Taguchi Methods and Design of Experiments (DOE) Key Terms

Understand the concepts used to design experiments in the Taguchi method

Originally developed to improve engineering productivity, companies are now using the Taguchi Method in direct marketing mail and web applications. By testing a variety of factors in exact combinations, both engineers and marketers can determine the positive or negative effect of each element. This allows engineers to choose the most effective selections to optimize the design. Although some have embraced this method, it is somewhat controversial. Supporters, however, say that it is easy to use, even for those without statistical training.

 

Orthogonal array selector

An orthogonal array selector is used to figure out how many trials will be necessary when conducting experiments. It also helps determine the factor levels for each parameter.
Try: George Sherwood provides in-depth information on constructing orthogonal arrays.

Parameter design

A parameter is a factor that might influence the final product. The main steps in parameter design are similar to those in any scientific experiment. First, the problem must be defined. Then an experiment is set up using controls to determine the factors that are influential in that problem. After gathering data, the best results are selected, and then additional experiments are conducted to be sure the optimum design has been chosen.
Try: iSixSigma explains the steps in robust parameter design.

Robustification

Robustification is used to adjust the design variables. Robustification makes the overall system less sensitive to noise or random variability. When done correctly, it reduces cost by allowing the equipment or operation to function more effectively.
Try: AbsoluteAstronomy.com defines robustification and explains its principles.

Optimization

Optimization refers to the process of making a product the best it can be. In engineering, it is the goal of design. In online marketing, it is often used to judge the landing page, or first page a customer sees. Because the purpose of online marketing is to attract and keep consumers, landing page optimization means making the page simple, appealing, understandable, easy to navigate and focused.
Try: Omniture provides an explanation of the purpose and benefits of optimization. Chase Online Marketing Strategies lists tactics for optimizing landing pages.

Multivariate testing and A/B testing

Multivariate testing, or multivariate analysis, is used to test multiple changes at the same time, whereas traditional A/B testing compares only one factor at a time or a limited number of factors. Many experts believe using multivariate testing can increase efficiency and profitability.
Try: MarketingExperiments compares multivariate with A/B testing. Incisive Interactive Marketing discusses ways to improve A/B testing.

Probabilistic design

Probabilistic design uses mathematical calculations to determine the probability distributions of each variable. Designers can use these to make more reliable products, which makes them safer, more economical, and of higher quality.
Try: John M. Browne explains probabilistic design. Clint Steele discusses putting these techniques into practice.

Copyright © 2011 Business.com, Inc. All Rights Reserved.

---