This preview has intentionally blurred parts. Sign up to view the full document

View Full Document

Unformatted Document Excerpt

10: Chapter Service Facility Location Chapter 10 Service Facility Location TEACHING NOTE This chapter begins using La Quinta Motor Inns to illustrate the strategic considerations in the facility location decision. Several strategies are explored including the impact of the Internet on service location. Returning to La Quinta we consider an application of regression analysis to motel site selection that is often used in business statistics classes. A discussion of geographical information systems follows. Because techniques for facility location analysis tend to be very complex, only a few simple approaches are discussed in length. The "cross median" approach to the location of a single facility using a metropolitan travel metric is intuitively pleasing because it is based on the statistical property of the median. This approach is easily implemented graphically and often produces counterintuitive results. Furthermore, the optimality of the solution, whether it is a point, a line, or an area, is easily verified. The Huff retail-location model has been used extensively for locating shopping centers, retail banks, convenience stores, and other profit-making endeavors. The location set covering model is a popular approach to the public sector multiple facility location problem. SUPPLEMENTARY MATERIALS Case: Southwest Airlines-1993 (A) (HBS case 9-694-023) Southwest Airlines, the only major U.S. airline to be profitable in 1992, makes a decision as to which of two new cities to open, or to add a new long-haul route. This case provides a window into Southwests strategy, operations, marketing, and culture. Lessons can be learned how an airline can simultaneously be a low-cost leader, service leader, and profit leader. Software: Excel spreadsheet of Huff model application to Athol Furniture case on CD-ROM LECTURE OUTLINE I. Strategic Location Considerations Competitive Clustering Saturation Marketing Marketing Intermediaries Substitution of Communications for Transportation Separation of Front from Back Office (Table. 10.1) Impact of the Internet on Service Location Site Considerations (Table 10.2) II. Regression Analysis in Location Decisions (Table 10.3) III. Geographic Information Systems (Figure 10.1) IV. Modeling Considerations 10-1 Geographic Representation (Figures 10.2 and 10.3) Number of Facilities Optimization Criteria (Figure 10.4) V. Facility Location Techniques (Figure 10.5) Cross Median Approach for a Single Facility (Fig 10.6 and Tab. 10.4, 10.5, 10.6) Huff Model for a Retail Outlet (Tables 10.7, 10.8, 10.9, 10.10) Location Set Covering for Multiple Facilities (Figure 10.7 and Table 10.11) TOPICS FOR DISCUSSION 1. Pick a particular service, and identify shortcomings of its site selection . [Student Frank Tyler provided the following response.] Many times the extent to which a poor decision regarding the location of a service facility affects the business is not apparent until a competitor locates on a better site. The location of Kemper Arena in Kansas City is a case in point. The arena is located in the old stockyard district near downtown and it is no longer a safe area. It is not served directly by any freeway or interstate system and most people travel to the arena through the downtown area. Therefore, events that are scheduled for early evening exacerbate the usual downtown evening rush hour traffic. An additional problem occurs when the event ends and people must exit the many small parking lots. Although no causality has been proved, it is often suggested that the poor attendance at games for the professional sports teams based at Kemper results from the stadium's location. In fact, attendance is a serious problem that has prompted two teams (the Kansas City Comets and the Kansas City Kings) to leave Kansas City. The Harry S. Truman Sports Complex offers a striking contrast to Kemper Arena. The Complex stadiums, which house the Kansas City Royals and the Kansas City Chiefs, are located in a suburban area next to a major freeway and two large multilane city