CVS Pharmacy Service Improvement PDF

5 -60 7-040 R E V: O C T O B E R 3, 2 00 7 TEACHING NOTE Pharmacy Service Improvement at CVS (A) and (B) Introduction This note is intended to assist instructors teaching the H...

5 -60 7-040 R E V: O C T O B E R 3, 2 00 7 TEACHING NOTE Pharmacy Service Improvement at CVS (A) and (B) Introduction This note is intended to assist instructors teaching the Harvard Business School case series, “Pharmacy Service Improvement at CVS (A) and (B)” (HBS No. 606-015 and 606-029). It gives a case synopsis, its teaching purposes, the analyses required to understand the cases’ issues and answer relevant questions, other teaching resources, a set of assignment questions for students, a proposed teaching plan, and a blackboard plan for the class. These cases are part of the course Managing in the Information Age (MIA); the course’s conceptual foundations are described in a course overview note for instructors (HBS No. 608-011) and its pedagogy is discussed in “Teaching the Course Managing in the Information Age: Note for Instructors on Pedagogy” (HBS No. 608-007). This case sequence is part of the course’s single-company enterprise IT module, which has its own note for instructors (HBS No. 608-008). The teaching plan for this case makes use of MIA’s principal capabilities-complements- responsibilities (CCR) framework. An abbreviated version of this framework is given as an appendix to this teaching note; the framework is more fully described in the course overview note. In this note the case’s teaching purposes and analyses are organized around the CCR framework, with additional sections devoted to case issues that are outside this framework. Case Synopsis In 2002, CVS assembled a team of operations executives and managers for its Pharmacy Service Initiative (PSI). Although CVS was one of America’s largest retail drugstores, problems in the company’s prescription fulfillment process had caused significant customer defection, hampering the company’s growth. The PSI aimed to streamline the fulfillment process. The fulfillment process had five steps. In the first step—the dropoff—customers dropped off a prescription and a pharmacy technician placed it into a wooden box that represented the customer queue. The next step, data entry, involved both an automated drug utilization review (DUR), which revealed any potential problems with the prescription, and an insurance check, which verified the amount of the patient’s copayment. In the production and quality assurance steps that followed, This note was prepared by Professor Andrew McAfee for the sole purpose of aiding classroom instructors in the use of “Pharmacy Service Improvement at CVS (A) and (B),” HBS Nos. 606-015 and 606-029. It provides analysis and questions that are intended to present alternative approaches to deepening students’ comprehension of business issues and energizing classroom discussion. HBS cases are developed solely as the basis for class discussion. Cases are not intended to serve as endorsements, sources of primary data, or illustrations of effective or ineffective management. Copyright © 2006, 2007 President and Fellows of Harvard College. To order copies or request permission to reproduce materials, call 1-800-545- 7685, write Harvard Business School Publishing, Boston, MA 02163, or go to http://www.hbsp.harvard.edu. No part of this publication may be reproduced, stored in a retrieval system, used in a spreadsheet, or transmitted in any form or by any means—electronic, mechanical, photocopying, recording, or otherwise—without the permission of Harvard Business School. 607-040 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) drugs were counted and reviewed by a pharmacist to make sure that all the details of the prescription were correct. Finally, in the pickup step, bags were sealed and stored in the pickup area in alphabetical order where customers could pick them up after payment. Problems for the customer could occur at almost every step. For example, during the data-entry process, problems with verifying insurance information might result in the customer being forced to pay the full price of a prescription at pickup, rather than the insurance plan’s dictated price. During production, insufficient inventory might prevent the prescription from being filled. The pickup step could also result in long wait times, especially if several customers in line had issues to resolve with their prescriptions. These service issues were among the problems that had caused an estimated 7.2 million regular pharmacy customers to leave CVS in 2000. In July 2002, at the close of the (A) case, members of CVS’s PSI team were trying to design a new fulfillment process that would retain the safety and thoroughness of the old process, while improving customer satisfaction. The team also wondered how its proposed changes might be received by CVS and its pharmacies. As described in the (B) case, in the latter half of 2002, the team members decided to change the process in three significant ways. First, data entry for each prescription was completed while the customer was still present, making it easier to verify customer insurance information. Second, the DUR process was folded into the quality assurance step, and both would be done after the prescription had been filled. Finally, the box that held the prescriptions was replaced by an online “virtual queue,” which could be displayed on all workstations in the pharmacy. As a result of the PSI, both customer satisfaction and pharmacy satisfaction increased significantly. Teaching Purposes This case provides a clear illustration of the capabilities conferred by enterprise IT. It shows how CVS was able to quickly and with high fidelity propagate a new business process across more than 4,000 separate locations by embedding process changes within EIT. The case illustrates all three of enterprise IT’s principle capabilities: design, standardization, and monitoring. It also highlights that shifted decision rights are an important organizational complement of EIT. As discussed below, these are decision rights about changes to the existing process. Finally, the case also indicates business leaders’ responsibilities during the work of EIT adoption. This two-case sequence is posed as a puzzle for students. The (A) case asks them to design a new prescription fulfillment process for CVS that addresses the problems identified by the PSI improvement team. I