This content is not included in
your SAE MOBILUS subscription, or you are not logged in.
Creating a Two Sided Customer Loss Function
Technical Paper
2015-01-1357
ISSN: 0148-7191, e-ISSN: 2688-3627
Annotation ability available
Sector:
Language:
English
Abstract
In the area of Human Factors and Usability research a desired output of many studies is identification of what value a specific Design Parameter should be set at to minimize customer dissatisfaction.
A Customer Loss Function is a simple way to graphically display the probability customers will be dissatisfied at different levels of a given design parameter, due to a given failure mode. Many design parameters however, have two distinct but related Failure Modes (customer disatisfiers), typically representing two ends of the parameter (i.e. too much/too little; too hot/too cold; too fast/too slow). Each of these Failure modes is represented by its own unique Customer Loss Function. This paper will introduce a technique to combine these two One-Sided Loss Functions into a comprehensive Two Sided Loss Function. The mathematics behind the creation of both one sided and two sided loss functions is based on Binary Logistic Regression [1,2,3] Analysis Techniques.
The benefits of incorporating both failure modes into one two-sided customer loss function include:
- 1)Being able to graphically display the Combined Customer Loss Function curve and utilize this visual aid to identify the level which minimizes customer dissatisfaction.
- 2)A secondary benefit is the ability to use the two-sided loss functions equation to develop an optimization tool for mathematically determining the minimum customer loss point. This is particularly useful if running a limit study (as opposed to a study with set levels).
Authors
Citation
Crowley, J., "Creating a Two Sided Customer Loss Function," SAE Technical Paper 2015-01-1357, 2015, https://doi.org/10.4271/2015-01-1357.Also In
References
- Hilbe , J. M. 2009 Logistic Regression Models Boca Raton, FL CRC Press
- Menard S. W. 2010 Logistic Regression: From Introductory to advanced concepts and applications Los Angeles SAGE
- Dodson , B. , Hammett , P. C. & Klerx , R. 2014 Probabilistic Design for Optimization and Robustness for Engineers. West Sussex, United Kingdom John Wiley & Sons, Ltd.