Microsoft Corporation: The Design of Microsoft@ Support Network 1.

0
Companion Case to ISBM 7-1994 James A. Narus
Wake Forest University

James C. Anderson
J.L. Kellogg Graduate School of Management

ISBM Report 151995

Teaching note for academics available from James Narus

Institute for the Study of Business Markets The Pennsylvania State University 402 Business Administration Building University Park, PA 16802-3004 (814) 863-2782 or (814) 863-0413 Fax

This publication is available in alternative media on request.

The Pennsylvania State University is committed to the policy that ail persons shall have equal access to programs, facilities, admission, and employment without regard to personal characteristics not related to ability, performance, or qualifications as determined by University policy or by state or federal authorities. The Pennsylvania State University does not discriminate against any person because of age, ancestry, color, disability or handicap, national origin, race, religious creed, sex, sexual orientation, or veteran status. Direct all inquiries regarding the nondiscrimination policy to the Affirmative Action Director, the Pennsylvania State University, 201 Willard Building, University Park, PA 16802-2801; tel. (814) 863-0471; TDD (814) 865 3175. . U.Ed. BUS 96-025

Microsoft Corporation: The Design of Microsoft@ Support Network 1.0

James A. Narus and James C. Anderson* August 8, 1995

*James A. Narus is a Babcock Research Professor and Associate Professor of Management, Babcock Graduate School of Management, Wake Forest University. James C. Anderson is the William L. Ford Distinguished Professor of Marketing and Wholesale Distribution, and Professor of Behavioral Science in Management, J. L. Kellogg Graduate School of Management, Northwestern University. He is also the AT&T ISBM Research Fellow at the Institute for the Study of Business Markets (ISBM), located at Penn State University. The authors gratefully acknowledge the financial support of the Babcock School s Research Program and the ISBM. Please note that this case was prepared as the basis for class discussion rather than to illustrate either effective or ineffective handling of an administrative situation. Some of the information presented has been disguised. Send correspondence to: James A. Narus Babcock Graduate School of Management Wake Forest University 7659 Reynolda Station Winston-Salem, NC 27104 (910) 759-5417 (phone) (910) 759-4514 (fax) jim_narus@mail.mba.wfu.edu (e-mail)

Microsoft Corporation: The Design of Microsoft@ Support Network 1.0

Trish May, the Director of Marketing for the Product Support Services Division (PSS) of the Microsoft Corporation, decided to take a well-earned break from the time-consuming and demanding analyses she was doing. It was about 8pm on the evening of December 28, 1992; yet, she and most of her task force were hard at work at PSS offices in Bellevue, WA. Trish and her boss, Patty Stonesifer, the Division s Vice President, had a major presentation scheduled for January 20, 1993 to ten top Microsoft executives, including Mike Maples, Executive Vice President of Products and Steve Ballmer, Executive Vice President of Sales and Support. At that meeting, Trish was to unveil PSS proposals for revamping Microsoft s entire support services network. Although her analyses had revealed significant insights and provided guidelines for a solution, she was still unsure as to the make-up of her final recommendations. As she headed for the Division s coffeemaker for a cafe latte, she reflected upon the financial and marketplace forces that necessitated a change in Microsoft policy. Efforts to redesign Microsoft s support service network were set in motion early in November 1992. At that time, Patty Stonesifer had noticed a particularly troubling upward spike in customer service costs in her periodic review of the division s Profit and Loss (P&L) statement. She was also becoming concerned with the rapid increase in the number of technical support personnel, which exceeded 1500. Although there was no immediate threat, Patty projected that service expenditures would become an acute problem in three years. When she brought the matter to the attention of Microsoft Chairman Bill Gates and Steve Ballmer, they readily agreed that something needed to be done. Bill Gates, in particular, had long advocated that Microsoft rethink its support service efforts and invest substantially in them to insure that the firm delivered high quality service to go along with Microsoft s high quality products. This appeared to be the appropriate time to review existing service policy and make that investment. Admittedly, Microsoft s support services were not as good as those offered by some competitors. Several factors contributed to the nondescript nature of Microsoft services. Previous support service policy had been determined at the product level. Annually, each product manager negotiated with PSS over the type, extent, and pricing of services to be offered to customers along with their products. Because Microsoft had the most extensive product line in the software industry, the result was a hodgepodge of service offerings. Some products had no support services, some offered unlimited free service that was accessed by phone via a toll number, and still others provided extensive telephone service for fee. For customers, particularly those that owned and used several Microsoft products, the service offerings were confusing because it was difficult to know which service came with which product. Moreover, expert users felt that they were paying for services they didn t need on basic applications. At the same time, they could not get sophisticatedsupport services on some of Microsoft s newly introduced line of highly technical advanced systems, even if they were willing to pay extra. Given the green light from top management to reformulate support service policy,Palty Stonesifer asked Trish May to take a temporary assignment as the leader of a task force that was chartered to investigate the matter. E&$nning on December 1, 1992, Trish and her task force put in long hours on the project. They worked closely with technical, business analysis, and marketing personnel from PSS and other Microsoft Divisions. In the process, they ran and evaluated countless financial and service request scenarios on Excel@ spreadsheets, conducted several focus groups and customer surveys, and evaluated competitive offerings. Furthermore, they relied on frequent discussions and candid debates with Microsoft managers from other divisions to sharpen their thinking and logic. Sipping her cafe latte, Trish pondered the numerous and often conflicting alternatives from which she would have to craft a comprehensive solution. Moreover, she wondered what implementation issues would have to be overcome to insure the success of the resulting support service network. It would be a late night.