streets. These thoroughfares expedite traffic into and out of the stadiums. Most people who live in metropolitan Kansas City can reach the Complex faster than they can reach Kemper. This is an excellent example of the situation where the shortest distance is not always the quickest route. The location of the Complex offers advantages other than the savings in travel time for patrons. Its location is safer than the Kemper area, it offers many restaurants and bars for pre-game and post-game dining, and has three hotels within walking distance so out-of-town people have a convenient place to stay when they come to cheer their visiting teams. One disadvantage that Kemper and the Complex share is a lack of good public transportation. A final irony may be noted: Kemper was invited to be a part of the Complex ... but declined! 2. How would you proceed to estimate empirically the parameter " " in the Huff retail location model for a branch bank? From equation 12 we have the following function that describes the attraction of facility j for customers located in area i: 10-2 Aij = Sj Tij Thus: Tij = Sj Aij Using logarithms solve for : log Tij = log S j log Aij = log S j log Aij log Tij The value of Sj could represent the size of the facility in square feet. The value of Ti j is the time for customers in area i to travel to a bank located at site j. The value of Aij could be measured as the percentage of potential consumers in area i who have accounts at bank j. An expected value for is determined by applying the above function to areas that are served by the branch, weighted by the number of customers in each area. 3. What are the characteristics of a service that would make communication a good substitute for transportation ? The service usually involves the creation, transfer, or storage of information rather than products. The cost of transportation must be greater than the cost of the communications system. Saving time is an important factor. The technological base must be available to the customers. A communication network is required. The quality of the service may need to be apparent in advance of the delivery of the service. No face-to-face contact is needed. Cash payment is not required. The most promising opportunities to substitute communications for transportation occur in services such as health care (paramedics, telemedics), education, government services (tornado warnings, satellite estimations of crop yields), extension of databases (interlibrary reference services), and computerized mail delivery. 4. What are the benefits of using intermediaries in the service distribution channel ? The distribution of services presents problems because the services are intangible and cannot be stored, stockpiled or moved. These factors restrict the geographical area for services. The use of marketing intermediaries eliminates the problems and creates an unlimited geographic service area. Therefore, a business may see an increase in the demand for its services. In addition, because of the greater service area, the entire service industry 10-3 may become more competitive, which in turn fosters increases in efficiency. Another benefit of the increased availability is the increased convenience for the customer. Examples of intermediaries include credit charge cards such as MasterCard and Visa. These cards may be used worldwide even though local banks may issue them. Many times the use of marketing intermediaries can directly affect the ultimate cost to the customer. For instance, group insurance policies can be written through businesses and labor unions, which increases the demand for services and provides insurance at a lower cost for the employees. Also, because the insurance is written through the business or labor union, the employees may be assured of quality service. INTERACTIVE CASE EXERCISE The class discusses the business opportunities of using Google Earth . Google Earth is often used in business websites to facilitate finding the service location. With GPS becoming available on cell phones, Google Earth can show what businesses, such as restaurants, are in the vicinity of the caller. EXERCISE SOLUTIONS 10.1 Copying Service Revisited y 5 w (5) 3 4 3 2 w (9) 2 A w (7) 1 1 1 Median = 2 3 w (7) 4 4x 28 = 14 2 No need to relocate because optimal location is line segment (2, 2-3), which includes location A at (2,2). 