Sunday, August 25, 2019

Planning Restaurants in Locations with Predominantly Generation-Y Demographics

Stacy, a college graduate with concentration in hospitality and food service, is interviewing for a Senior Analyst position.

Question # 1
Interviewer: One of our clients, a national restaurant chain, wants us to a find a new location in our city where Generation-Y demographics predominate. What additional factors would you consider in your initial plans?

Stacy: I would recommend customizing the interior and exterior décor as well as other family attractions in order to be demographically more appealing to the Gen-Y population; for instance, for a fast food chain I might recommend annexing a kids’ play area.   





Question # 2
Interviewer: Our interns randomly surveyed a group of Gen-Y students and presented the summary results at a management meeting. Explain to me their findings as shown in the Generic Taste table.  

Stacy: They like higher than average salt content in their meals. They are young folks so they still have the sweet teeth. Obviously, they love the savory taste, but they have not developed much taste for sour or bitter yet.


Question # 3
Interviewer: What is the reason we are publishing the stats between the 5th and 95th percentile? If you have to extend the curve, how would you go about doing it?

Stacy: Since the survey results are going to be used in important business decision-making, you avoided using the outlier data points, which is extremely prudent. If I were to extend the distribution, I would use 1st to 99th percentile, avoiding the minimum to maximum range.





Question # 4
Interviewer: After surveying the generic taste, our interns asked them to sample and score the actual food sample we received from our restaurant client. This survey comprised an overall score, as well as scores in three relevant categories, on a scale of 1 to 5, where a score 5 was considered excellent. Interpret the actual scores.

Stacy: The overall median score is 4.0, meaning 50% thought the food was good to excellent while the other 50% rated it average to good. Given their higher liking for salt and sweet, they had some issues with the salt and sweet components in the samples they tried, thus suggesting that the foods could use a notch more salt and sweet to meet their tastes. Obviously, it’s typical of this age group. In other words, while developing the menu items for this location, the salt and sweetness levels need to be permanently bumped up.


Question # 5
Interviewer: What is the statistical difference between generic savory and actual savory tastes?

Stacy: The distribution of the generic savory taste is more compact ranging between 4 and 5, while the distribution reflecting the actual score has a wider range of 3 to 5, signifying that there is some room for improvement in catering to their actual taste, absent which the bottom 10% that gave a score of 3 might fall off.





Question # 6
Interviewer: We know that the folks who rated the food samples below 4 would come along once the menus are adjusted to their tastes, but we wanted to find out if their higher-score counterparts who rated the food 4 and above would continue to feel that way. How would you interpret the actual vs. the predicted scores?

Stacy: The table shows that the predicted scores are as good; for example, the median predicted score was 4.32 relative to the actual median of 4.25, while the standard deviation fell from .31 to .28, further authenticating the stability of the higher scores. In other words, their high-score counterparts would continue to like the food.





Question # 7
Interviewer: What is the purpose of inserting the graph here?

Stacy: The graph is more telling. When you graphed the same table data, it makes the inconsistencies stand out more vividly. The predicted score after the 75th percentile has flattened out, proving their generic tastes didn’t justify the abrupt spike in scores from 4.50 to 5.00. That jump from the 90th to 95th percentile is irrational. Likewise, the sideways movement at the lower end of the distribution, i.e., between the 5th and 25th is not logical either.





Question # 8
Interviewer: Any idea how we developed the regression model?

Stacy: Yes, you developed the regression model by joining the two pieces of data from the two questionnaires. Specifically, you used the actual scores as the dependent variable and their generic tastes as independent variables in the regression model.

Question # 9
Interviewer: Interpret the regression output and the impact of the variables.

Stacy: The regression model shows that sweet and savory are stronger than salt, thus indicating that even the high-score group would require more adjustment to the salt level in individual entrées than the sweet and savory. The strength of an independent variable in regression models is evidenced by the higher T-stat and lower P-Value. Clearly, the sweet has the highest T-stat and lowest P-value, closely followed by the savory.     

Thank you,

Sid Som, MBA, MIM

Homequant, Inc.
homequant@gmail.com   


No comments:

Post a Comment