The Microsoft Corporation Microsoft is the leader in the microcomputer, software industry. In fiscal year 1992, net revenues topped $2.7 billion, while net income exceeded $708 million (see Exhibit 1 for financial information). Its revenues almost matched the combined totals for firms ranked 2 to 10 in the industry. Furthermore, Microsoft was a truly global competitor. It operated in 41 countries around. the world and over 55% of 1992 revenues came from outside the US. The story of Microsoft s success and the career of its founder, Bill Gates, are both inseparable and legendary in business. As a student at Harvard, Gates rewrote Kurtz & Kemeny s BASIC for the MITS Altair. In 1975, he founded Microsoft. At that time, Gates articulated Microsoft s mission as, to make the software that will permit there to be a computer s first major break came in 1980, when on every desk and in every home. Perhaps the company IBM asked Gates to design an operating system for its personal computer. The product, MSDOS was introduced in 1981 and quickly became the standard operating system. During the 198Os, the company developed an assortment of popular applications software including Word, Excel, and Works. Another critical success came with the introduction of an operating system shell named Microsoft@ WindowP, that was designed to facilitate multi-tasking (i.e., simultaneous use of several programs) and to dramatically increase personal computer (PC) ease of use. By the early 199Os, sales of Windows 3.1 skyrocketed and it too became standard for PC users. Not only did Windows spawn additions to the Microsoft product line including Windows NTrM, WindowsTM for Workgroups, and Windows NTTM Server, it also prompted other firms to adapt their own software and/or develop new programs to run in a Windows environment. Products & Competitors, Segments, and Service Usage Patterns Products & Competitors Products targeted to business and institutional customers could be grouped into five major lines as a function of how and where they were used in the customer s computer system (see Exhibit 2 for brief summary of products). Application software performed specific tasks (e.g., word processing, spreadsheet analyses, presentation slide preparation) for users, typically at a desktop. PC operating systems software, on the other hand, provided an essential interface between the PC and software that allowed the computer to run applications programs. These programs could function on a desktop PC or off of a network server. Advanced systems, notably servers, permitted organizations to network large numbers of PCs, access large databases, undertake multi-tasking, and utilize high-end, technical programs (e.g., computer-aided design). Development products included programming languages and kits to help and encourage other companies to develop software that was compatible with Microsoft products and/or could run in a Windows environment. These entailed programs such as the Windows Software Development Kit, C/C++, Visual C++TM, Visual BasicTM, and Basic. Lastly, hardware, referred to a very narrow line of data input and control devices that enhanced PC usage and effectiveness. Because of this diverse product mix, Microsoft faced a myriad of competitors. In applications software, Microsoft competed against firms such as Lotus, WordPerfect, Borland, and Claris, among many others. IBM s OS/2 and Apple s System 7 (for Macintosh computers only) were viewed as principal operating systems competitors. And in the server marketplace, Microsoft was pitted against the likes of IBM s OS/2 server, Apple s MAC-X server, and a multitude of UNIX servers (e.g., SUN Solaris and Hewlett-Packard s HPLJX). As contrasted s broad line made it with most software houses that tended to be specialists, Microsoft vulnerable to flanking attacks from many sources.

3
Major Market Segments

By late 1992, Microsoft had an installed base of tens of millions of users worldwide. The firm divided its diverse customers into three market segments -- end-users, corporate accounts, and developers. End-users were those individuals that routinely applied Microsoft solutions on their desk-top computers to complete specific tasks. The category covered a broad range including: consumers, home businesses, small firms, resellers, and help-desks (i.e., inhouse, management information system or MIS professionals who provided assistance to company employees). However, the firm divided end-users by technical proficiency into two groups -- novices and experts. Both groups were likely to use PC operating systems and hardware. Novices, though, more commonly owned basic applications (e.g., Microsoft Office), while experts relied more heavily upon advanced applications. scale information processing requirements and thus, broad and highly technical software needs. They were prospective buyers of all Microsoft lines, except the most basic applications. Developers were firms that designed, marketed, and/or furnished technical support for software packages. They included other software companies (e.g., Lotus, WordPerfect), valueadded resellers (VARs), independent service vendors (ISVs), and firms that developed software for their own internal use. They were prospects for all Microsoft products. Service Usage Patterns1 Microsoft s ABC Costing and service monitoring systems provided a wealth of information that would be useful to the task force. Some key findings included the following: l Over 80% of all service calls came within go-days of software acquisition. l About 71% of all customers never called. Of those that did, most called only 1 or 2 times. More than 80% of all calls were generated by about 15% of all callers. Frequent callers needed highly customized and personalized support. l Less than 15% of all callers required extended hours of service availability (i.e., more than Bhours per day). About 5% of all callers needed 24hour service availability. Costs varied as l The cost for all support services averaged about 9% of retail dollar sales. follows: Office Family of Products (5%), Advanced Applications (5.5%), PC Operating Systems (7.5%), Hardware (8%), Developer Products (lo%), Advanced Operating Systems (12%), and Servers (12.5%). Table 1 provides more detailed forecasts of costs for 1993.
Corporate Accounts included businesses, institutions, and governments that had large-