10-4 10.2 Temporary-help Agency y 10 . . c (2) 2 c (4) 4 c (1) 5 . 5 .c 3(2) c (3) 1 5 10 W i Median = . 2 = 15 x 12 =6 2 Optimum location lies along line between end points of (7,7) and (7,11). 10.3 Hospitals y 10 . H (450) 1 .H 2 (1200) 5 H3 . . (300) 5 Median = H 4 (1500) 10 15 x 3450 = 1725 2 Optimum location is at (7,3). 10-5 10.4 Pizza Delivery y 6 c (4) 2 5 c 4 3 2 5 (5) c (3) 3 c (5) 1 1 1 3 2 c (1) 4 5 4 x 18 =9 2 Median = Optimum location is area defined by corner points: (2,3), (2,4), (3,3), and (3,4). 10.5 Airport Terminal y 10 5 . .. A(28) C(36) . D(18) B(22) 5 Median = 10 15 x 104 = 52 2 Optimum location is (8,3) or terminal C. 10-6 10.6 Catering Service Y 15 P 5 (1) P( 4 4) 10 . P3 ( 2) 5 P( 4) 1 P2 ( 3) 10 5 X 15 Locate site in area defined by: 7 x 11 4 y7 10.7 Copying Service Revisited Using Huff Model For the existing store at A assume that S1 = 1 and for the new store at B assume that S 2 = 3 (i.e. three times the capacity of existing store) and = 2. Assume new demand weights from question 7.1. First calculate: Aij = Area New (S=3) Old (S=1) A Sj Tij2 1 3/2 = .75 1/12 = 1.0 1.75 2 3/2 = .75 1/12 = 1.0 1.75 2 ij Second, calculate: Pij = 3 3/3 = .3333 1/42 = .0625 0.3958 2 2 4 3/2 = .75 1/32 = .11 0.86 2 Aij A ij Store New Old Orders Sales ($) 1 .43 .57 700 7,000 Customer Area 2 3 .43 .84 .57 .16 900 500 9,000 5,000 Sales ($) 4 .87 .13 700 7,000 10-7 Share (%) 17,170 10,830 .61 .39 28,000 1.00 10.8 (a) Department Store Customers 1 2 Avg. budget (Bi) 1 $ 100 $ 150 $ 125 No. customers (Ci) 2 $ 75 $ 100 $ 87.50 1500 2500 Areas 3 $ 125 $ 125 $ 125 4 $ 100 $ 120 $ 110 1000 3000 5 $ 120 $ 125 $ 122.50 2000 Annual home appliance sales (in thousands) 5 T = Ci Bi = 187.5 + 218. 75 + 125 + 330 + 245 i =1 = $1106. 25 10.8 (b) For store 1 (existing locally owned store) assume that S1 = 1 and for store 2 (Bull's Eye new store) assume that S2 = 3 (three times the size) and = 1. First calculate: Aij = Area Local (S=1) Bull's Eye (S=3) A Sj Tij 1 1/20 = .05 3/15 = .20 0.25 ij Second, calculate: Pij = 2 1 /35 = .0286 3/20 = .15 0.1786 3 1/30 = .0333 3/25 = .12 0.1533 4 1/20 = .05 3/25 = .12 0.17 5 1/25 = .04 3/25 = .12 .16 Aij A ij Store 1 Local 2 Bull's Eye *Sales (1000s) * C B i i 1 .2 .8 $187.5 Customer Area 2 3 4 .16 .217 .29 .84 .783 .71 $218.75 $125 $330 using data from problem 7.8 10-8 Sales (1000s) 5 .25 .75 $245 $ 257.98 $ 848.27 $1106.25 10.9 (a) Gas Stations For the old station assume that S1 = 1 and for the new station assume that S2 = 2 (double the number of pumps) and = 2. First calculate: Aij = Area Old (S=1) New (S=2) A ij Sj Tij2 1 1/52 = .04 2/202 = .005 0.045 Second calculate: Pij = 2 1/102 =.01 2/82 = .0312 0.0412 3 1/92 =.0123 2/122 = .0139 0.0261 Aij A ij Station Old New Customers Customer Area 1 2 3 4 .88 .24 .47 .07 .12 .76 .53 .93 100 150 80 50 (b) Market loss = .12(100) + .76(150) + .53(80) + .93(50) = 215 215 Percent loss = = 57 percent 380 10-9 4 1/152 = .004 2/62 = .055 0.059 10.10 Rural Medical Clinics Community 1 2 3 4 5 6 7 8 9 Set Served by Community 1,2,3 1,2 1,3,4,5 3,4,5,6 3,4,5,6 4,5,6 7,8 7,8 9 Sites that Serve Community (#6 omitted) 1,2,3 (1,2) 1,3,4,5 3,4,5 3,4,5 (4,5) 7,8 (7,8) (9) Possible solutions: {1 or 2, 4 or 5, 7 or 8, 9} 10.11 Banks Locating ATMs Site 1 2 3 4 5 6 7 Could Reach 1,2,3 1,2, 4,5 1, 3,4 2,3,4, 6,7 2, 5,6 4,5,6 4, 7 Solution {2 and 4} 10.12 (a) Volunteer Fire Department Site 1 2 3 4 5 6 7 Could Reach 1,2,3 1,2, 4,5 1, 3,4 2,3,4, 6,7 2, 5,6 4,5,6,7 4, 6,7 Possible solutions {2 and 4} or {1 and 6} (b) Consider the population at the different communities (i.e., fire potential). 10-10 CASE: HEALTH MAINTENANCE ORGANIZATION (C) First calculate the median value: (81)/2 = 41 enrollees (in 1000s). Next with the x and y axes oriented in the direction of the major streets shown on the map below, we can see that a location for the facility in census tract 5 or 6 would be appropriate. It is interesting note to that St. Davids Hospital is located in census track 5. 