Does Assortment affect Restaurant Sales? For the partial completion of the course ECON F435 Marketing Research
Submitted by: Group 15 Shubham Arora - 2012B3A7522P Gautam Singh - 2012B3A3467P Aditi Sinha - 2014B3PS630P
Submitted to: Prof. Anil K. Bhatt
Problem Definition The owner of a restaurant located in Connaught, Pilani is wondering whether his sales are declining because of a reduced assortment of menu items.
Management Decision Problem Should the menu assortment mix of the restaurant be changed/improved upon to attract more customers?
Marketing Research Problem Definition 1. Determine the criteria for the choice of restaurants in Pilani. 2. Why do consumers eat at the restaurants they do? 3. How does restaurant menu assortment affect the customer in the choice of the restaurant?
Research Questions: 1. What are the factors on which consumers evaluate the choice of restaurant at a given time? 2. Is the price a significant factor in determining restaurant sales? 3. Is the restaurant menu assortment a significant factor in determining restaurant sales? 4. Is the patron loyalty a significant factor in determining restaurant sales? 5. Is the quality of food a significant factor in determining restaurant sales? 6. Is the veg/non-veg offering a significant factor in determining restaurant sales? 7. Does having large number of unique items for a restaurant affect the choice of restaurant for a customer?
Theoretical Framework The number of available options can influence consumer choice in multiple ways. Decades of research suggest that choice increases satisfaction and that larger assortments increase the likelihood that consumers will find an option that matches their preference. People actively seek variety whether to satisfy the need for stimulation (Berlyne 1960) or for other reasons, and larger assortments help consumers satisfy these needs. Would the number of alternatives influence the process by which people choose and, consequently, what they select? When faced with difficult decisions, consumers often search for reasons to justify their choices. While normative theories focus on the utility associated with various aspects of the decision, other research notes that decision making is often driven by a reason-based analysis (Shafir, Simonson, and Tversky 1993). This framework suggests that consumers seek reasons to resolve conflict and justify the options they select. In such instances, consumers often focus their decision processes on the choice of good reasons rather than on the choice of good options (Simonson and Nowlis 2000).
Hence, we set out to explore if assortment in restaurant menus affects sales of restaurants in Pilani.
Common assortment metrics involve issues such as assortment size, reflecting both the breadth (i.e., number of categories) and the depth (i.e., number of items within a category) of the available product lines; the type of items (e.g., overall attractiveness); the relational properties of the items (e.g., item similarity); and the variety of items over time. (Bell, Tang, and Ho 1998) This research has contributed to significant advancement in understanding the impact of assortment on consumer choice. It takes a consumer’s perspective to examine how restaurant menu assortment influences restaurant sales. The report investigate the impact of product assortments, along with prices, patron’s loyalty, food quality as well as the vegetarian/non-vegetarian offerings of the restaurant.
Data Our dataset is a multi-restaurant cross-sectional data from the students of BITS Pilani collected from restaurant-goers from Connaught after they have had their meals over a 3 week period. Data was collected on weekdays so as to avoid large groups of restaurant-goers.
Along with respondent data, we also collected the exhaustive list of menu items offered by each restaurant with their prices.
Individuals who dined in groups of 2-3 were only surveyed, in order to account for a larger role of every individual in the decision making for the restaurant to dine out in.
Diners from 9 restaurants were surveyed, since these are the restaurants receiving a significant footfall and account for the major amount of transactions at the BITS campus. These restaurants are namely: Restaurant
Percentage of Visits
Sharma’s
14.60%
Kamal's
16.90%
Noble's
20.70%
Annapurna
6.20%
Golden Dragon
12.20%
Greenfield
12.20%
Vijay's
8.10%
Café Coffee Delite
5.70%
Blue Moon
3.40%
Before the survey, each respondent was asked the approximate number of visits to each restaurant out of the last 10 trips they made to Connaught. Respondents who could not be certain about their responses were dropped.
The resultant data set contains responses from 210 respondents, who had volunteered to share their restaurant bills with us. The items consumed by the respondents were recorded.