Marketing & Technical Support Systems

Responsibilities for planning and implementing marketing activities for each of Microsoft s 150 software products were held by product managers and their assistants. Importantly, their jobs entailed the coordination of all promotional, advertising, and sales efforts for their respective products. In addition, they were actively involved in product line development and/or modification. Traditionally, each product manager determined the extent and price of technical support service that would be offered along with his or her product and then worked closely with the technical support division to insure its delivery. Product managers were evaluated and compensated, in part, based upon the market share and the operating profits of their respective products. Microsoft had long distributed its software through a two-tiered marketing channel that included approximately fifty software distributors and over 3,000 software resellers or dealers across the US and Canada. The firm considered this channel to be an essential ingredient in its success and offered a comprehensive and substantial program of reseller support, training, and incentives. Due to the size of its network, Microsoft reported little problems with product
lNote: The information presented in this section has been disguised to reflect industry averages. It does not necessarily represent actual usage patterns and/or costs for Microsoft technical services.

4 availability. For the most part, resellers were eager to stock Microsoft products because of their widespread popularity and because they provided a platform on which resellers could bundle other products and services; thus dramatically improving their profitability per transaction. Microsoft also sold its products through direct marketing efforts. These customers were most often located in geographic regions, particularly internationally, where there were no resellers. The Microsoft sales force included account representatives (AR@, systems engineers, and district managers who were assigned to seventeen geographic districts across the United States. Each AR was responsible for Microsoft s entire product line and made periodic calls on resellers, corporate users, developers, and select end-users in their territories. ARs also made presentations at regional computer shows and gave training seminars at reseller stores. Their tasks were largely evangelical in nature; they introduced, promoted, and spread the word about Microsoft software, without taking orders. Instead, they directed prospective customers to the reseller network. ARs could take orders for technical support contracts that were fulfilled by the Microsoft Corporation. They also had the option of directing customers to third party service providers. Each AR was paid a salary plus a bonus based upon reseller sales of Microsoft products within his or her geographic territory and/or assigned accounts. Systems engineers complemented the efforts of ARs by providing technical pre-sales and evaluation support to prospective and existing customers. The Product Support Services Division of Microsoft furnished all technical support services to customers. The division was headed by a Vice President. General Managers for each market segment (i.e., Developers, End-Users, and Corporate Accounts) as well as Directors of International Product Support and Marketing reported to the Vice President (see Exhibit 3 for PSS Organization Chart). Approximately 1,500 support engineers staffed the division in the North America and another 900 provided service in 36 countries around the world. North American service engineers were deployed at three sites - Bellevue, WA, Charlotte, NC, and Irving, TX - to facilitate uniform market coverage across time zones.
.

The Forces of Change

By the early 1990 s, changing marketplace conditions had caused all software producers to re-examine and reformulate their technical support offerings. Four key drivers were at work - declining financial performance, a fragmenting customer base, competitive initiatives, and customer demands for greater simplicity in service offerings. Financial Performance Although many industry participants hated to admit this, by 1990, many software products had reached the mature phase of the life cycle. As a consequence, word processing and spreadsheet programs which had once sold for in excess of $795 were being priced well below $149.2 With shrinking margins, producers had no choice but to scrutinize all costs. One prime target for cost reduction was support service. With the average cost of support services for high-end products exceeding 10% of dollar sales, industry pundits opined that the practice of offering unlimited, free service was doomed.3 At Microsoft, a defining event came on July 1, 1991 when an Activity Based Costing (ABC) System was implemented. The system had taken two years to develop. It tracked service usage and associated costs by product. These costs, in turn, were charged back to their respective products. For the first time, product managers found the expense item, Product Support Costs, on their monthly operating income statements. Many reacted immediately to a

2Schneider, Jerry (1993), Vendors Create New Fee-Based Options to Meet UsersGrowing Needs, PC Week, 10 (November 15), 123. 3Benchmark Report (1993), in Soff-Letter, Jeffrey Tarter, ed., (!%ptember 21), 1.