use assignment and in-class questions to highlight different features of a “complete” solution, many of which are easy to overlook. I then hand out the (B) case, which describes the PSI team’s solution and its positive impact on performance. I have found this to be a straightforward case to teach, one that works well in the classroom and resonates with students. It is one of the central cases in MIA’s module on enterprise IT. It also contains both qualitative and quantitative analysis, and so requires good preparation from students. Capabilities The work design, standardization, and monitoring capabilities offered by enterprise IT are all clearly on display in this case. As discussed below in Analyses: Capabilities, the case also shows how these capabilities can reinforce each other; case discussion reveals that CVS’s design and 2 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) 607-040 standardization efforts were aided by its enterprise software’s monitoring capabilities. EIT’s capabilities are my main emphasis when teaching this case. Complements I also spend some time in the CVS class on one of the organizational complements of IT, namely reallocated decision rights. As discussed below in Analyses: Complements, I do this by highlighting an apparent paradox. PSI team members believe correctly that their proposed process changes will be resisted by the company’s retail pharmacists, even though the new process shifts no responsibilities and takes no rights or responsibilities out of anyone’s hands. The key to understanding their reluctance is the realization that EIT adoption might not shift decision rights within the process itself, but it does centralize decision rights about changing the process. Responsibilities The cases themselves spend little time on the rollout of the new pharmacy process. Even so, they can still be used as a base for discussing business leaders’ roles in EIT adoption. The cases illustrate why relatively heavy-handed “top-down” adoption strategies are often both viable and desirable with EIT efforts. Process Analysis and Improvement In addition to serving as the base for discussing enterprise IT topics, the CVS cases are also effective for teaching about process analysis and improvement. The (A) case presents students with a business process problem that needs to be solved and the information required to solve it. Furthermore, case data can also be used to assess the financial impact of an effective solution to the company’s pharmacy service challenges. I start the class discussion by asking students to present their process analyses and improvement recommendations. Analyses Capabilities Redesign The (B) case describes the prescription fulfillment process changes designed and implemented by the PSI team. As that case describes, there are three main changes: Switching the order of insurance check and DUR (drug utilization review) steps Moving the insurance check forward Moving the DUR after the insurance check The most obvious required change, and the one that most students identify, is moving some process steps forward so that they take place at the pharmacy’s dropoff window. As part of this 3 607-040 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) change, most students also advocate staffing the dropoff window throughout the day. They correctly perceive that a broad problem with the process described in the (A) case is that CVS is missing an opportunity to identify and resolve problems early in the process, when the customer is still present. The ability to surface and correct problems close in space and time to their occurrence is fundamental within the Toyota Production System, and students understand its importance here. I have found, however, that most students do not make sufficiently detailed improvement recommendations. In particular, they don’t see the need to switch the order of the two steps, and to move the DUR to the script’s quality assurance (QA) step. Instead, they feel that simply “front- loading” the existing process will yield substantial improvements. There are two main reasons why this is not likely to be the case. First, as the (A) case mentions, there are compelling reasons why the customer should not be present during the DUR, so if it remains as the first computerized step in the process, then no steps can benefit from customer involvement. Second, even if this problem were somehow addressed, it is highly inefficient for the DUR to be the first step in the process. At that point, it involves three people—the dropoff window technician, the customer, and the pharmacist, and is an interruption to the pharmacist’s normal flow of work. If the DUR were moved to QA, it would involve only the pharmacist, who would not have to move between stations (as shown in Figure B of the (A) case). Another difference between most students’ proposals and the changes actually made by the PSI team is that IT changes are unnecessary for most student plans, but essential for the PSI team’s new process. It would be possible to front-load the existing fulfillment process by simply asking pharmacies to perform the initial three steps—data entry, DUR, and insurance check—at dropoff instead of waiting. This is what many students suggest, and it would require no changes to CVS’s pharmacy information systems. The PSI team’s new process, on the other hand, requires two significant changes to pharmacies’ existing enterprise systems. First, the insurance check needs to take place before the DUR. Second, the DUR itself needs to occur at the pharmacist’s station during the QA step. These differences are important not because of the technical difficulty of any required system changes, but because of the difficulty of accompanying organizational changes. As the (A) case stresses, pharmacists, technicians, and other employees are deeply concerned with the health and safety of their customers, and will resist any process changes that they feel have the potential to compromise safety. Rearranging process steps so that the DUR is not the first thing done is likely to be just such a change. Pharmacists could argue that it makes no sense, and is in fact dangerous, to physically dispense medication for a prescription before doing a DUR on that prescription. As discussed below in Analyses: Responsibilities, an important responsibility for the PSI team and CVS leadership is countering these arguments and overcoming pharmacies’ objections to process changes. Monitoring Regardless of what the first step in the new fulfillment process is, it is important that it take place at dropoff for reasons discussed above. However, it is not immediately obvious how IT can ensure that this process change will be implemented at the various CVS locations. It would be possible for pharmacies to continue to use the wooden boxes described in the (A) case to sort prescriptions by desired pickup time, and continue to do data entry only one hour before pickup rather than at the time of dropoff. Most students realize that this would be counterproductive, but few of them offer a plan to combat the possibility that goes beyond communication and training about the proposed new process and its advantages. 