Abstract/Resume
Several sophisticated menu analysis methods have been compared in studies using theoretical restaurant menus. Institutional and especially hospital cafeterias differ from commercial restaurants in ways that may influence the effectiveness of these menu analysis methods. In this study, we compared three different menu analysis methods - menu engineering, goal value analysis, and marginal analysis in an institutional setting, to evaluate their relative effectiveness for menu management decision-making. The three methods were used to analyze menu cost and sales data for a representative cafeteria in a large metropolitan hospital. The results were compared with informal analyses by the manager and an employee to determine accuracy and value of information for decision-making. Results suggested that all three methods would improve menu planning and pricing, which in turn would enhance customer demand (revenue) and profitability. However, menu engineering was ranked the easiest of the three methods to interpret.
Plusieurs methodes raffinees d'analyse des menus ont ete comparees dans des etudes utilisant des menus de restaurant theoriques. Les differences entre les cafeterias d'etablissement en particulier celles des hopitaux - et les restaurants commerciaux sont telles qu'elles peuvent influencer l'efficacite de ces methodes d'analyse des menus. Dans la presente etude, nous avons compare trois methodes d'analyse des menus - l'analyse en fonction des revenus, l'analyse en fonction des buts et l'analyse marginale pour evaluer leur efficacite relative dans la prise de decision. Les trois methodes ont ete utilisees pour analyser les donnees sur le cout des menus et les ventes dans une cafeteria representative d'un grand hopital urbain. Les resultats ont ete compares a des analyses informelles menees par le gestionnaire et un employe pour verifier la justesse et la valeur de l'information necessaire a la prise de decision. Ces resultats portent a croire que les trois methodes amelioreraient la planification des menus et l'etablissement des prix, qui en retour accroitraient la clientele (revenus) et le profit. Des trois methodes, l'analyse en fonction des revenus a ete consideree comme la plus facile a interpreter.
INTRODUCTION
Menu analysis has been defined as the systematic evaluation of a menu's cost and/or sales data to identify opportunities for improved performance (1). Several sophisticated menu analysis methods have been described in the literature, such as menu engineering (2), goal value analysis (3), marginal analysis (4), and analyses that use two or more criteria to evaluate menus (5-8). All except goal value analysis and marginal analysis use matrix approaches based on portfolio analysis (9), which is used in strategic marketing to classify products or business units by comparing them to the means on two or more dimensions such as share of the market and growth potential (10,11). These methods have been described and compared in published studies using theoretical commercial restaurant menus, but have not been studied in a hospital cafeteria setting.
Decision-making is a key function of management, and involves identifying and selecting alternatives (12) by using a combination of intuition, information, and experience (8). Although managers always use their intuition and experience in decision-making, better information would enable them to evaluate the consequences of decisions more accurately, and to consider all alternatives (12).
Morrison (13) noted that menu planners were reticent about using menu analysis, despite the availability of information technology. Planners considered traditional criteria, such as product availability and staff skills, more important than profitability or customer demand. In addition, despite the long-term use of portfolio analysis in business, no empirical evidence in the literature proves that it has positive effects on decision-making. Armstrong and Brodie (9) found that it might even have a negative effect. Although menu analysis methods are based on less controversial dimensions than those used in strategic marketing, their role in decision-making must still be evaluated empirically. This will enhance their credibility and use by menu planners.
Institutional and, more specifically, hospital cafeterias differ from commercial restaurants in ways that may influence the effectiveness of menu analysis methods. Hospital cafeterias cater to a relatively captive and diverse clientele, which traditionally has expected low-priced menu items. Commercial restaurants cater to a more distinct target market that will pay higher prices for special occasion or infrequent meals. The clientele of hospital cafeterias requires more variety than does the clientele of commercial restaurants, because hospital clientele tend to eat in the same place every day.
Canadian healthcare reform has brought and will continue to bring many changes to healthcare management and delivery. Reform has also led to changes in hospital cafeterias. Traditionally, non-patient foodservices were financed by a government subsidy formula based on dollars per number of non-patient meals. This subsidy and revenues from low-priced menu items were intended to cover food expenses. Labour and operating expenses were generally expected to be part of the patient funding, as the same staff prepared the cafeteria meals in the same facilities. Now, non-patient foodservices must be self-sufficient, if not profit-generating. An opportunity has been created to maximize profit through greater marketing of cafeteria menus and growth of branded products. This opportunity, along with low profit margins, means that hospital cafeteria managers must be able to analyze their menus effectively perhaps even more effectively than do commercial restaurant managers.