From the restaurant specific data, we made an account of: 1. Cuisines served 2. Variety of each type of food available 3. Number of serving size options available per item 4. Number of menu items
Model The model assumes that the sales of a restaurant depend upon the following factors: 1. Pricing 2. Patron loyalty 3. Location 4. Menu assortment 5. Quality of food served 6. Ambience 7. Vegetarian/Non-Vegetarian
Pricing PRestaurant = a. PWholesome + b. PBeverage + c. PDessert. PRestaurant is the weighted average price of all menu items at the restaurant PWholesome is the weighted average price of menu items sufficient for wholesome menu items PBeverage is the weighted average price of beverages in the menu PDessert is the weighted average price of beverages in the menu
Patron loyalty LRestaurant = Σ (Li, Restaurant . Billi, Restaurant). Li, Restaurant = (nRestaurant + 1) / nRecent. LRestaurant is the weighted average of each individual patron weighted by the bill amount Li, Restaurant is the individual patron loyalty of patron i to the restaurant Billi, Restaurant is the bill amount borne by the patron i at the restaurant nRecent is the number of recent visits made to dine out at Connaught nRestaurant is the number of visits to the restaurant out of the last nRecent number of visits
Location Location parameter was dropped out of the model since all the restaurants in Connaught are within 40m of each other. Hence, the effect of location is nullified in the present context.
Menu assortment ARestaurant = v. nuniqueRestaurant w. CuiRestaurant + x. nitemsRestaurant + y. VarRestaurant. ARestaurant is the assortment metric for each restaurant nuniqueRestaurant is the number of unique items served at the restaurant CuiRestaurant is the number of cuisines served at the restaurant nitemsRestaurant is the number of items in the menu VarRestaurant is the number of items of each menu item category available The normalised mean of all components of assortment is taken as a proxy for Menu Assortment.
Quality of food served QRestaurant = Σ (Qi, Restaurant . Billi, Restaurant). QRestaurant is the weighted average of the self-evaluated food quality ratings by individual respondents, weighted by their bill amounts Qi, Restaurant is the individual quality rating of patron i to the restaurant Billi, Restaurant is the bill amount borne by the patron i at the restaurant
Ambience Since every restaurant at Connaught has a similar seating style, i.e. furniture, location of seating, etc., the effect of Ambience is nullified. Hence, Ambience has been dropped from the dataset.
Vegetarian/Non-Vegetarian VNVRestaurant = 0, if Vegetarian;
1, if Non-Vegetarian food also served
Sales SalesRestaurant = b0 + b1.PRestaurant + b2.LRestaurant + b3.ARestaurant + b4.QRestaurant + b5.VNVRestaurant + e. The sum total of all bills from all patrons interviewed for each restaurant is taken as a proxy variable to train the model.
The model is comprehensive because of the following assumptions:
The decision to eat at a restaurant in Connaught might be planned as well as impulsive.
We consider menu assortment is defined at the category level which shows that choices are realised at the category level.
The choices are independent of the restaurants.
Connaught is assumed to be a closed market, hence neglecting the competition from All Night Canteen and Food King.
Research Approach A competitive analysis problem was presented to us. The approach we took is described below:
Exploratory Phase An exploratory research was carried out, wherein the team members set out to talk to frequent restaurant goers in BITS Pilani. These were individual casual setups where we enquired about the reasons why people go to the restaurants they go to and how they make the choice to eat at a specific restaurant. A total of 12 respondents were interviewed for this activity.
Based on the exploratory research a number of candidate factors were jotted down and data was collected upon those factors. Data Collection and Fieldwork We set out to collect restaurant level data from each restaurant at Connaught. Menus were scanned to gather data about the pricing and the assortment for each restaurant. Also, patrons were interviewed using systematic random sampling, i.e. every third restaurant goer was interviewed after they had just ate. They were asked about the items they consumed at the restaurants and their evaluation of food quality and their own loyalty to the restaurant. 185 respondents were surveyed after their meals, out of which 14 were dropped on suspicion of providing wrong information.
Data was only collected over a period of 4 weeks on weekdays, so as to avoid large crowds of restaurant-goers where the decision making of the individual is reduced. Data Analysis and Preparation Since the restaurant specific variables are all intervals, we used a multivariate linear regression based analysis to evaluate the significance of each of the candidate factors determining the sales of restaurants. Results from the multivariate linear regression are reported below.