5 new cost item that served to reduce their bonuses, calling for a re-examination of support efforts. Some even demanded that all support services be eliminated. Customer Base As was the case in other industries, the marketplace for software was becoming more s extensive product line, customers ranged across a and more fragmented. Given Microsoft broad spectrum from home users to global corporations. Each group not only wanted more customized software but also more customized technical support. Whereas most industry analysts dwelled on the potential ill-effects of charging endusers for support service, Microsoft managers were becoming more concerned with the growing dissatisfaction of corporate accounts and developers with the availability and quality of highend support. For instance, callers to several well-known software firms free 8OO#s reportedly got a busy signal 60% of the time. When they did get through, these corporate customers had to wait in a telephone queue for an average of 10 minutes, often behind home users who had very simple questions. What made things even more frustrating for these customers was that they were willing to pay more for highly responsive assistance on missioncritical problems. Microsoft managers expected that demands for high-end service on its own network would become even more acute in the next few years with the introduction of a new generation of technically complex products including Windows NT and Windows NT Server. (See Exhibit 4 for sample profiles of corporate prospects for high-end service.) The nature of technical support service requested was also changing. In the past, when customers only bought one or two packages, it was relatively easy for support engineers to understand and resolve problems on a product-by-product basis. In 1992, the sophisticated user was more likely to utilize a greater variety of software programs and more complicated systems that were purchased from various vendors. More than ever before, the entire system had to be fully integrated to function properly. For example, applications software had to be compatible with the operating system and the network server. This made trouble-shooting very difficult because a given problem might originate in a number of places. Not only did each service engineer have to understand the workings of all 150 Microsoft products, he or she had to have first-hand knowledge of all competitorsproducts and of every industry-specific product application. Clearly, Microsoft would have to make hard decisions and restrict its service offerings to a subset of problems that its personnel could productively and competently resolve. The Competition The industry as a whole was beginning to come to grips with the issue. Although most competitors continued to give technical service for free via the telephone, several firms -Lotus, Borland, and Aldus - had initiated service charges as early as 1991. It was clear that these firms were experimenting with a variety of offerings and charges. Lotus in 1992, for example, provided unlimited and free service for 90 days following a customer s first service call. After that time period, the customer could choose between paying $129 for an annual contract or paying $2 per minute for each service call. Borland s services revolved around the nature of customer problems. It chose to offer an 800# and free service only for installation and start-up difficulties. For all other problems (e.g., usage and custom design), customers could select from several service and payment options: 1) $2 per minute on an 800# for each service call, 2) an annual service contract for application and language products for a fee of $129, or 3) a premium and top-end service contract for $2,000 for each product with a maximum of $5,000 for four applications. Interestingly, it was reported that Borland was using its service contracts tactically, throwing them in for free at the last minute to close deals. The Aldus Corporation, a producer of layout, illustration, and presentation software, offered seven levels of customer service. CustomerFirst service entailed free technical support

via a toll number for the first go-days following purchase for all new products. A basic support package, CustomerFirst Basic, included toll-free technical support and product discounts for an annual fee of $99. A comprehensive electronic support service, CustomerFirst Online, included technical support, information, and other resources. Access via modem was available 24-hours a day. Annual fee for this service was $249. CustomerFirst Premier Online combined the previous two services at an annual fee of $429. CustomerFirst Corporate Library was top-end, customized support for large customers. Charges were not posted but determined as a function of what needed to be done. Finally, CustomerFirst 900 was a pay-as-you-need, 900#, telephone support option that went for $15 per call. Competitors were relying upon a variety of approaches to charge customers for telephone service calls. Some sent an invoice following the call. Others charged via a 900# or had service engineers take credit card numbers. Alternatively, some competitors either included in software packages or sold separately incident coupons that entitled the bearer to make a number of pre-specified technical support calls. Adobe Corporation took a different approach. They provided customers with service credits as a function of the dollar value of Adobe software that they purchased. The more software purchased the more credits received. When needed, the customer could redeem service credits for technical support. For the most part, software suppliers offered low-end, telephone support service by phone. There were exceptions. For example, Novell, which marketed premium-priced and highly technical operating systems and networking products, provided top-end telephone support for around $150 per hour. However, most of their problem-solving services were furnished on-site by third party service vendors called, Certifid Netware Engineers (CNEs). CNE fees varied widely, ranging between $100 and $500 per hour of assistance. Clearly, there was considerable opportunity for the development of high-end technical support.
.

The Need for Simplicitv Prior to late 1992, Microsoft did not have a consistent support service offering. Instead, service levels varied by product from no service to premium service. Neither service levels nor fees appeared to be correlated to the technical complexity of products. And, the service rendered was of adequate quality. The problem was underscored by a 1992 marketing research survey in which customers, particularly those that used several Microsoft products, expressed frustration and confusion over Microsoft s technical support services. The sales force simultaneously reported difficulty in communicating the service offering to prospective customers. Obviously, a more comprehensive and consistent technical support offering was needed. And, it had to be simple enough to communicate, understand, and execute.

Development and Consideration of Alternatives

Prior to formation of the PSS task force in December 1992, Patty Stonesifer, Trish May, and key team members, along with the input of Bill Gates, Steve BaUmer, and Mike Maples, devised a set of goals that were to guide efforts and serve as the basis for all technical support programs. The goals that emerged can be summarized as follows: Microsoft technical support services should reinforce and enhance the high quality of Microsoft products. Bill Gates insisted on this goal. He maintained that Microsoft needed to deliver a comprehensive offering to the marketplace that was comprised of both high quality products and support services. Importantly, Gates was willing to invest significant resources into technical support services to improve their overall quality and customer perceptions of them.
l l

All technical support programs should be characterized by three attributes - Choice, Quality, and Simplicity.

7 Based upon several marketing research studies and experience, the task force concluded that customer satisfaction would be sustained by delivering technical support that met customer needs in three ways. First, support programs should offer customers a choice. Those that did not want to pay for support they didn t need should be able to buy software alone, At the same time, customers that wanted more at a price that did not include service costs. support and were willing to pay extra for it could also do so. Second, all technical support T h i r d , a l l provided by Microsoft should meet or exceed industry standards for qualitv. technical support programs should be marked by simplicitv; that is, customers should have little difficulty understanding Microsoft s offering and how it applied specifically to their needs. Technical support services should be a cost center for the Microsoft Corporation. On this point, there was uniform agreement in the company. As managers saw it, Microsoft s core competence was its technology. Though important, technical support services were not seen as a potential source of corporate profits. Instead, Microsoft was prepared to invest in them in the short run, with the eventual goal of breaking even.
l

