Final Analysis

-Toy Horse Conjoint Experiment

​

Team 2

Executive Summary

• Key point and findings:
• Four distinct customer segments: 1) Modern beauties 2) Charming Rockers 3) Live fasts 4) Wallet Watchers
• Male and female have significant different preference on attributes
• Both segments prefer 26” toy horses
• Males are more sensitive to price than females
• Females prefer rocking horses and males prefer bouncing horses
• Females prefer glamour style while males prefer racing style
• Dropping, altering, and adding a product will increase market share and compete against current competition effectively.

​

• Recommendations
• Drop 18” Rocking Racing Horse and price at $139.99
• Drop 26” Glamorous Bouncing Horse to save $20,000 in fixed costs
• Add 26” Bouncing Racing Horse and price at $119.99

​

Cluster Analysis- 4 clusters

We find there are four main different customer segments.

Names of four clusters: 1) Modern beauties 2) Charming Rockers 3) Live fasts 4) Wallet Watchers

​

• Price Sensitivity: Consumers in Cluster 4 are more price sensitive than others.
• Height Preference:. Consumers in Cluster 1 prefer 26’’, while other consumers prefer 18’’.
• Motion Preference: Consumers in Cluster 1 and 4 prefer bouncing, while others prefer rocking.
• Style Preference: Consumers in Cluster 1 and 2 prefer Glamour, others prefer racing.

​

• Mean Value is calculated by K-means Clustering Method of the Part Utilities of different attributes.
• Attribute Level: price(1) = $119.99 price(0) = $139.99; height(1) = 26" ;height(0)=18“; motion(1) = Rocking;motion(0)=Bouncing; style(1) = Glamour;style(0) =Racing

Choosing Strategy

• Competitor concerns

Competitor’s Profile:

26’’ Racing Rocking Horse $139.99

​

Utility of the consumers receive by using competitor’s product:

​

​

​

​

​

​

​

We should add or modify products that could generate more utility than the competitors for consumers in each clusters.

​

• Business concerns
• Based on the competition concern, we selected five profiles to each consumer segment that could give higher utility

​

2. Since different combination of height and motion has different variable costs, we would select the ones with lower variable costs.

​

• Finally, we would have to trade off between the utility and the cost. We would use the simulation to test the scenarios and decide the best one with the highest profit.

​

Targeted Profiles

Profile 6, 4, 13 are the best to target for each segment.

• These profiles can give the target customers higher utility than the competitor’s product.
• Even though “Live fasts” will like to have 139.99,18’’ rocking horse, this profile has lower utility than competitor. We should lower the price in order to increase customer’s willingness to buy and get more profit.
• Even though “Wallet Watchers” is price sensitive, customers still will purchase our toys since our toys have higher utility than competitors.

​

•Details of the profiles of each product are in the appendix.

•Revenue = Wholesale Price – Cost; The wholesale prices are $111.99 for the $139.99 retail price and $95.99 for the $119.99 retail price.

A priori segmentation-gender

Run regression to test whether gender have different preference on the attributes. As table shown above, we concluded that all preference for attributes are significant different.(p-value <0.05). “Price” and “Style” are the most significant different attributes.

​

Therefore, we can try to figure out specific profile for Male and Female. Regression conclusion as below:

​

• Male weighs price attribute more than female.
• Both segments have similar preference for 26” inches toy.
• Female slightly prefer rocking horse, while male prefer bouncing horse.
• Female prefer glamour style, while male prefer racing style.

​

​

* Test whether each gender has significant preference on attributes.

* Get the coefficient of attributes for each gender.

Preferred Profiles

Use each gender’s coefficient to calculate each profile’s rating.

Profile choosing strategy:

1. Customers prefer our profile more than the competitor’s. (higher utility)
2. Profile that brings the highest profit. (price-variable cost= profit)

​

​

​

Even though male are more price sensitive, once other attributes match their preference, they would still like to purchase the toy. Similar, female will still buy the profile even though they prefer rocking horse.

​

We should launch profile 5 to Male, profile 6 to Female to gain higher profit and extract as more consumer surplus as possible.

Market Simulations

Based on cluster analysis, we choose profile 4, 6, and 13 for scenarios.

Based on a prior segmentation, we choose profile 5 and 6 for scenarios.

Market Share

* Blue row= current market. Red words= low market share.

• The current highest profit from a scenario is: $292,212.
• One of our products, profile 3, gives us 0% market share. Therefore, we decided to drop this product. Profile 6 only gives 0.0375% market share, so dropping that profile saves $20,000 in fixed costs.

Recommendations

​

• Drop 18” Rocking Racing Horse and price at $139.99 (profile 3)
• Add 26” Bouncing Racing Horse and price at $119.99 (profile 13)
• Drop 26” Glamorous Bouncing Horse to save $20,000 in fixed costs (profile 6)

​

Then we can obtain 93.75% market share and $310012.5 profits.

​

Appendix

Segmenting the Customers into Three Clusters�

​

Consumers’ Preference on Different Attributes:

• Price Sensitivity: Consumers in Cluster 1 is more price sensitive than those in Cluster 2 and 3.
• Height Preference: All the consumers would prefer 26’’ horse, especially those in Cluster 3.
• Motion Preference: Consumers in Cluster 2 would prefer Rocking, while others in Cluster 1 and 3 would prefer Bouncing.
• Style Preference: Consumers in Cluster 2 and 3 would prefer Glamour, especially Cluster 2. But consumers in Cluster 1 would prefer Racing.

​

Notes:

• Mean Value is calculated by K-means Clustering Method of the Part Utilities of different attributes.
• Attribute Level: price(1) = $119.99 price(0) = $139.99; height(1) = 26" height(0)=18“;

motion(1) = Rocking motion(0)=Bouncing; style(1) = Glamour style(0) =Racing

Reasons for Choosing the Targeted Products

Targeted Products to Three Clusters

Profile 5, 4 and 6 are the best to target each segment.

•These profiles can give the target customers higher utility than the competitor’s product.

•Even though “Charming Rockers” will like to have 26’’ horse, but we choose to provide them with 18’’ one. It’s because the cost of 18’’ bouncing will be smaller than 26’’ bouncing.

•Even though “Wallet Watchers” is price sensitive, the utility of other attributes of profile 5 is big enough to justify the higher price.

Notes:

•Mean Value is calculated by K-means Clustering Method of the Part Utilities of different attributes.

•Attribute Level: price(1) = $119.99 price(0) = $139.99; height(1) = 26" height(0)=18“;

motion(1) = Rocking motion(0)=Bouncing; style(1) = Glamour style(0) =Racing

Revenue = Wholesale Price – Cost; The wholesale prices are $111.99 for the $139.99 retail price and $95.99 for the $119.99 retail price.

Reasons of Selecting the Chosen Segment

Based on the K-means Clustering Method, it’s better to have difference between the groups and similarity within the groups. As we can see from CLUSPLOT for three clusters and four clusters, when you group the consumers into four groups, it will be better to distinguish the group 2 and group 3 according to the second component.

Cluster Analysis in R

Regression - test gender preference

Gender preferred profile

* Use the coefficient to calculate each profile’s utility for each gender.

* Red=Competitor profile. Blue= Original profile. Red words=Better rating than competitor’s profile.