10-11 CASE: ATHOL FURNITURE, INC. Athol Furniture Inc. is preparing to open a new store in Bluff Lake, a small community that already has two stores that sell furniture and large appliances. Athol has identified three possible sites for its new outlet. The designated sites and their respective store size options are (1) Site X: 10,000 or 15,000 sq. ft.; Site Y: 10,000, 15,000, or 20,000 sq. ft.; Site Z: 10,000 sq. ft. An analysis of the six options using the Huff Model is performed to select the site and store size that will maximize the net operating profit before taxes. Bluff Lake is divided into 12 census tracts and detailed information about the population including incomes and spending habits (relative to purchases of furniture and large appliances) is available. The probability that a customer in a given census tract will shop at one of the new stores can be computed based on the travel time (of the customer to the store) and the size of the proposed store. These calculations provide a probability matrix (6 variables across 12 census tracts) that is used to estimate customer expenditures and the pretax profits for each variable. When the probabilities obtained from the model are applied to the detailed information on expenditures, we can calculate estimates of the gross sales at each proposed site. The equation for this procedure is shown below: m Tj = ( Pij Ci Bik ) j =1 The Ci Bik term remains constant within a census tract for each decision option. The estimates of customer expenditures for each store are shown in the spreadsheets that appear with the following answers. 1. Utilizing a ite Y wspreadsheet version of the Huff location model (with = 1.0), recommend a store size and location for AFI that will maximize expected net operating profit before taxes. Assuming that AFI does not wish to consider a store smaller than 10,000 square feet, assess the store sizes (based on increments of 5,000 square feet) up to the maximum allowable sales area for each potential site. Using the Excel spr "annua eadsheet HUFF Model for Retail Location, the results shown below indicate that a 15,000 sq. ftl profit" of $23, 39 3 m. store at sith anaximizes the expected net operating profit before taxes. 10-12 HUFF MODEL FOR RETAIL LOCATION PROFIT MARGINS --------------SIZE MARGIN 10000 0.039 15000 0.036 20000 0.029 PROFIT CALCULATIONS ; LAMBDA = | | | | SIZE ------10000 15000 MARGIN BEFORE TAXES | ------| 0.039 | 0.036 | | | | | | 10000 15000 10000 15000 20000 10000 | | | | | | 1 ANNUAL PROFIT ------------21127 40144 MARKET SHARE ------0.327 0.673 0.039 0.036 0.039 0.036 0.029 0.039 18521 21867 19683 23393 22055 16748 0.2866 0.3666 0.3046 0.3922 0.459 0.2592 Head to head comparisons A | 10000 | 0.039 B | 15000 | 0.036 X | 10000 | 0.039 16275 27527 18521 25.19% 46.15% 28.66% PROPOSED SITE A B X X Y Y Y Z A B X | 10000 | 15000 | 15000 | 0.039 | 0.036 | 0.036 14752 24162 21867 22.83% 40.51% 36.66% Y | 10000 | 15000 | 10000 | 0.039 | 0.036 | 0.039 15587 27088 19683 24.12% 45.42% 30.46% Y | 10000 | 15000 | 15000 | 0.039 | 0.036 | 0.036 13859 23459 23393 21.45% 39.33% 39.22% Y | 10000 | 15000 | 20000 | 0.039 | 0.036 | 0.029 12497 20732 22055 19.34% 34.76% 45.90% Z | 10000 | 15000 | 10000 | 0.039 | 0.036 | 0.039 14801 30523 16748 22.91% 51.17% 25.92% A B A B A B A B 10-13 2. What is the expected annual net operating profit before taxes and market share for the outlet you have recommended? Defend your recommendation. Although a 15,000 sq. ft. store at site Y yields an expected annual net operating profit of $23,393, a 20,000 sq. ft. store at the site would maximize the market share at 45.9 percent. With the larger store the market share of the competition is severely eroded (see head-to-head comparison). Store A with a 10% loss in market share (33% to 23%) will probably not survive. Furthermore, with store B and Athol located in close proximity the effect of clustering will draw even more customers to shop both stores simultaneously. 