4 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) 607-040 Such training is certainly important, but it alone is unlikely to convince all pharmacies and employees to do away with their wooden boxes and start doing data entry at dropoff. In fact, with over 4,000 retail locations it is likely that at least a few will try to ignore the new process as much as possible and continue with the status quo. Even after it is pointed out to them, most students struggle to come up with effective means to combat this tendency. They suggest store audits, but these are expensive and time-consuming. What students typically fail to grasp is the enterprise systems already in use within CVS’s pharmacies provide timely and highly detailed audits of process execution at every location. All pharmacy transactions are recorded, and are visible and analyzable at company headquarters. This is all the information required for comprehensive auditing of compliance with a policy of doing data entry at dropoff. Because the distribution of prescription dropoff times at a typical pharmacy is known from experience, it can be compared with each actual pharmacy’s distribution of data-entry times. If these two distributions are dissimilar, an auditor can be reasonably confident that data entry is not taking place at dropoff at that pharmacy (see the figure below). In particular, if a pharmacy’s distribution of data-entry times looks like the distribution of requested customer pickup times shifted back one hour, an auditor can be confident that the pharmacy is still using its wooden box and only starting the fulfillment process one hour before the requested pickup time. Data-entry transaction timing within a pharmacy that is not doing data entry at dropoff In this pharmacy, data entry is still done approximately one hour before the desired pickup time. Since most customers want to pick up their prescriptions at lunch and after work, peak data-entry times are at 11 a.m. and 4 p.m. 5 607-040 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) Data-entry transaction timing within a pharmacy that is doing data entry at dropoff In this pharmacy, data entry is done at the time of prescription dropoff. Data-entry peak times are before work, at lunch, and after work. This example illustrates the powerful monitoring capabilities brought by enterprise IT. Because these systems are used to execute transactions and business processes, and because they capture and store data in consistent formats, they provide rich repositories of information. This information can be examined at multiple levels of aggregation, and for many purposes. As the CVS cases show, one such purpose is monitoring compliance with specified processes. Standardization One of the central themes of MIA’s enterprise IT module is that EIT is a powerful tool for ensuring that business processes are executed as intended by their designers (who are typically generalist business leaders). The CVS cases show how EIT delivers both ex ante and ex post compliance. Ex ante compliance stems from the fact that after the PSI team’s new process is put in place, it is simply not possible to do the DUR before the insurance check. Ex post compliance derives from EIT’s monitoring capabilities; as discussed in the previous section, auditors can easily determine if a remote pharmacy is in fact doing data entry at dropoff. These two types of compliance combine to yield a high level of business process standardization across CVS’s more than 4,000 pharmacies. Without enterprise IT, it is difficult to envision how such a high level of standardization could be achieved. Complements At first it seems that of the organizational complements of IT, the only one clearly present in this case is changed workflows. The PSI team changes the sequence and timing of a couple tasks in the pharmacy fulfillment process, but otherwise makes few alterations to the work of the pharmacies. It doesn’t seem that pharmacy employees are any more (or less) interdependent in the new process than in the old. It also seems that their decision rights are also preserved. Pharmacists still clear DUR exceptions, for example, and the sensitivity of the DUR itself is not altered in the new process. Given this, it seems strange for the PSI team to anticipate in the (A) case that pharmacists and other employees will resist their proposed process changes. If very few of the organizational 6 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) 607-040 complements of IT are being altered by the technology change, why would this change be at all problematic? Part of the answer, as mentioned in the case, is that pharmacy employees are deeply concerned about customer safety, and so will resist any change that has the potential to impact safety. But another reason for resistance, and perhaps a more fundamental one, stems not from the comparison of the old and new processes, but instead from the decision rights about the parameters of the new process and the transition to it. As the (A) case makes clear, the new process is defined by a small team. Many functions were represented on the PSI team, and it sought input from the field, but it couldn’t possibly canvass or solicit the opinions of even a fraction of the company’s more than 4,000 pharmacies. This means that most of the people and locations affected by the new process will have had no voice in defining it. They have, in other words, no decision rights about the change. Furthermore, they have no decision rights about whether or not to comply with the change because, as discussed above, the new technology itself ensures process compliance to a large degree. Even though pharmacy employees’ decision rights will