The purpose of this study was to identify and analyze the relative effectiveness of the three methods - menu engineering, goal value analysis, and marginal analysis - for decision-making in hospital cafeteria menu management. Effectiveness was evaluated in terms of information value, accuracy, and ease of interpretation.
Menu analysis methods defined
Information about the various menu analysis methods is essential to a full understanding of the study reported in this paper.
Menu engineering - This matrix menu analysis method was developed by Kasavana and Smith (2). Its two dimensions are the contribution margin (CM) and the menu mix (MM). The CM is the difference between the selling price and direct, usually food, costs for each menu item. The MM is calculated as the percentage of an item's sales relative to total menu item sales. The CM and MM percentages for each item are compared with a mean menu CM and with an MM popularity rate, respectively. The MM popularity rate is calculated by taking 70% of the total MM percentage (100%) divided by the number of menu items. Kasavana and Smith (2) state that this 70% rule was based on extensive practical experience, and that it is unrealistic to expect each menu item to share equally in sales. It is unclear whether this rule applies to hospital cafeterias.
Once the comparisons with the two dimensions have been made, menu items are classified as stars (above the mean CM and MM popularity rate), dogs (below the mean CM and MM popularity rate), plowhorses (below the mean CM but above the MM popularity rate), or puzzles (above the mean CM but below the MM popularity rate). Various strategies to increase item CM and/or MM, such as repricing (changing the cost or selling price), repositioning (promoting or renaming the item), replacement (with a new item), or retaining the item, have been described (2).
Hayes and Huffman (3) and Atkinson and Jones (1) compared several matrix menu analysis methods and concluded that each could give different results. Beran (4) and Hayes and Huffman (3) pointed out that categories may be influenced by high or low menu item performers that can distort the true picture. Removing poorly performing items from the menu also changes the means for subsequent menu analysis, with the potential for only one menu item remaining (3). Although this is an unlikely scenario, it illustrates the limitations of menu analysis in management decision-making. Goal value analysis - Hayes and Huffman (3) propose goal value analysis as an alternative to matrix analysis. The formula is as follows: Goal Value = (1 - food cost %) x average numbers sold x average selling price x 1 - (direct cost % + food cost %) .
The use of this formula establishes an index based on expectations for or past performance of menu items. The score for each menu item is determined from actual sales and costs and compared with the goal value.
A potential problem is the establishment of a goal value based on expectations that may be unrealistically high or low. This could lead to unnecessary changes in the menu, or failure to realize the full potential of the menu.
Marginal analysis - Beran (4), also proposes marginal analysis as an alternative to matrix analysis. With this method, the menu engineering spreadsheet is modified to provide the calculations necessary for graphing individual item total CM (on the Xaxis) to total sales volume (on the Y-axis). The slope of each graph shows the CM in relation to other items. Items with more horizontal lines have a greater CM than items with more vertical lines. The length of the line, or the distance of the item's plot from the intersection of the X- and Y-axes, indicates the volume of sales. The graph shows how menu items actually compare with each other. Matrix analysis, on the other hand, uses comparisons with means or rates. Beran (4) suggests that a manager could evaluate the individual item and the overall menu performance with marginal and matrix analyses.
METHODS
Study site
The site selected for this project was a 25-seat, cafeteriastyle coffee shop in an 800-bed, acute care hospital and community health centre. The coffee shop is open weekdays from 10 a.m. to 6 p.m., is staffed by three employees, and serves approximately 260 customers a day. The clientele is representative of the larger foodservice operations in this institution, which serve a mix of staff, visitors, and outpatients. The menu is a combination of both static and ten-day cycle items, which again is representative of the larger cafeterias at this institution.
Sales data collection - Sales data were collected from the daily sales summary reports printed from the cash register at closing, and entered on a Microsoft Excel 5.0 worksheet (Microsoft Corporation, Redmond, WA, 1993). (At the time of data collection, the cash register was not able to produce weekly or monthly reports.) Sales data were collected for 12 weeks in spring 1995. Food costs for menu items were determined from current ingredient and recipe cost reports from Carex Dietary Information Systems Inc. (London, ON N6A 1A9). Pre-tax selling prices were used. Budget information and other operating percentages were used as applicable to the specific menu analysis methods.