The Microsoft Corporation should be the technical service provider of last resort. Microsoft managers realized that they could take progressive steps to leverage the work of PSS engineers. First, continuous efforts were to be directed at making all software as userfriendly as possible. Resulting reductions in user problems would lower demand for service. Second, special efforts were to be initiated to give customers tools and information, so that they could resolve as many problems themselves as possible. Third, customers were to be informed about and encouraged to use third party service providers (e.g., VARs, dealers, ISVs) in their Finally, if the first three methods proved ineffective, then communities and industries. customers were actively directed toPSS personnel.
l

Technical Support Service Alternatives Early on in their work, the task force hit upon the idea of using a matrix to summarize the various service offerings. Members referred to it as the Microsoft Support Network. The rows would consist of major service groupings, while the columns of the table would contain significant product or market segment groupings. In turn, each element (i.e., cell) in the matrix would describe a service offering and include a fee structure. To avoid customer confusion, the group concluded that 4 rows and 4 columns (i.e., 16 elements) should be the largest size of the matrix; however, no research had been done to confirm this. Although the concept of a service offerings matrix received enthusiastic and universal approval among Microsoft managers, there was widespread disagreement over the make-up of the rows; columns, and elements. Trish noted that the matrix could be constructed in several ways. Product or Market Segment Groupings As Trish saw it, service offerings (i.e., rows) could be devised either around major product lines or key market segments (i.e., columns). In the former case, services would _be deployed as a function of products owned, while in the latter, by customer groupings. Product Groupings. Analyses of service usage patterns showed considerable similarity within and differences between a handful of product clusters. Initially, the task force proposed four groupings - Desktop Applications, PC Operating Systems (along with hardware), Development Products, and Advanced Systems. This breakdown appeared to make sense because as one moved across the categories, the value, cost, and complexity of service increased dramatically. In particular, product line service costs clustered tightly around these four categories. However, the decision concerning product groupings was politicalIy charged. In general, product managers for best-selling software wanted the groupings to be built around their individual products or at a minimum around the product lines given in Exhibit 2. And, some product managers raised legitimate concerns. For instance, it was proposed that Access and FoxPro software be placed into a fifth category, Data-Base Management Programs.

8 Proponents argued that data-base marketing was a large and growing business activity that required highly specialized and integrated support. They did not want it to be considered just another application like word processing. Trish observed, however, that Access and FoxPro service costs were very similar to those of other applications and that data-base management problems became severe only in those instances where firms had customized FoxPro (the highend product) to their particular requirements. Such needs would have to be addressed with high-end, service contracts. Finally, product managers for niche products such as hardware feared that their lines would get lost in the shuffle. These items didn t fit well with other products; yet, they were not important enough to merit a separate grouping. Market Segments. It could also be argued that service requirements varied as a function of the type of customer. Four segments could be readily identified -- Novice End-Users, Expert End-Users, Corporate Accounts, and Developers. Novice end-users rarely called. When expert end-users called, their technical service requirements overwhelmingly focused on very. general, installation and usability issues. End-user problems tended to be application-specific, document-specific, and function-specific. For the most part, these difficulties were a source of inconvenience rather than a job-threatening event. In fact, many end-users could address their own problems if they read the program manual that came with the software or a commercially available how to book. Critically, end-users were highly price sensitive. And, most learned to use software when technical support was expected to be unlimited and free, forever. Corporate accounts were more likely to have mission critical environments where systems failures and down-time could have catastrophic effects on both costs and customer service. As a consequence, corporate accounts demanded responsiveness, for example, in the form of disaster recovery service. As contrasted with end-users, their support concerns gravitated around usage rather than installation problems. Moreover, these problems could be due to the use of diverse hardware and software, and unique applications. Obviously, such problems were difficult and time consuming to resolve. Fortunately, because corporate accounts were use to an IBM mainframe environment, they fully expected to pay extra for third party installation and technical support contracts. Moreover, they were not as price sensitive. Developers needed highly technical assistance on applications programs, operating systems, servers, and programmin g languages. Software companies, for instance, might need assistance in adapting their proprietary software to Windows standards. ISVs, on the other hand, might have problems applying Microsoft C/C++. Alternatively, large corporations might have difficulties with more complex problems with Visual Basic,FoxPro, and advanced servers. For developers, responsiveness and resolution of difficult problems were of paramount importance. Because developers commonly pass through technical support costs to their own customers, they were the least price sensitive. Responsiveness-Based or Problem-Based Service Groupings At the same time, Trish believed that service offerings could be based upon either the speed of response or the type of assistance provided. In the first case, customers would-pay for responsiveness; while in the second, they would pay for specific problems solved. Responsiveness-Based Services. While several competitors, most notably Borland, had created offerings based upon type of software problem, the task force thought that _ responsiveness might provide a more strategic foundation. To begin with, responsiveness was a key value-driver for major customers. These firms were highly dissatisfied with the level of free service they were receiving from other software houses, disliked waiting in queues behind novice end-users, and were willing to pay a premium for responsiveness. Furthermore, given the complexities of computer systems, it was becoming very difficult to diagnosis, let alone forecast, the source of problems. It might be easier for both customers and Microsoft, if service was sold as an insurance policy in which Microsoft promised to give priority to certain accounts and address anv and all problems within specified times. With this notion in mind, the task