be almost completely unchanged between the old and new fulfillment processes, the PSI team’s work does in fact greatly affect another important set of decision rights—those about whether and how to change the process. These rights are highly centralized within the PSI team and the company’s senior leadership, and are inextricably linked to CVS’s enterprise IT. This case therefore illustrates an important point about enterprise IT and its organizational complements. Even though a proposed enterprise technology might not significantly alter users’ decision rights within the new processes it brings, it could still greatly change decision rights about the process itself, and in particular about making changes to it. EIT adoption efforts like the one described in this case can take this latter category of decision right away from the parties that actually participate in or execute the process, and vest these rights instead with a remote central group like the PSI team. Responsibilities The case focuses on adoption, and spends little time on selection or exploitation. It stresses two separate adoption tasks for business leaders. The first, and most obvious, is leading the rollout of the new pharmacy fulfillment process across CVS’s thousands of retail locations. As the (A) case stresses, at least some pharmacists in these locations were likely to object to any process changes. When discussing the (A) case, the instructor can ask students for their plans on how to overcome this resistance and “sell” the new process. The (B) case details how the PSI team did this, and how they communicated the need for change and the efficacy of the process changes they were proposing. As this case describes, these communication efforts were persuasive and effective. They were also backed up by widespread awareness that the process changes proposed by the PSI team were not optional for the pharmacies; they were instead mandatory and inevitable modifications to the existing fulfillment process. The cases thus show how the “hard” and “soft” elements of an EIT adoption strategy can productively interact. The second adoption task stressed by the CVS cases is a business leader’s responsibility to get the new process right—to ensure that the new workflows and decision right allocations embedded in and deployed by the enterprise system will actually be improvements, and will lead to desired results. This point seems so obvious as not to be worth even mentioning, but a quick scan of enterprise IT “horror stories” reported in the press reveals that in many cases, even in high-profile projects within large and successful companies, enterprise systems went live with process configurations that 7 607-040 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) actually harmed the business. Examples include Hershey’s (Nelson and Ramstad 1999), Foxmeyer Drug (Rose and MacDonald 1998), and Nike (Konicki 2001); (Wilson 2001). As these examples make clear, the assumption that business leaders will “get it about right” with EIT is a dangerous one. As discussed below, instructors can stress this point when teaching the CVS cases by asking students how they can be sure that the new processes they design will actually be improvements. Process Analysis and Improvement The (A) case asks students to analyze an existing process and recommend improvements to it. It therefore has a strong operations management component to it, and at least two insights from that discipline are helpful. The first is that variability kills—that it is a huge impediment to effective process execution. If some steps display large variability in their throughput time, overall process performance deteriorates markedly. The problems brought by variability are greater if, as in the case at CVS, it is not possible to build up buffer inventories between process steps so that blocking and starving do not occur. A dominant theory of operations management has been called the “theory of swift, even flow” (Schmenner and Swink 1998). As the (A) case makes clear, the pre-PSI pharmacy fulfillment process at CVS experiences so many exceptions that it can be neither even nor swift. This process may have worked fairly well at some point in the past and experienced fewer exceptions. As the (A) case makes clear, however, the environment within which prescriptions are fulfilled has become much more complex in recent years. The number of available medications has increased greatly, as have the complexity of approval and payment processes. The mix of patients, payors, PBMs (pharmacy benefit managers), formularies, and generic medications combines to generate a potentially high volume of exceptions for any retail pharmacy’s fulfillment process. The fulfillment process described in the CVS (A) case is particularly ill-equipped to avoid or respond to these exceptions because exceptions become evident late in the process, rather than as early as possible. The process begins one hour before the customer’s desired pickup time, rather than when the customer drops off the script. This means that any problems encountered during data entry or later steps become evident when only little time remains to correct them, and when the patient herself is not available to assist with exception handling. The second operations management insight relevant for this case is that potential problems should be identified as early as possible in a process, and corrective actions should be taken as near as possible (in both time and space) to the point at which the exception occurs. This focus on rapid exception identification and resolution is core to the Toyota Production System (see, for example, Spear and Bowen 1999) and the case “Toyota Motor Manufacturing, U.S.A., Inc.,” HBS No. 693-019), and should be part of any process redesign at CVS. Many students understand this point, but do not perceive that to do this effectively CVS will need to not only move data entry to dropoff, but also move the DUR after the insurance check. Magnitude of Opportunity The case provides sufficient data to let students estimate quantities of interest such as: The percentage of CVS’s 7.2 million defecting customers in 2000 who are light versus heavy pharmacy users The volume of scripts lost annually to light and heavy defectors 8 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) 607-040 Total scripts filled annually by CVS pharmacies Average revenue per script Total annual lost revenue due to defecting