Menu analysis - The menu was analyzed on Microsoft Excel 5.0 spreadsheets, using menu engineering (2) and goal value analysis (3). Table 1 summarizes key sections of the spreadsheets to facilitate comparison. Sales data for menu items were summarized from the worksheet in fourweek periods to ensure that the data collection period was representative of the operation (2) and that cycle menu items were represented equally. Sandwiches, soups, entrees, and combinations were analyzed together because they were common lunch selections and had similar prices. The cycle menu items were grouped under the headings "specialty sandwich" and "special entree" for this analysis, and the mean CM was used. The individual cycle menu items were also analyzed separately to determine their performances within their groupings.
Scatter plot charts were prepared from the menu engineering spreadsheet (Table 1) for the menu engineering matrix (Figure 1) and for the marginal analysis graph (Figure 2). It was not necessary to prepare a separate spreadsheet for marginal analysis (4). For the menu engineering matrix, CM by item (Table 1, Column F) was selected for the X-axis and MM percentage by item (Table 1, Column C) formed the Y-axis. For marginal analysis, total CM by item (Table 1, Column L) formed the X- axis and total number of sales by item (Table 1, Column B) formed the Y- axis.
We compared the four- and eight-week analyses for differences in results and to establish the optimum time for data collection. We also compared the analyses for groupings to determine the effects of evaluating the whole menu rather than menu groups.
Manager and employee evaluations - Before they viewed the menu analysis spreadsheets and graphs, the foodservice manager and a staff member were asked individually to identify menu items as good, poor, or uncertain, and to decide whether to retain, reprice, reposition, or replace them. We assumed that they would base their evaluations on intuition, experience, and recent demand and/or cost information. We then compared their evaluations with the results of the three menu analyses.
The menu analysis spreadsheets and graphs were then explained to the manager, who ranked the three menu analysis methods according to their ease of interpretation. Menu decisions were revised as a result of more detailed information. Ideas to improve menu item prices, change portion sizes, or replace items were identified, and the potential for improved performance was tested with the menu analysis. The manager then selected three menu changes to introduce. Because of seasonal fluctuations in demand, it was not practical to compare actual menu analyses before and after changes.
RESULTS
There was no difference in menu item classifications in the four- and eight-week sales data sets, and we therefore selected the former to facilitate data collection. The analyses of the cycle menu items within their own groups-specialty sandwiches and specialty entrees - indicated which items were relatively poor performers.
The menu engineering spreadsheet (Table 1, Column S) classifies items in four categories, which enables the decisionmaker to identify the problem area(s) of a menu item easily. Goal value analysis (Table 1, Column Y) classifies items as above or below the goal value so that the decision-maker must analyze and compare each variable in the formula to identify menu item problem areas. In this study, the menu items classified as plowhorses and stars with menu engineering had scores above the goal value, while the puzzles and dogs had scores below the goal value. With goal value analysis, the reason for high or low scores is not readily apparent, but menu engineering clearly shows that a plowhorse has above average popularity and a below average CM.
The menu engineering matrix (Figure 1) presents the spreadsheet in a visual format. The degree of a menu item's classification is portrayed by the distance the item is plotted from the MM popularity rate or from the mean CM. For example, item 2 (designer sandwich) is very popular but only slightly below the mean CM. With this information a decision could easily be made to, either moderately increase the price or reduce the portion size, without affecting demand significantly.
The marginal analysis graph portrays the data differently. A diagonal line dissecting the chart facilitates identification of high slope (lower CM) versus low slope (high CM) menu items. The popularity, or total sales by item, is easily seen bv the distance of the plotted item from the Y-axis.
Similar decisions are possible using both menu engineering and marginal analysis. For example, item 7 (chili) is classified as a puzzle with menu engineering (Table 1 and Figure 1), which means that it had poor sales compared with the other items but an above average CM. Similarly, marginal analysis (Figure 2) shows that it has a low slope, or higher CM, and a relatively short plot from the Y-axis, or low sales. A decision strategy would focus on promoting sales of this item by lowering the price or by repositioning (i.e., name or accompaniment changes).
Manager and employee evaluations The manager and coffee shop employee generally agreed in their evaluations of the menu items, with some differences in the evaluations of individual cycle sandwiches and entrees. However, their evaluations for items 1, 3, 6, and 9 differed from the menu engineering spreadsheet (Table 1, Column S) classifications. The manager and the employee considered egg/tuna sandwiches good sellers with poor profit margins (plowhorses), cycle sandwiches poor in both categories (dogs), cycle entrees uncertain (puzzles), and soup with a roll a star. Menu engineering classified these items as dogs, puzzles, stars, and a plowhorse, respectively.
