11
Decisions Based On Science Optimizat ion Statistic al Modelling Data Mining 1

Decisions Based on Science

Embed Size (px)

DESCRIPTION

Decision science involves several social and behavioral science disciplines. It draws on insight from psychology, neuroscience, economics and management, among others by using statistics to improve decision making.

Citation preview

Page 1: Decisions Based on Science

Decisions Based On Science

Optimization

Statistical Modelling

Data Mining

1

Page 2: Decisions Based on Science

Dataset contains characteristics of the problem,

the rep, and the client.

Page 3: Decisions Based on Science

How might customers’ perception of priceimpact customer loyalty?

First we build a model to understand how customers’ perception of price and overall customer satisfaction predict loyalty.

Customer Loyalty (in years and

months)

Customer Satisfaction

Customer Perception of

Price (somewhat high versus priced fair)

Page 4: Decisions Based on Science

How might customers’ perception of priceimpact customer loyalty?

We use regression procedures to determine how important both perception of price and customer satisfaction predict customer loyalty.

Customer Loyalty (in years and

months)

Customer Satisfaction

Customer Perception of

Price (somewhat high versus priced fair)

Page 5: Decisions Based on Science

We use regression procedures to determine how important both perception of price and client satisfaction predict customer loyalty with the company.

How might customers’ perception of price

impact customer loyalty?

Page 6: Decisions Based on Science

Customer satisfaction explains almost 50% of the client’s loyalty to the company, while customer perceptions of price only about 4.4%.

49%

4.4%

How might customers’ perception of priceimpact customer loyalty?

Customer Satisfaction

Customer Perception of

Price (somewhat high versus priced fair)

Customer Loyalty (in years and

months)

Page 7: Decisions Based on Science

These results suggest that overall, the clients’ satisfaction with the company is the primary engine driving customer loyalty.

Customers seem to be satisfied with the company for reasons beyond just how they perceive the price of the product.

In fact, perceptions of price does not seem to be the major factor in promoting customer loyalty.

But might there be subgroups of customers who show a different pattern?

How might customers’ perception of priceimpact customer loyalty?

Page 8: Decisions Based on Science

Cluster Analysis to Examine Subgroups of Customers

Cluster analysis used to determine whether or not there are distinct subgroups of clients.

The first cluster shows that those who consider price to be very reasonable also show the highest satisfaction and loyalty to the company.

The last cluster shows that there are customers who are very loyal to the company despite also viewing the price of the product as being somewhat high.

Page 9: Decisions Based on Science

Cluster Analysis to Examine Subgroups of Customers

In fact, over 22 thousand (or about 46% of surveyed customers) were highly satisfied and highly loyal despite considering the price of the product somewhat high.

Page 10: Decisions Based on Science

Conclusions and Further Directions

Clearly customer satisfaction is an important component of loyalty to the company.

Two unexpected but useful findings for further exploration.

First customers’ perceptions of price were not nearly as important as customer satisfaction in determining customer loyalty.

Almost half of the survey respondents were highly satisfied even when they considered price to be somewhat high.

Further data mining needed to determine what factors predict customer satisfaction and loyalty.

Page 11: Decisions Based on Science

Business Intelligence

We use advanced tools and technologies to provide analytical power that requires expertise and experience to deliver.

Probability and Statistics – data mining to find valuable connections and insights, test conclusions, and make reliable decisions. Expertise in Regression Analysis, Cluster Analysis,

Multivariate Statistics, and Hierarchical Linear Modeling.

Optimization – understanding which factors best predict the intended outcome by narrowing your choices to the very best when there are virtually innumerable feasible options.

Simulation – giving you the ability to try out approaches and test ideas for improvement. 11