Results
Restaurant Blue Moon Annapurna Greenfield Vijay's Café Coffee Delite Sharmas' Golden Dragon Kamal's Noble's
Restaurant Blue Moon Annapurna Greenfield Vijay's Café Coffee Delite Sharmas' Golden Dragon Kamal's Noble's
Restaurant Blue Moon Annapurna Greenfield Vijay's Café Coffee Delite Sharmas' Golden Dragon Kamal's Noble's
Total Bill Amount Standard Deviation ₹ 56.79 ₹ 43.46 ₹ 58.45 ₹ 69.44 ₹ 89.37 ₹ 66.77 ₹ 98.45 ₹ 96.34 ₹ 137.47
Mean ₹ ₹ ₹ ₹ ₹ ₹ ₹ ₹ ₹
120.88 98.67 178.44 140.13 110.50 153.34 278.49 277.56 310.56
SalesRestaurant LRestaurant PRestaurant ARestaurant QRestaurant VNVRestaurant 70.12 -0.450789 7567 0.25634 3.1959 0 73.77 0.4856676 8560 0.30036 3.0145 0 76.98 0.2773343 10003 0.33035 3.0345 1 78.23 1.5006067 10560 0.36256 3.2715 0 84.76 2.6259567 11567 0.31807 3.2745 1 105.56 8.1405457 15786 0.36753 3.6033 0 120.23 11.523547 18916 0.39486 3.5645 1 124.4 15.239532 23005 0.53012 3.6493 1 140.12 18.14371 27050 0.51367 3.8413 1
No. of unique items
No. of cuisines 1 2 2 2 2 2 3 4 4
5 0 0 0 4 6 7 7 11
No. of menu items 61 88 83 70 78 134 135 151 148
No. of item categories 5 7 9 9 7 10 12 14 18
Conclusions Regression Statistics Multiple R R Square Adjusted R Square Standard Error Observations
0.997506049 0.995018319 0.986715516 792.2588675 9
ANOVA df Regression Residual Total
Intercept LRestaurant PRestaurant ARestaurant QRestaurant VNVRestaurant
5 3 8
SS MS F 376106297.7 75221259.53 119.8412647 1883022.34 627674.1132 377989320
Coefficients Standard Error -6583.3047 14875.3492 15181.2167 10574.6029 73.4995 139.0271 421.9018 559.0983 1612.8556 3664.0336 685.9177 850.4572
t Stat -0.4426 1.4356 0.5287 0.7546 0.4402 0.8065
P-value 0.6880 0.2466 0.6337 0.5053 0.6896 0.4789
Lower 95% -53923.3047 -18471.8892 -368.9468 -1357.3986 -10047.7346 -2020.6167
Significance F 0.001188483
Upper 95% 40756.6954 48834.3226 515.9457 2201.2021 13273.4457 3392.4522
The restaurants taken for our research are Blue Moon, Annapurna, Sharma’s, Kamal’s Golden Dragon, Greenfields, Nobles, Vijay’s and Cafe Coffee Delite., We ran our regression taking the sales of a particular restaurant as the Y value and the variables like price of the food in the restaurant, assortment of the food in the restaurant, quality of food in the restaurant and patron loyalty in the restaurant and veg/non-veg offering as X values. At 95% confidence level, If the p values are greater than 0.05, this means that our null hypothesis that the assortment does not affect consumer preferences is rejected. If it is less than .05, this means that our null hypothesis that the assortment does not affect consumer preferences is accepted. In our regression output, the p values for the variables:
1. Prices significantly affect the restaurant sales positively. 2. Quality of food significantly affect the restaurant sales positively. 3. Menu Assortment significantly affect the restaurant sales positively. 4. Patron Loyalty significantly affect the restaurant sales positively. 5. Serving Non-vegetarian food significantly affect the restaurant sales positively.
Our Recommendation to the client was to increase the menu assortment levels, in order to boost their sales.
Limitations Limitations of this research include anything that is beyond the control of the researcher: 1. This is true in research where the variables are uncountable and continuous such as adequacy, effectiveness, efficiency, extent etc. Therefore, all the factors involved in determining the factors involved that fixates a student’s decision to visit a particular restaurant and order a select item from the menu are unaccounted for. 2. The data might be influenced by extraordinary events which might affect restaurant goers’ decision making during the data collection phase. 3. The data used in the study is basically the consequent visit of a student visiting Connaught for over 4 weeks. Therefore the quality of the data may not be extremely accurate but our research is quite extensive which compensates it to the fullest. 4. The results calculated are indicative and have to be interpreted with utmost caution.