The manager placed more weight on her knowledge of food costs or production difficulties, and the employee based her evaluations on customer demand. A potential limitation may be that only two people evaluated the menu, and they based evaluations on intuition or experience rather than referral to current cost and/or sales data. Morrison (13) also reported that menu decisions were commonly based on intuition and experience.
Possible study limitations - The accuracy of the menu sales data depended on the staff's accuracy in keying in sales on the cash register. We expected that some errors, such as one item being keyed in as another, would be made daily, but that this should not affect the results as the analysis period was four weeks. To further control for this limitation, we regularly checked cash register key codes for accuracy with the menu. Staff members were told that correct key entry was important for the menu study. The student research assistant observed cash register operation for several days and promptly investigated possible errors on the daily sales summary. Ease of using the methods - The matrix format of menu engineering was found to be the easiest to interpret and use. The labels of menu engineering, with their associated meanings - dog, star, puzzle and plowhorse - facilitate communication of the results to all staff.
Marginal analysis was ranked the second easiest method, and essentially was deemed to be another way to present the information. Both the menu engineering and the marginal analysis charts can be produced from the spreadsheet.
Goal value analysis was found to be more difficult to interpret for menu decision-making. The manager could not easily identify the cause of a menu item's low score, and this method was therefore not found to be useful for menu planning and improvements.
The 70% MM popularity rate of menu engineering was determined to be acceptable for this operation, although for this particular menu a 100% MM popularity rate of 10% would not have resulted in any more dogs or puzzles. As 10% of all revenues must be paid to the provincial government by non-patient foodservices, this was considered to be a direct cost, like food cost, and was included in the calculation of the CM. (Other direct costs, such as labour, could also be included.)
Effects of menu analysis on decisions and profits - When she reviewed the information from the menu analyses, the manager identified four menu changes for testing. Table 1, Column T, summarizes the manager's decisions. When these changes to prices and/or menu description were tested using the same data, except for a projected small increase in sales for item 7 (chili), menu CM increased by 1%. This may seem small by commercial standards, but is significant in hospital cafeterias, which have low profit margins ( < 5%).
Because she was concerned about customer dissatisfaction, the manager did not make other changes suggested by the menu analysis, such as replacing dogs or repricing plowhorses. Her decision illustrates that intuition and experience remain important to management decision-making, even with improved information from menu analysis.
DISCUSSION
Maximizing profits with a low CM - Traditionally, government funders, hospital administrators and staff have viewed non-patient foodservices as a service to employees. The expectation that hospital cafeterias should offer items at cost or low prices persists and is accentuated by staff members' job insecurity and cautious spending. Sophisticated menu analysis is, therefore, perhaps even more necessary for hospital cafeteria managers than for restaurant managers; hospital cafeteria managers must try to maximize profits with a low CM, while high-priced restaurant or hotel menu items have a high CM and tend to be classified as stars or puzzles, depending on customer demand.
Menu analysis of static and cycle items - A hospital cafeteria menu differs in structure from those of restaurants, hotels, and commercial operations. To meet the needs of the diverse clientele, hospital cafeteria menus are a combination of cycle and static menus (12). The static menus correspond to those in commercial establishments or franchises, and are often part of a food fair concept designed to serve specific market segments. Cycle menus are intended to satisfy a diverse and captive clientele, which wants variety. Because they correspond to patient menu production schedules, cycle menus also enable hospital foodservice operations to benefit from economies of scale. This study demonstrated that combination cycle and static menus could be evaluated by menu engineering.
Profitability and demand as pricing components - Prices in hospital cafeterias traditionally have been established as a mark-up on the item food cost (12,14). This mark-up is based on desired or budget food cost percentages. Problems with this approach are that each menu item must bear the same multiplier for fixed expenses and profit, which may result in unfair prices for outsourced or convenience items that require little or no labour. This pricing approach also does not take into consideration the marketing potential of setting prices higher or lower than the mark-up to stimulate greater demand for certain items or combinations of items. This study identified an efficient and systematic method of menu analysis that would enable managers to evaluate profitability and customer demand as equal components of menu item pricing.
Ideal food cost percentage -The menu engineering spreadsheet also calculates the ideal food cost percentage (Table 1, Box K). This can be compared with the actual food cost percentage, calculated from inventories and purchases, for variance analysis. Thus, this menu analysis method provides additional and valuable information for foodservice managers.