9 force had initially proposed four responsiveness-based service categories -- Premier Support, Priority Support, Standard Support, and F&Tips & Electronic Services. Premier support would be offered primarily to corporate and developer customers whose computer systems operated in mission-critical environments where system failure and downtime could have catastrophic consequences. Mission-critic al environments were of ten found in investment banks, hospitals, warehousing and logistics operations, and large-scale computer-controlled manufacturing. Almost always, mission-critical applications required customized software and support service. Premier customers would be given top-priority when problems developed. Each account would receive highly personalized attention from an permanently assigned team of the best service engineers. Because of their distinct requirements, corporate and developer accounts would subscribe to separate premier programs. And, for reasons of practicality, they would receive service for all of Microsoft s products. Priority support would be directed primarily toward a second group of corporate and developer accounts, those that did not have mission-critical requirements yet wanted reasonably prompt response. Although they were willing to pay extra, they were more price sensitive and many wanted to buy service only for a subset of Microsoft product lines. For this reason, they would be allowed to choose between: Priority Comprehensive Support which covered all products; Priority Desktop which only addressed applications, PC operating systems, and hardware; and Priority Development with Desktop that included development products, applications, and PC operating systems and hardware. Although all customers claimed to want their problems addressed promptly, many would tolerate delays and wait in a queue if the service was free or offered at a nominal fee. This was particularly true if they had either minor problems or those that could be considered an inconvenience. Many problems that occurred within the first go-days of ownership were prime examples. End-users, particularly those that owned desktop applications were likely to want this so-called standard service. Because the nature of PC operating system and hardware, and development products problems tended to get very complex after 90-days of ownership, limiting the amount of standard service to 90days appeared to make sense. At the same time, given the complexity of advanced systems and their history of paid service fees, it did not appear logical to offer standard service to owners of advanced systems. FastTips & Electronic Services would provide commonly-requested information in several electronic formats. These would include: an automated phone system that provided callers with a digital menu of topics from which to choose, bulletin boards on the Internet or other on-line services (e.g., CompuServe@ and America On&n&), CD-ROMs, and Microsoft data-bases that could be accessed via modems. They could be accessed 24hours a day. Problem-Based Services. Recently, Microsoft had implemented an Activity-Based Planning System that monitored the types of software problems that customers encountered. Findings, in turn, could be used to plan service offerings and staffing requirements. The system pointed to several types of assistance that would be required by customers to solve particular problems - General Information, Installation & Start-Up, Usage & Productivity, Systems Integration, and Customization. To reduce the number to four, Usage & Productivity and Systems Integration could be combined. The most basic service needed was access to general information. It would address questions such as: Does Microsoft sell a project management program?, When will the next 7 What value-added resellers market turkey systems for the upgrade of Windows be released., 7 and What quick tips can you give me to banking industry that rely on the Windows NT Server., improve my Excel skills? Ail customers needed general information to some extent. Importantly, such inquiries could be anticipated and stock answers for them could be prepared. The fact that over 80% of service calls came within go-days of software acquisition suggested that installation and start-up assistance was important to many customers and in

10

particular, end-users. On the other hand, developers and corporate accounts experienced these problems less often because their software was typically installed by professionals. For the most part, installation problems related to application software, PC operating systems, and development products were routine in nature and could be resolved quickly and effectively. Due to their complexity, advanced system installation and start-up problems tended to be far more difficult and time-consuming to address. For this reason, advanced system customers had always been charged for installation and start-up assistance. Assistance with software usage and productivity problems represented the next type of service. All customers had these problems to some extent; however, they were more critical to developers and corporate accounts. In the case of application software, usage problems tended to be specific to tasks, documents, and functions. For example, a customer might be having difficulty printing out a document. III general, these issues were very basic in nature and readily managed by service engineers. PC operating system, hardware, and development product problems associated with usage and productivity tended to fall into two categories. Those occurring the first 90 days of ownership revolved around basic functions and could be handled quickly. On the hand, those problems that occurred after 90-days, could often be attributed to either interactions with other software or difficulties with advanced features. They were far more challenging for service engineers. Advanced systemsusage and productivity problems were very difficult to address and customers had always been charged for their resolution. Owners of advanced systems (i.e., corporate accounts) were most likely to encounter systems integration problems. This was due to the fact that their computer systems networked numerous user sites, were comprised of hardware from multiple vendors, and ran diverse software from numerous software houses. Depending upon the nature of the software being designed, users of development products might also need this type of assistance. This was particularly true in the case of firms that were tailoring their products to a Windows environment or creating advanced systems. To a lesser extent, PC operating systems and hardware owners might experience systems integration difficulties due to their use of diverse programs and hardware. End-users that ran applications on individual PCs were the least likely to have any of these concerns. In general, a more experienced and technically competent service engineer would be needed to resolve this category of problems. For this reason, the cost of providing service in these situations could be quite expensive. Customized assistance represented the topend of technical support for those corporate accounts and developers that demanded extraordinary treatment. Conceivably, such assistance might be needed for any of Microsoft s products. However, the best prospects for top-end support needed to have applications or advanced system software customized to meet the demanding needs of their industries and businesses. Examples might include a direct-mail catalogue firm that had to have an extensive data-base linked to its order-Mfillment system, and an investment bank that had to have access to current transactions on major stock and bond markets and the ability to route such information to hundreds of managers for analysis. Given the extensive systems they employed, these customers would need to be assigned to a team of highly trained and experienced service engineers. A long-term relationship would have to be established with the team. In addition, whenever problems arose outside of the team s expertise, they would have to be able to obtain immediate help from Microsoft engineers who had knowledge in that area. Because of divergent requirements, two types of high-d support had to be offered - one for corporate accounts and another for developers. Service Deployment. Reviewing the list of potential services, Trish wondered which should be offered as standard for free, which should be sold as an option for a fee, and which should not be sold. Given the relatively low costs involved, a strong case could be made that General Information or F&Tips & Electronic Services should be offered for free. Because these services could be accessed by anyone, they would also generate goodwill and spark interest in Microsoft products among prospective customers.