scripts Revenue improvement from reductions in defection rate Improvements to CVS’s revenue growth rate from reductions in defection rate These calculations are shown and explained below in Table A. 9 607-040 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) Table A Data from Case Total 2000 customers: 29,500,000.00 p. 3 annual scripts, light users: 5.00 p. 3 annual scripts, heavy users: 40.00 p. 3 2000 revenue 20,100,000,000.00 p. 3 fraction of revenue from pharmacy: 0.67 p. 2 total light + heavy defectors 7,200,000.00 p.3 total defecting scripts: 55,000,000.00 p.3 Defection Calculations total light defectors: 6,660,000.00 see algebra below total heavy defectors: 540,000.00 see algebra below pct. Defectors who are light: 0.93 B21/(B22+B21) pct. Defectors who are heavy: 0.08 1-B23 total light defecting scripts 33,300,000.00 B21*B11 total heavy defecting scripts 21,600,000.00 B22*B12 Total scripts / year, 2000: 224,937,500.00 B9*B11*B23+B9*B12*B24 total pharmacy revenue, 2000: 13,400,067,000.00 B14*B15 Revenue per script, 2000: 59.57 B30/B29 revenue lost from lost scripts 3,276,482,067.24 B32*B18 algebra: light + heavy = 7.2 5*light + 40*heavy = 55 heavy = 0.54 Estimated defection savings % of light defections that could be saved w/ better service 0.13 Figure A % of heavy defections that could be saved w/ better service 0.44 Figure A Savings calculations total light that could be saved if 60% defection rate reduction 2,597,400.00 0.6 B$25*B$41*C47 if 70% defection rate reduction 3,030,300.00 0.7 B$25*B$41*C48 if 80% defection rate reduction 3,463,200.00 0.8 B$25*B$41*C49 if 90% defection rate reduction 3,896,100.00 0.9 B$25*B$41*C50 total heavy that could be saved if 60% defection rate reduction 8,791,200.00 0.6 B$25*B$43*C52 if 70% defection rate reduction 10,256,400.00 0.7 B$25*B$43*C53 if 80% defection rate reduction 11,721,600.00 0.8 B$25*B$43*C54 if 90% defection rate reduction 13,186,800.00 0.9 B$25*B$43*C55 total improvement in scripts This represents what % of total lost scripts? if 60% defection rate reduction 11,388,600.00 20.7% B58/B$18 if 70% defection rate reduction 13,286,700.00 24.2% B59/B$18 if 80% defection rate reduction 15,184,800.00 27.6% B60/B$18 if 90% defection rate reduction 17,082,900.00 31.1% B61/B$18 total improvement in revenue if 60% defection rate reduction 678,446,248.56 B58*B$32 if 70% defection rate reduction 791,520,623.32 B59*B$32 if 80% defection rate reduction 904,594,998.08 B60*B$32 if 90% defection rate reduction 1,017,669,372.84 B61*B$32 Impact on CVS Revenue Growth 2002 2001 2000 1999 1998 Net Operating Revenues 24181.5 22241.4 20087.5 18098.3 15273.6 revenue growth 1940.1 2153.9 1989.2 2824.7 average rev growth over this period 2,226.98 % of this represented by: 60% defection rate reduction 30.5% (B64/1000000)/B$74 70% defection rate reduction 35.5% (B65/1000000)/B$74 80% defection rate reduction 40.6% (B66/1000000)/B$74 90% defection rate reduction 45.7% (B67/1000000)/B$74 Source: Casewriter. 10 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) 607-040 Assignment Questions 1. “What changes do you recommend to CVS’s existing pharmacy fulfillment process? What IT changes, if any, are required to implement your changes?” 2. “How can you be sure that the new process you propose will be an improvement over the existing one? How can you be sure that it won’t make things worse?” 3. “What groups, if any, are likely to have problems with your proposed solution? How will you deal with their objections?” 4. “How will you ensure that there’s no backsliding—that there won’t still be wooden boxes in use six months from now? How can technology be used to prevent or inhibit backsliding?” 5. “Does PSI represent a significant opportunity for CVS? Would improving customer service be of significant financial benefit to the company?” 6. “What percent of pharmacy defectors from CVS in 2000 were light versus heavy users?” Teaching Plan Because this is a two-case sequence, and because it contains some quantitative analysis, I find this to be a “full” class—one that easily fills up 75–80 minutes. My teaching plan typically opens with an assessment of the financial impact of pharmacy service improvement at CVS. I then move into an analysis of the problems with the current process and a discussion of student plans to improve it. In this latter discussion, I focus on whether students’ plans address all of the root causes of the poor current levels of service—whether they’re complete or piecemeal. I then hand out the (B) case describing the new fulfillment process defined by the PSI team, ask students to comment on it, and finish class by asking how CVS can ensure compliance with the new process. This is usually a successful class. Students appreciate its insights, particularly the fact piecemeal solutions won’t work and that IT is essential for ensuring compliance. I have found that the “right” answer for this case is neither too obvious nor too subtle for students. I often ask students to submit their redesigned processes prior to class. This forces them to spend some time thinking about them and to commit in advance to one course of action. It also gives me the opportunity to review their proposals before class, and to focus in class on students, plans, and details that support the case’s teaching points. Opportunity (15 minutes) I start the class by asking a question like: “Does PSI represent a significant opportunity for CVS? Would improving customer service be of significant financial benefit to the company?” The goal of this section of class is to help students understand that improving pharmacy fulfillment will significantly improve CVS’s revenue, profits, and growth rates, and so the PSI effort, far from being peripheral, is core to the company. I walk them through the calculations given above in Table A in roughly the order shown. I have found it important to take students through the 11 607-040 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) algebra required to calculate the percentages of light and heavy defectors, and often assign this calculation as a pre-class assignment question. Process Analysis/Problems with Current Process (15 minutes) I then ask students to describe the problems with the CVS’s current pharmacy fulfillment process. They usually home in fairly quickly on the facts that, as discussed above in Analyses: Process Analysis and Improvement, exceptions are identified too late in the process, and that