Labour costs - Labour costs are another dimension of menu analysis. However, we did not know the labour costs per item, and the manager could not justify the time and expense required for this determination. A study of menu planning criteria used by upscale restaurants showed that, for similar reasons, labour cost was not a major criterion (13).
A three-dimensional matrix by LeBruto et al. (5) and the software program Menu Dynamics 2.0 (MP Talbert, Ithaca, NY 14850, 1996) do not use labour costs for specific items, but rather, classifications of high and low. These classifications could be based on intuition and experience, or calculations of direct menu item production and service. If based on intuition or experience, the information used in menu analysis could be wrong. As decisions usually involve varying elements of intuition, experience, and information, care must be taken to ensure that the latter is not compromised.
CONCLUSION
In many ways, the distinctions between commercial and institutional foodservice operations are becoming less apparent. However, differences in clientele, management, and menu planning in hospital cafeterias indicate a need to evaluate menu analysis methods originally developed for commercial ventures. Determining the relative effectiveness of the three methods should also prove valuable to commercial restaurants.
This study demonstrates that information is interpreted more easily with menu engineering than with marginal analysis or goal value analysis. The use of menu engineering is therefore more likely to improve management decisions about hospital cafeteria menus and to maximize sales and profit.
RELEVANCE TO PRACTICE
Under government subsidized funding guidelines, hospital cafeteria managers evaluated menu items by their adherence to a desired food cost percentage (14) and by customer surveys (12). When menus are evaluated only on the basis of food costs, the profit potential of individual items and the interactive roles played by menu items receive less attention. A low food cost percentage does not necessarily mean that a menu item will have a high profit margin (14). Customer surveys are expensive and time-consuming, and their value may be limited by low response rates and the influence of a recent good or bad meal.
Menu engineering, which uses a computer spreadsheet program such as Microsoft Excel or Lotus 1-2-3 plus accurate cost and sales data, can enable managers to evaluate menu performance more effectively and to forecast the effects of menu changes easily. However, the final decisions must remain with the manager, who can make the best judgments with a combination of intuition, experience, information and creativity.
[Reference]
References
[Reference]
Atkinson H, Jones P. Menu engineering: Managing the foodservice micromarketing mix. J Restaurant Foodservice Marketing 1994;1(1):37-55. Kasavana MK, Smith DI. Menu engineering. Okemos, MI: Hospitality Publications Inc, 1990.
3. Hayes DK, Huffman L. Menu analysis: a better way. Cornell HRA Q 1985; 25:64-70.
[Reference]
Beran B. Menu sales mix analysis revisited: an economic approach. Hospitality Research J 1995;18(3)/19(1):125-44. LeBruto SM, Quain WJ, Ashley AA. Menu engineering: a model including labor. FIU Hospitality Rev 1996;13(1):41-51. Merricks P, Jones P. The management of catering operations. London, UK: Cassell Publishing,1988.
Pavesic DV. Prime numbers: finding your menu's strengths. Cornell HRA Q 1985;26(3):70-7.
[Reference]
Miller J. Menu pricing and strategy. Boston, MA: CBI Publishing Company Inc,1980:77, 18.
Armstrong JS, Brodie RJ. Effects of portfolio planning methods on decision making: experimental results. International J Research Marketing 1994;11:73-84.
10. McCain G. Black holes, cash pigs, and other hospital portfolio analysis
problems. J Healthcare Marketing 1987;7(2):56-64. 11. Hedley B. Strategy and the business portfolio. Long Range Planning 1977;10:9-15.
[Reference]
12. Spears MC. Foodservice organizations: a managerial and systems approach, 3rd ed. Toronto, ON: Prentice-Hall,1995:589,185. 13. Morrison P. Menu engineering in upscale restaurants. International J Contemp Hospitality Management 1996;8(4):17-24. 14. Keiser J, DeMicco FJ. Controlling and analyzing costs in foodservice operations, 3rd ed. Don Mills, ON: Maxwell Macmillan Canada Inc,1993:380.
[Author Affiliation]
LINDA L. MANN, MBA, PDt, Department of Applied Human Nutrition, Mount Saint Vincent University, Halifax, Nova Scotia; DONNA MACINNIS, PDt, Food and Nutrition Services, Queen Elizabeth II Health Sciences Centre, Halifax, Nova Scotia; NICOLE GARDINER, PDt, Clover Group, Halifax, Nova Scotia