11 At the same time, industry pundits speculated that it would be impossible to eliminate free Insfailation & Start-Up service on application, PC operating systems, and developer products. Customers would interpret fees for this period as an unethical way for software vendors to pad their profits. Many would wonder if suppliers deliberately made their software difficult to use so that customers would have no choice but to buy service.4 Usage & Productivity assistance would be a tougher call. While most problems associated with application software could be addressed at low cost, those for PC operating systems, development products, and hardware could be costly; particularly those that occurred after the first 90-days of ownership. Clearly, Systems Integration and Customized support services would be quite expensive to deliver. However, the task force did have a choice. They could include them as standard and cover their costs via a hefty price premium on software. Alternatively, they could sell them as an option at a significant fee. If services were offered as a function of responsiveness, fee decisions would be a little more straightforward. Standard Support would probably be offered free on an unlimited basis for desktop applications. Because usage and productivity questions after 90-days of ownership could get costly, it would probably be necessary to charge personal operating systems, hardware, and development products owners a fee after 90-days. Priority and Premier Supporf would come with significant charges due to the staffing needed to provide them. A series of intriguing decisions would have to be made concerning what services not to offer. Given that most applications and PC operating systems software was being preinstalled by original equipment manufacturers (e.g., IBM, Compaq, Dell), it could be argued that installation and start-up problems should be handled by them. At the same time, a valid case could be made for having local computer dealers handle basic usage problems, particularly those that occurred within 90-days of purchase. As for the highad services, the firm would have to make some difficult choices. For example, in the case of cross-platform or integrative problems caused by another vendor s products, should the problem be handed-off to the 7 There would also be limits to the amount of customized vendor responsible for the problems. and mission-critical service that could be offered. At some point, it made more sense to direct customers to VARs or ISVs that specialized in an industry. If Microsoft decided to hand-off these problems to 3rd parties, it would have to determine which firms would provide the appropriate services and how these firms would be compensated for doing so.

Conclusion
Returning to her office, Trish settled back into her desk chair and re-opened her project file. She had a clear idea of what the alternatives were. It was now time to evaluate their merits and propose a comprehensive solution. In addition, she would have to address some thorny implementation issues. For instance, how would fees be structured? Obviously, timing might be one way, with 90-days as an initial free period. Alternatively, customers might pay by the incident or buy an annual contract. Communicating the details of the Microsoft Support Network 1.0 would also be a challenge. Above all, Trish did not want customers to respond to the new program by saying, You mean I have to pay for something I used to get for nothing? If this happened on a widescale basis, sales could suffer. Instead, she wanted them to delight in their ability to choose from a variety of high-end services while concluding, I don t have to pay for what I don t need! It was not at all clear how these communications goals would be achieved. Although these decisions would be difficult to make, Trish felt inspired by the knowledge that the efforts of her task force could gain a competitive advantage for Microsoft and influence an entire industry. She looked forward to her presentation on January 20th.
4Tarter, Jeffrey (1993), Point-to-Point: Paying for Support, Computer Reseller News, (November 29), 78.

Exhibit 1 A SUMMARY OF MICROSOFT S FINANCIAL PERFORMANCE: 19894992

(In millions of dollars except employee and per share data)

1989

Year Ended Tune 30 1990 1,183 253 930 181 317 39 393 17 410 131 279 533 1,105 919 1.04 5635 1991 1,843 362 1,481 235 534 62 650 21 671 208 463 735 1,644 1,351 1.64 8226 1992
2,759

Net Revenues Cost of Revenues Gross Margin Research & Development Sales/Marketing General & Administrative Income from Operations Nonoperating Income (loss) Income before Tax Income Tax Net Income Working Capital Total Assets Stockholders Equity Net Income per Share Number of Employees

804 204 600 110 219 28 243 8 251 80 171 310 721 562 0.67 4037

467 2,292 352 854 90 996 45 1,041 333 708 1,323 2,640 2,193 2.41 11542

Source:

Microsoft s 2 993 Anntral Report

Exhibit 2 Major Product Lines

I. Application Software: For more serious users, businesses, l Microsoft Office Familv of Products for Windows. and institutions, Microsoft offered a line of professional products (e.g., the word processing program Microsoft Word, the spreadsheet application Microsoft Excel, the presentation graphics program Microsoft PowerPoint@, and a very basic data-base management program Microsoft Access@).
l Advanced Applications for Windows. A portfolio of more sophisticated and high-end application programs were targeted primarily at business users. These included several versions of a project management program, Microsoft Project, and of an advanced database management system, FoxPro@.