CVS is giving up on the opportunity to involve customers in exception handling and problem resolution. I then ask a question like: “Why would you design a process like this?” The answer, as discussed above in Analyses: Process Analysis and Improvement, is that CVS’s process was designed when the prescription fulfillment environment was much less complex than it was in 2002. In every class, I have found, some students are articulate about describing the drivers of this complexity, from the proliferation of medications and generics to the rise of formularies and PBM companies. To help this discussion along, I ask questions like: “What’s a formulary? Why do they exist? Whose purposes do they serve? What’s a PBM?” Improvement Plans (20 minutes) I then ask one of the standard questions of the MIA course: “So what IT-based capabilities does CVS need at this point in time?” Students quickly and accurately respond that the company needs to redesign its pharmacy fulfillment process and standardize it across its thousands of retail locations. I then ask them to present their redesigned processes. As discussed above in Analyses: Capabilities: Redesign, many students recommend moving data entry to dropoff, but most do not see the need to make the insurance check the first step and to move the DUR to QA. Since I ask students to submit their redesigned processes before class, I can select appropriate students to highlight these points. If any student has submitted a plan that does in fact correctly call for the order of these two steps to be changed, I call on her last. I then ask the second of the three standard questions: “What new organizational complements need to be put in place to gain these capabilities?” or more specifically “What new organizational complements need to be put in place for the new pharmacy fulfillment process at CVS?” The point of this question is to raise the point that, as discussed above in Analyses: Complements the new process apparently requires very few new complements. Interdependence is not greatly increased (pharmacy employees were already highly interdependent), data is already consistent, and decision rights are not at all reallocated by the new process. It seems that the only novel complement is a changed workflow. This analysis, however, ignores the fact that, as discussed above in Analyses: Complements, with PSI the decision rights about changing the pharmacy fulfillment process move from the pharmacies themselves to the PSI team at CVS headquarters. This is a fundamental shift, and one that is not likely to be welcomed by at least some pharmacists, especially because they realize that, as discussed above in Analyses: Capabilities: Monitoring, they will not be able to simply ignore the new process and continue with “business as usual.” I initiate this discussion with a question like: “So if so little change is 12 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) 607-040 required—if so few new complements need to be put in place—why is the PSI team so worried about pharmacists’ reactions? The new process sounds like it will actually be less work for them!” (B) Case (10 minutes) After discussing how the PSI effort will centralize an important decision right—the right to change the fulfillment process—I hand out the (B) case and give students a couple minutes to read it. I then ask them to comment on whether or not it addresses the root causes identified earlier in the class. Most students agree that it does. The last portion of the class is a discussion of how the PSI team can ensure that pharmacies don’t ignore important elements of the new process, in particular the requirement to do data entry at dropoff. I initiate this discussion by asking, “How will you ensure that there’s no backsliding—that there won’t still be wooden boxes in use six months from now? How can technology be used to prevent or inhibit backsliding?” A variant of this question is “How can a PSI team member at headquarters determine if a pharmacy is still using a wooden box after they’ve been switched over to the new process?” This question is intended to bring up the fact that, as discussed above in Analyses: Capabilities: Monitoring, EIT’s monitoring capabilities can be used to help ensure compliance with new business processes. Most students don’t list better monitoring as one of the capabilities CVS needs to acquire, but come to see after this discussion that monitoring is in fact critical for ensuring process compliance in this and many other cases. Video (15 minutes) I then show students a video of Jon Roberts and Josh Flum speaking to a class of my MBA students (HBS No. 607-717). The video is approximately 15 minutes in length and touches on the following topics (in the list below, numbers indicate the time at which discussion of the topic begins): 0:00 Other proposed solutions to pharmacy problems. Need to find root causes. Statistics about how bad “symptoms” were. 1:28 Description of new process. Data entry at dropoff. Customers can help solve problems. DUR hard-stopped the process, but pharmacist could almost always clear them without consulting anyone. 5:00 Other measures to improve communication. Other pharmacy process improvement measures. 8:20 “Selling” the new process. Convincing employees that their jobs would get easier. 9:50 IT and compliance. “Technology drives compliance to a process.” 11:35 Team was set up to succeed. Skepticism about new process. 12:45 Pharmacies have to do ALL seven elements of the new process. 13:40 Internal salesmanship. 14:12 New process measurement. Systematic measurement of improvements. If results weren’t there, team went back into the store. 13 607-040 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) Conclusion (5 minutes) In my concluding comments on this case, I stress the positive impact a relatively small process change can have when it’s replicated across thousands of locations. I also highlight two of the less obvious aspects of the case—that resistance to new EIT often comes not because of the new processes it embeds, but instead because it centralizes decision rights about process change, and that EIT-based monitoring can be very effective in ensuring compliance with business process execution. Blackboard Plan The following pages capture the blackboards at the end of a typical class devoted to this case. The blackboard content follows the teaching plan given above fairly closely. I usually use five boards in a class, one of which is devoted to a diagram of MIA’s organizing CCR framework. I refer to this diagram during the class discussion, and frequently make use of it during my concluding remarks when I stress the insights and course themes that emerged from the class discussion. 