3. Try two other values of (e.g., 0.5 and 5.0) to measure the sensitivity of customer travel propensity on your recommended location. Store Size 0.5 5.0 Y Y 15,000 15,000 Max Profit $23,122 $22,450 Store Size Y Y 20,000 20,000 Max Market Share 46% 41% 4. Briefly state any shortcomings you may perceive in this model . (a) Size may not be an adequate approximation of the attractiveness of a furniture store. Other qualitative aspects such as perceived quality of the furniture, friendliness of the sales staff, and the customer's loyalty to an existing store may confound the size variable. (b) At first glance, a long travel time to the store may be seen as a disincentive to visit a furniture store. To counteract such a disincentive, furniture stores can emulate automobile dealers who establish "motor miles" in many communities. This strategy of locating all or most dealers in one area is effective in drawing many customers from great distances away. (c) Other factors, most of that are not quantifiable, may also alter a customer's perception of the utility of the shopping center. For example, consider the effect of an established firm's reputation or the lure of a new store. The visibility of a store and the nature of adjoining stores or neighborhoods may also influence the store's ability to attract customers, but these factors are not easily measured. CHAPTER QUIZ QUESTIONS True/False 1. Finding a unique set of sites in a multi-location problem is a simple extension of the methods used for single facility location. (F) 2. When a customer travels to the service facility, the direct cost that is incurred is the decrease in potential customer demand that occurs as a result of the distance traveled. (F) 3. The metropolitan metric method of measuring distance is described as travel "as the crow flies." (F) 4. Selecting a site for a public service center, locating in an area that contains a large number of individuals for whom distance is a strong barrier would maximize facility utilization. (T) 5. The choice of location criterion has little effect on the site selected. (F) 6. Saturation clustering occurs when service firms locate their business near their competitors because they expect a high level of customer traffic. (F) 10-14 7. The Huff model for retail outlet location postulates an inverse relationship between the attractiveness of a location and the travel time to that location for a particular customer. (T) 8. The location set-covering problem locates facilities to minimize the distance that the most distant customer would travel to reach a facility (known as maximal service distance). (T) 9. The advantages of substituting communication for transportation become possible when face-to-face interaction between server and customer is not necessary. (T) 10. The absence of quantifiable criteria considerably diminishes the importance of facility location methods in the public sector. (F) 11. Geographic information systems are a visual method of displaying data. (T) 12. Cross-median is an approach to the location of a single facility using the metropolitan metric to maximize the total distance traveled. (F) 13. Marketing intermediaries are business entities in the channel of distribution between the final customer and the service provider. (T) 14. Use of the center of gravity formulas for solving the single location problem is incorrect. (T) 15. The greedy algorithm is a solution procedure that is used for measuring the attraction of a customer to a facility. (F) 16. The Euclidian metric is used for locating a facility in an urban area. (F) 17. For a multisite location problem the services to provide at each site may vary. (T) 18. The median has the property of minimizing the sum of absolute deviations from it. (T) 19. The Euclidian location model is used when vector travel is involved. (T) 20. The success of L.L. Bean questioned the necessity of always finding a location that is convenient for customers physical access. (T) Multiple Choice 1. A strategy that may be used by urban retailers to locate multiple sites close to each other in high-density areas despite the potential for cannibalization is called a. competitive clustering. b. marketing intermediaries. c. saturation marketing.* d. saturation clustering. 2. The optimization criteria governing the location of a service facility is determined mainly by the a. ownership of the service, i.e., public or private.