing. business side of its ecosystem shows some signs of instability due The chipsets were compact integrated circuits with physical con- to opposition from a number of its licensees.Founded in 1985, nectors that made it easy to plug them inside cell phone handsets, Qualcomm started out designing communications technology and Qualcomm's licensing of its patents made it possible for op- for satellites and military applications and went on to establish its erators to use CDMA protocols.This strategy enabled dozens of proprietary wireless communications technology as a platform companies to include Qualcomm technology in most second- for the cellular phone industry.s generation and many third-generation cell phones,as well as in Qualcomm solved a basic technical problem of the late 1980s and hundreds of other wireless devices. early 1990s:incompatible and inefficient wireless cell phone tech- Qualcomm has a more checkered performance in its relation- nologies.This problem negatively affected other industry players ships with other companies in its ecosystem.In the company's such as telecom operators and handset manufacturers.Qualcomm business model,an important source of revenue is from licensing its invented the code division multiple access technology,which breaks intellectual property.Qualcomm therefore filed thousands of pat- phone calls into small bits and then reassembles them,much as the ents and regularly and aggressively challenged any potential violators Internet does with data packets.Key industry players such as AT&T in court.Its customers may not always have appreciated this litigious (later Lucent)and Motorola licensed Qualcomm's technology.By approach.However,since Qualcomm owned approximately 80%of addressing an essential technological problem in its industry,Qual- the patents for CDMA and CDMA2000 technology,they had little comm satisfied the first condition for platform potential. choice for many years.Also,Qualcomm lessened the conflicts with It was also easy for other companies to connect to and build some of its key ecosystem members in the late 1990s by selling its cell SLOANREVIEW.MIT.EDU WINTER 2008 MIT SLOAN MANAGEMENT REVIEW 31

SLOANREVIEW.MIT.EDU WINTER 2008 MIT SLOAN MANAGEMENT REVIEW 31 Qualcomm: Coring in Wireless Technology Another company that has done very well in the technological aspects of coring is Qual￾comm Inc. in the wireless technology industry. It has been extraordinarily successful in terms of profitability, although the business side of its ecosystem shows some signs of instability due to opposition from a number of its licensees. Founded in 1985, Qualcomm started out designing communications technology for satellites and military applications and went on to establish its proprietary wireless communications technology as a platform for the cellular phone industry.5 Qualcomm solved a basic technical problem of the late 1980s and early 1990s: incompatible and inefficient wireless cell phone tech￾nologies. This problem negatively affected other industry players such as telecom operators and handset manufacturers. Qualcomm invented the code division multiple access technology, which breaks phone calls into small bits and then reassembles them, much as the Internet does with data packets. Key industry players such as AT&T (later Lucent) and Motorola licensed Qualcomm’s technology. By addressing an essential technological problem in its industry, Qual￾comm satisfied the first condition for platform potential. It was also easy for other companies to connect to and build upon Qualcomm’s technology — the second prerequisite condi￾tion for platform potential. To facilitate third parties’ adoption of its technology, Qualcomm invested in chipset designs embedding its technology and made CDMA widely available for licensing. The chipsets were compact integrated circuits with physical con￾nectors that made it easy to plug them inside cell phone handsets, and Qualcomm’s licensing of its patents made it possible for op￾erators to use CDMA protocols. This strategy enabled dozens of companies to include Qualcomm technology in most second￾generation and many third-generation cell phones, as well as in hundreds of other wireless devices. Qualcomm has a more checkered performance in its relation￾ships with other companies in its ecosystem. In the company’s business model, an important source of revenue is from licensing its intellectual property. Qualcomm therefore filed thousands of pat￾ents and regularly and aggressively challenged any potential violators in court. Its customers may not always have appreciated this litigious approach. However, since Qualcomm owned approximately 80% of the patents for CDMA and CDMA2000 technology, they had little choice for many years. Also, Qualcomm lessened the conflicts with some of its key ecosystem members in the late 1990s by selling its cell Over the past decade, we have investi￾gated dozens of companies that have attempted to formulate and implement platform strategies. These companies operated in a variety of industries includ￾ing computing, telecommunications, electronic appliances, semiconductors, enterprise software, data storage, auto￾mobiles, Web portals and electronic payment systems. The major companies we studied in the first phase of our re￾search included Intel, Microsoft, Cisco, Palm, and NTT DoCoMo, the Tokyo￾based mobile communications company. We interviewed hundreds of managers and engineers and comple￾mented the interviews with analysis of companies’ archival records and com￾pany and industry data. This first research stage aimed at uncovering the drivers of success at established platform leaders. The results of that work were published in MIT Sloan Management Review in 2002, as well as in our book Platform Leadership (HBS Press, 2002). The focus of our initial work was on how Intel, Microsoft, Cisco and other companies had been able to drive indus￾try innovation and sustain positions of platform leadership. We identified four “levers” or mechanisms through which successful platform leaders were able to “architect” or influence external innova￾tion. The first lever was company scope: the choice of what activities to perform in-house versus what to leave to other companies — in particular, whether the platform leader should make at least some of its own complements in-house. The second lever was technology design and intellectual property: what func￾tionality or features to include in the platform, whether the platform should be modular and to what degree the plat￾form interfaces should be open to outside complementors and at what price. The third lever covered external relationships with complementors: the process by which the platform leader manages complementors and encour￾ages them to contribute to a vibrant ecosystem. The fourth lever was internal organization: how and to what extent platform leaders should use their organ￾izational structure and internal processes to give assurances to external complementors that they are genuinely working for the overall good of the eco￾system. Taken together, the four levers offer a template for sustaining a posi￾tion of platform leadership. This article presents findings from the second stage of our research, which draws heavily on public information. It has been inspired primarily by several consulting engagements (such as with Nokia, EMC, Tokyo-based information technology company NTT Data and e frontier, the 3D computer graphics de￾veloper based in Santa Cruz, California), contacts with managers at organiza￾tions using our original framework (such as enterprise resource planning soft￾ware provider SAP, the Internet Home Alliance and Siemens Automation) and numerous MIT master’s theses and Ph.D. dissertations, as well as class projects. About the Research