Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

Assignment 3

Data and Analysis for Marketing Decisions

Formatting: Your report should follow the following formatting guidelines:

· Start the answer to each question on a new page. By question, I mean the numbered “big” questions, not the lettered sub-questions.

· Your answer to each question should be no more than five single-spaced, 11-12pt non-condensed font (e.g., Times), 1-inch margin pages. (Again, this refers to the numbered “big” questions, not the lettered sub-questions.)

· Concise answers are better than long, wordy ones

· We expect you to “write up” your responses, and submit as a properly formatted document – do not submit raw R code or Excel workbooks. In short, your submission should be easily readable by a human. (For those of you who have used them before, R Markdown notebooks are fine, and encouraged, but 100% not required.)

Submissions that do not follow these guidelines will be penalized. Make sure you check your final write-up document prior to submission. If you are collaborating with your teammates using Google docs, please double check that the length is still correct after exporting the file.

Grading: Unless otherwise noted, each sub-question is worth 2 points:

· A fully correct and justified answer will receive the full 2 points.

· An incorrect but thoughtfully attempted answer, or a correct but not fully justified answer will receive 1 point.

Data Access:

· Several of our data sets are provided by Wharton Customer Analytics, in conjunction with partner companies. These are real data sets, provided to us by these partner companies for educational purposes. Before you can download these data sets, you must sign a Data User Agreement (DUA), available on Canvas under the Modules tab. After you sign the DUA, you will be able to download these data sets from the Modules tab.

· All non-WCA data sets are available on Canvas under Modules > Non-WCA Data.

Question 1 (30 points)

Note: The five-page limit excludes the required appendix.

The purpose of this exercise is to learn conjoint analysis by doing it. You will develop, administer, and analyze a ratings-based conjoint analysis study. There are three parts to the assignment: designing the study, collecting the data, and analyzing the data.

Scenario: You are developing a more eco-friendly version of a product or service. You want to understand how much customers value and are willing to pay for this new eco-friendly option.

Study Design (5 points)

a. Identify a product or service you’d like to study, given the scenario above. Then develop a ratings-based conjoint survey to measure consumers’ preferences about this product. Specifically, in your write-up, you should:

o Describe your product.

o Describe three to five attributes that define products/services in this category, to be tested in your conjoint survey. You must include price as an attribute, as well as an attribute that matches the scenario. If you believe that price cannot be an attribute in your case, please email the professor before starting work on this assignment, or choose a different product/service to study.

o Identify levels for each of these attributes (at least two per attribute).

Then, based on the above, create a questionnaire composed of product profiles that can be rated by consumers. You should base these questions on a fractional factorial design, using the tool “Conjoint_Experimental_Designs.xlsm” (available on Canvas). The rating scale can be whatever you like, although a good and convenient option is a Likert scale from 1 (Very Unfavorable) to 7 (Very Favorable). Include a copy of this questionnaire as an appendix to this question.

Data Collection

Administer your questionnaire to at least 10 target consumers. You should collect this data using a tool like Qualtrics or Google Forms. If you’d like to use Qualtrics, you can create a free account with your Wharton email address at wharton.qualtrics.com.
You do not need to submit the survey responses as part of your report or appendix.
Note: These “customers” may be your group members. If you have trouble finding 10 people to survey, feel free to take your own survey multiple times, pretending to be a different person each time (e.g., “how would my brother/sister/friend/classmate/etc. respond to these questions?”). This assignment is not meant to test how many favors you’re able to call in from your friends.

Data Analysis (25 points)

b. Using regression in Excel, estimate the part-worths separately for each of your ten respondents. Report the coefficient estimates in a single table, where the rows are the coefficients, and the columns are the respondents. In an appendix, include the regression results for a single respondent. You do not need to report the full regression output for the rest of the respondents (even in the appendix). (5 points)

c. Using the data from one respondent: (8 points)

· Report the part-worths for that respondent.

· Report the relative percentage importance of each attribute for that respondent.

· Compute and report the willingness-to-pay for each of the non-baseline product features. (If you need to make an assumption about the $/util conversion rate, state it explicitly.)

· Determine and report the best and worst possible products for that respondent.

d. Plot the part-worths for the price variable for five respondents. Are these plots consistent with your expectations? Note: you may overlay the lines in a single chart to save space. (2 points)

e. Imagine you are planning to actually launch a new product in this product category. Generate two product profiles, X and Y, that represent two main competitors in the market (current and/or potential). Based on your ten respondents, what would you estimate the market share to be for X and Y if they were the only two products offered in the market? Assume that each consumer would pick the product with the highest utility for him or her (i.e., NO logistic probabilities).  Report what the two products are, and the expected market share. Include any details of the market share calculation in an appendix, not in the answer to the question. (4 points)

f. Generate three additional product profiles (A, B, C) that you are considering introducing to the market to compete with X and Y. Select these three profiles to match the scenario, and based on the new product candidates you judge as most promising based on your prior knowledge and intuition. Based on the ten respondents, which one of these three products/services will capture the greatest market share from X and Y? Report the candidate products, the market share that A would obtain if it competed with X and Y, the market share that B would obtain if it competed with X and Y, etc., and identify the new product/service with the highest share. Any details of the market share calculations should be included in an appendix. (6 points)

Final deliverable: Create a report containing the answers to all of the above lettered subquestions, and any specifically asked for results (and excluding any specifically not-asked-for results). In an appendix, provide a copy of your questionnaire, the regression analysis results for one respondent, and a copy of your market share analyses.

Remember: do not just copy/paste stuff from Excel. If you do not process the output into a “report,” with sentences or clearly labeled and described figures that directly answer the questions, you will get a zero.

Question 2 (10 points)

The file “car_data.csv” contains consumers’ responses to questions about attitudes toward buying a new car. The details of these questions can be found in the data key, “car_data_key.docx.” Both files are available on Canvas, under Modules.

We are interested in using these psychographic measures to understand the new car market, and segment customers into different buyer types. For the following questions, you should ignore the Purchased variable.

a. Using k-means clustering on the psychographic questions (Q1-Q17) and the Ideal_Price variable, identify and justify an appropriate number of attitudinal segments for new car buyers. For each of the segment, report how many respondents belong to it, and characterize the mean difference in the responses to each of the 17 questions across segments. (Hint: make sure you standardize the variables first. In R, this can be done using the scale() function.) (3 points)

b. Using the results from your cluster analysis, name each segment, and create a short, 1-3 sentence, descriptive profile of each. (2 points)

c. Ignoring, for the moment, the segments you just described, use principal components analysis (PCA) to identify and justify an appropriate number of factors (at least 2) describing consumers’ responses to the psychographic questions (this time NOT including the Ideal_Price variable). Report the number of factors chosen, and why you chose that many, and the (varimax rotated) factor loading matrix. Then interpret the dimensions. (4 points)

d. Use the dimensions you described in part c to describe the segments you identified in parts a-b.  (1 point)