14  607-040 -15- Board Plan Enterprise IT Capabilities Design Standardization Monitoring Responsibilities Complements Selection Decision rights Adoption Interdependencies Exploitation Workflows (Bold font indicates that adoption is a particularly Important responsibility with enterprise IT) 607-040 -16-  607-040 -17- Board Plan (continued)  607-040 -18- Board Plan (continued)  607-040 -19- Board Plan (continued) 607-040 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) Appendix Managing in the Information Age’s Capabilities-Complements-Responsibilities Framework Introduction Managing in the Information Age (MIA) focuses on IT deployments that are intended to change how work is accomplished within or between companies.1 The course divides them into three categories: deployments of function IT (FIT), which affect the execution of a single function, job, or task; of enterprise IT (EIT), which integrate multiple functions by imposing new work structures across them; and of network IT (NIT), which integrate multiple functions but do not impose new ways of working. Instead, some network technologies allow new work structures to emerge over time (see Table A-1). Within the context of the course the phrase “work structures” primarily means attributes of a business process including workflows, interdependencies, and decision right allocations. However, course materials also use the phrase “work structures” to encompass higher-level attributes of an organization such as degrees of centralization or decentralization. Finally, MIA also uses “work structures” to capture low-level data formats and definitions, which are important for many types of IT. These three technology categories and the differences among them are discussed in more detail in the course overview note (HBS No. 608-011). Other course materials include a note on pedagogy (HBS No. 608-007) and notes on each of the course’s three principal modules: single-company enterprise IT (HBS No. 608-008), multiple-company enterprise IT (HBS No. 608-009), and network IT (HBS No. 608- 010). Table A-1 MIA’s Three IT Categories IT Category Definition Examples Function IT (FIT) IT that assists Simulators execution of a discrete Spreadsheets function or task CAD/CAM software Statistical software Enterprise IT (EIT) IT that integrates Enterprise Resource Planning (ERP) systems multiple functions by Supply chain management (SCM) systems imposing new work Customer Relationship Management (CRM) systems structures Sourcing/procurement software Network IT (NIT) IT that integrates E-mail multiple functions Instant messaging without imposing new Wikis work structures Prediction markets Source Casewriter. The concepts of capabilities, complements, and responsibilities provide the course’s organizing framework. MIA’s point of view can be summarized as follows: The three technology categories offer distinct and non-overlapping capabilities. To obtain these capabilities, organizations must not only deploy technologies—they must also put in place 1 The course does not consider IT efforts that are invisible to end users. Database and server upgrades, for example, are important projects, but they are not included in MIA’s materials. 20 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) 607-040 organizational complements, or changes in how work is accomplished. These complements are identical to the work structures defined above. During IT deployments, general managers have three main responsibilities: Selecting IT based on desired capabilities. Adopting the new technology by putting in place appropriate complements. Adoption is defined as preparing both the technology and the organization prior to the “go live” date. Exploiting technologies once they’re up and running to maximize the capabilities delivered. Course module notes provide detail on the applications that fall into each category, their characteristics, and how technologies and general managers combine to deliver each category’s unique capabilities by imposing complements (in the case of EIT and FIT), or letting them emerge (in the case of NIT). Capabilities MIA spends little time on the functionality, feature sets, or “look and feel” of the many technologies covered. Instead, course materials emphasize the higher-level capabilities that information technologies offer to the organizations that adopt them. MIA holds that these capabilities are largely consistent within each technology category. All enterprise information technologies, for example, deliver similar capabilities, even though the EIT category comprises applications developed by many vendors and positioned for distinct markets and customers. The course also holds that while capabilities are similar within the categories of FIT, EIT, and NIT they are largely different across them. Function IT, for example, does not offer a subset of enterprise IT’s capabilities, or even a significant intersection with them. In other words, the Venn diagram of the capabilities offered by the three IT categories would consist of three circles with little overlap between them: Capabilities from Function IT Increased experimentation capacity. FIT lets knowledge workers like scientists, engineers, and financial analysts conduct huge numbers of trials and quickly explore the “solution space” of their work. Greater precision. If a system is well understood it can be modeled very accurately with FIT. This lets knowledge workers conduct very precise experiments, understanding the effects of even small changes and finding optima. Capabilities from Network IT Facilitating collaboration. Network technologies let people work together without mandating or trying to predict how they will interact. This flexibility means that network technologies support many modes of collaboration. Allowing expression of judgment. Network technologies give their users a voice and, in many cases, a forum to express themselves. Prediction markets, for example, let people express their judgment about the likelihood of future events. 