* b. kind of service, e.g. food, banking, or postal. c. number of facilities to be located. d. average size of the facilities to be located. 3. A strategy defying the curse of cannibalization is called 10-15 a. competitive clustering. b. customized marketing. c. saturation marketing.* d. relationship marketing. 4. If you are going to open four facilities and have identified six candidate locations, how many different combinations of site selections are there? a. 4 b. 6 c. 15* d. 24 5. The following marketing phenomenon occurs when there exists within two city blocks the following hamburger outlets: McDonald's, Burger King, Sonic, and Wendys a. competitive clustering.* b. customized marketing. c. saturation marketing. d. relationship marketing. 6. A cruise line has the option of locating its dock in Miami or Pensacola because both cities have adequate port facilities. Which of the following considerations would seem to be the most germane? a. Adequate parking b. Location of competitors c. Freeway access to the port d. City served by major airport* 7. Which of the following items does the Huff retail location model not consider? a. The degree of accessibility* b. Travel time from the customer's location to the facility c. Facility size d. A parameter reflecting the importance of travel time by customers 8. The following are important in estimating spatial demand except a. defining the target population b. selecting a unit of area c. developing a visual representation or map d. distance traveled* 9. With respect to service facility location it is not prudent to open 10-16 a. a new facility next to one's competitors. 10-17 b. multiple facilities within the same area thus cannibalizing one's own business. c. an information based service in a downtown area with high real estate prices.* d. all of the above. 10. A strategy of locating sites in a number of states (or countries) can reduce the overall risk of financial crisis from a regional economic downturn. This strategic approach to facility location is known as a. competitive positioning. b. flexibility.* c. demand management d. focus 11. In the Huff retail location model, the value of estimates ______________. a.amount of travel time to reach a service facility b. the importance of convenience in attracting customers c.the importance of travel time in attracting customers* d. the cost of transportation for a customer 12. In the decision to locate a public health clinic, which one of the following criteria is not appropriate? a. Maximize utilization b. Minimize distance per capita c. Minimize distance per visit d. Minimize cost* 13. Car dealerships are often located near one another along a motor mile to allow customers to make comparisons. This location strategy is known as a. competitive Clustering* b. saturation Marketing c. marketing Intermediaries d. competitive Positioning 14. The concept of focus in location decisions considers all but one of the following issues: a. Cannibalization b. Multisite locations c. Formula facility d. Competitive changes* 15. Site selection considerations include all but one of the following a. Access b. Environment 10-18 c. Government d. Weather * 16. The disadvantage of saturation marketing lies in which of the following areas? a. Increased advertising costs b. Cannibalization * c. Increased requirement for supervision d. Reduction in customer awareness 17. The ___________ is a variation on the location set-covering problem. a. Huff model b. Greedy algorithm c. maximal covering algorithm* d. shortest path tree 18. Cross-median approach is used for _________ location. a. single facility * b. multiple facility c. (a) and (b) d. none of the above 19. Competitive clustering is common the ________ industry. a. healthcare b. lodging* c. communications d. banking 20. ArcView, a GIS for business applications, finds applications in all but one of the following: a. Mapping customer databases b. Determining site locations c. Forecasting demand * d. Improving delivery service 10-19 ... View Full Document

End of Preview

Sign up now to access the rest of the document