II. PC Operating Systems Software: Microsoft marketed several widely-used operating l Microsoft PC Operating Svstems. systems including MS-DOS and the shell (i.e., it functioned with MS-DOS), Microsoft Windows that typically ran off of desktop PCs. Another shell for MS-DOS, Windows for Workgroups, was designed for simple networks comprised of a limited number of desktop computers, printers, e-mail systems, and fax modems. III. Advanced Systems Software:
l Advanced Operating Svstems. Among these was Windows NT., an operating system for client-server computing and a workstation solution for business-critical computing. l Servers. Windows NT Server allowed firms to increase the speed with which large da= could be accessed and multi-tasking could be run from a large number of networked desktop computers and workstations without the use of a mini-computer based, central server. Microsoft SQL Server was another highly sophisticated network server that enabled companies to transfer data from large-scale databases across their organizations.

IV. Development Products:

l Development products, included programming languages and kits such as the Windows Software Development Kit, C/C++, Visual C++T, Visual BasicTM, and Basic.

V. Hardware:

l Hardware items included the Microsoft BallPoint@ Mouse and several versions of the Microsoft Mouse.

Exhibit 3 Organization Chatt Product Support Services Division

WCC President

EUSfnfqkStrvkts
hit hlJ~Ct

. KS Wtsmn Rtgba
WA Mmrga

KS Emkrn Rtgbo
Mf Mmagct & L

PSS Ctnld Rtgion

TX Mmrgct

Exhibit 4 Profiles of Potential Customers for High-End Service Contracts The following profiles identify the types of customers who are ideal prospects for highend service contracts. Each profile describes the benefits that each prospect would gain from the contracted service. Customer Profile #l: An oil company headquartered in the US has global operations for petroleum refinement and distribution. Its ten subsidiaries are currently responsible for their own hardware and software technical support. In total, the company has over 50 thousand, networked personal computers (PCs). The Microsoft Office suite of applications and Windows NT servers are being installed companywide. The firm is also considering converting to Microsoft Mail across all of these PCs. With a high-end service contract, the oil company would save money by not having to hire a consultant to coordinate the initial migration of its users to Microsoft Office. Moreover, the service would immediately support the firm s helpdesk. As a consequence, training time for internal service personnel would be reduced and employees would not have to wait for technical support. If the firm selected a contract that designated a technical account manager, the mean time for repairs would also be reduced. This would be due to the designated technical account manager s more thorough knowledge of the account s profile, installed base, support history, and recurring technical problems. Customer Profile #2: A large aerospace firm with over 75,000 PCs is trying to establish Microsoft Office as the standard software for those users. Management would like the applications to be installed on Windows for Workgroups workstations. The firm is attempting to establish a standard to facilitate the ready use of programs, document sharing among employees, and PC support. A high-end service contract would help senior management sell the idea of a companywide software standards because it would simply existing practices. For instance, it would reduce service costs by eliminating the need for multiple software support vendors. As with Profile #l, the service contract would smooth the transition to Microsoft applications. As part of the contract, Microsoft personnel could maintain a historical record of past and future support interventions. Additional funds could also be saved through the design of standard company forms. Finally, down time could be reduced if Microsoft can identify common problem areas and leverage technical solutions across the entire organization. Customer Profile #3: One of the largest financial institutions in the US operates some 30 thousand PCs. It is in the process of converting all of its customer services (e.g., new accounts, loans, teller line applications) to PCs. Both customized software and Microsoft Office applications will be run on Windows NT. All 30 thousand PCs will have access to these programs. Because of the tremendous amount of data that will have to be moved across the network to a myriad of locations, a high-end service contract will be critical to reduce downtime and to insure that applications work properly every time. Moreover, the 24hour mission critical support reduces the risk of system outages during the late evening and early morning hours. Importantly, Microsoft service can insure that each application runs smoothly on Windows NT and that all programs on the system are integrated. And, Microsoft personnel can work with company developers to reduce software development cycles. Microsoft will also provide access to microcodes and sample data for the software design process.

Table 1
Forecasted 1993 , Costs per Service Call and Dedicated Service Engineer (SE) Accountsa Forecasted Service Costs by Product Categories
Desktop . &nlicatiom On* Developt 7 million 0.16 1,120,000 3,809 294 $95,000 $27,930,000 $25 12 million 0.14 1,680,OOO 3,750 448 $95,000 !$42,560,000 $25
b

1993 Forecas Unit Sales Service Calls per Unit Sold Total Service Calls Service Calls per SE per Year SEs Needed Average SE Cost Total SE Cost Average Cost per Call

P rodu& Adced Svstems 250 thousand 0.1 25,000 1000 25 $150,000 $3,750,000 $150
b

500 thousand 0.2 100,000 1,315 76 $125,000 $9,500,000 $95

Forecasted Costs per Accounts Needing Dedicated SEs

Accounts Needing Dedicated SE d # of Accounts per Dedicated SE SEs Needed Average SE Cost Total SE Cost Average Cost per Account 500 8 63 $160,000 $10,080,000 $20,160

Note: The information presented in this table has been disguised and does not reflect actual Microsoft Corporation service costs.
This simplified chart has been presented for discussion purposes. Actual Microsoft data analyses featured detailed regression and queuing models. Average SE Cost represents annual compensation, workstation, and problem-solving tools expenses. I A dedicated SE would be permanently assigned accounts and provide customization assistance and system integration support (e.g., Premier Support).