21 607-040 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) Enabling self-organization, or the appearance of visible high-level patterns and structure as the result of many low-level interactions, without central guidance or direction. Capabilities from Enterprise IT Design and redesign of business processes. Enterprise technologies allow business leaders to define and then deploy novel business processes, even ones that cut across many organization groups. Standardization of workflows across an arbitrarily large “footprint.” Once a business process has been embedded within EIT, it can be replicated and distributed cheaply, widely, and with high fidelity. One of the problems of distributed companies is a reduced ability to control what happens far from headquarters; EIT provides a way to address this problem. Monitoring of activities and events, often in something close to real time. Another problem of corporate scale has been managers’ inability to observe the performance of remote operations; available reports are often both out of date and too aggregated. With EIT in place, managers can get a more precise and current picture of what’s happening throughout an enterprise. Complements MIA takes the view that acquiring capabilities from IT is not simply a matter of purchasing and installing the software. The course instead views information technology as a general purpose technology (GPT), the value of which is maximized by complementary innovations. As discussed in the course overview note, the complements to IT are organizational; they are changes in organizations’ configurations that maximize the positive impact of the new technology. Within MIA, “complements” and “work structures” are synonymous and are distilled from previous scholarship on GPTs. MIA stresses two points about IT’s complements. First, although the complements are the same for all three IT categories, each category has a distinct relationship with them. Enterprise IT imposes complements. An ERP deployment, for example, defines many cross-functional business processes within software. Each of these processes makes the groups it covers highly interdependent, specifies the precise process workflow and the decision rights of all involved parties, and relies on consistent data. Network technologies like e-mail and wikis, on the other hand, are not used (and in fact typically cannot be used) to impose their complements. They are instead vehicles to let IT’s organizational complements emerge over time. A wiki, for example, makes its users more interdependent, but doesn’t specify in advance how they should work together. It also doesn’t assign them into roles or assign them specific decision rights in advance, nor does it mandate what data should be included, or where. However, roles, interdependencies, and consistent data appear over time in most wikis. With FIT, finally, the technology’s organizational complements are separate from the technology itself. Because a function IT operates only on a single task, it cannot by itself increase interdependence, reallocate decisions right, etc. In the presence of a new function technology it might be desirable to alter some organizational complements, but the technology itself is not central to this work—managers are. The second point MIA stresses about IT’s organizational complements is the lack of a “one best way.” As cases within and between modules make clear, there is no single set of workflows, decision right allocations, interdependencies, and data structures that will work in all organizational contexts, or even most of them. Instead, both the appropriate information technologies and the appropriate set 22 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) 607-040 of organizational complements that maximize the capabilities delivered by the technologies are heavily dependent on circumstance and context. Responsibilities The course holds that generalist business leaders are the most important constituency for success with IT, and specifies three responsibilities for them. As shown in the figure below, these are selection, adoption, and exploitation. Buy IT “Raw “Live” IT “Finished Selection Adoption Exploitation materials” System Goods” Build Configuration Use Testing Training Time Selection is the work of choosing new information technologies. MIA advocates that business leaders select IT using an “inside-out” approach in which they start inside by determining what capabilities are most important to acquire before looking out at the landscape of available technologies. While this sounds like common sense, it is actually quite far from the “outside-in” approach taken by many businesses in which they select technology based on what’s novel or popular in the present landscape. After a new technology is selected and bought or built, a business leader’s second responsibility is adoption, which is defined as the work of preparing the technology for initial use. Adoption is usually a minor activity for function and network technologies, but a major responsibility during enterprise IT deployments. This is because EIT adoption (which includes system configuration, testing, and user training) is essentially the work of defining in advance the organizational complements that will be imposed by the new technology. As several of the course’s cases make clear, this is time-consuming, difficult, and often contentious work. Once any new technology is up and running, a business leader’s final responsibility is exploitation, or ensuring that the technology is delivering its capabilities to the maximum extent possible. MIA’s materials show that there are a wide variety of ways to do this, and that IT exploitation is, in fact, closer to an art than a science. 23 607-040 Teaching Note—Pharmacy Service Improvement at CVS (A) and (B) References Konicki, S. “Nike Just Didn’t Do It, Says i2 Technologies.” Informationweek, March 5, 2001, p. 32. Nelson, E. and E. Ramstad. “Trick or Treat: Hershey’s Biggest Dud Has Turned Out to be Its New Technology.” Wall Street Journal, October 29, 1999, p. A6. Rose, M. and E. MacDonald. “Foxmeyer Trustee Sues SAP, Deloitte; $500 Million Is Sought from Each Firm.” Wall Street Journal, August 27, 1998, p. B8. Schmenner, R. W. and M. L. Swink 1998. “On Theory in Operations Management.” Journal of Operations Management 17(1): 97–113. Spear, S. and H. K. Bowen 1999. “Decoding the DNA of the Toyota Production System.” Harvard Business Review 77(5): 96–107. Wilson, T. “Supply Chain Debacle—Nike Faces Yearlong Inventory Problem after i2 Implementation.” InternetWeek, March 5, 2001, pp. 1–3. 24