Genichi Taguchi began his design of experiments method in the 1950s, before there were computers. Born in Japan, Taguchi studied statistics and engineering. He founded the methods originally to improve manufacturing methods. Today his methods are popular in biotechnology as well as in marketing and advertising.

There are three basic factors of the Taguchi Method. The first one is system design, which is the creative and innovative process. The second is parameter design, which involves robustification. The third is tolerance design, which focuses on control and reduction in variation of critical dimensions.

Statisticians have never wholly embraced Taguchi methods. However, these methods are very helpful in the Internet-marketing field today to test and measure various components of web pages and Internet campaigns. Manufacturers still apply the methods as well. Consider the following when deciding how to work with Taguchi Methods:

1. Use Taguchi Methods in manufacturing.

2. Find DOE companies for Internet marketing.

3. Attend workshops and get training for Taguchi Design of Experiments.

Action Steps
The best contacts and resources to help you get it done

Use the Taguchi Method to improve your manufacturing bottom line

Taguchi Experimental Design and DOE use testing and measurement processes and methods to scientifically compare and contrast various ways to get the best results and outcome for a product or process.

I recommend: ReliaSoft analyzes experimental data in detail from metrics, ratios, reliability factors, interactions, comparisons, response and error that may affect the process, service or product. Integral Concepts works with mixtures and formulas. They help in design and modeling of complex manufacturing techniques.

Find an Internet marketing Taguchi DOE company

Making the most of Taguchi Methods and Design of Experiments (DOE) is very popular in e-business. Whether a company wants to compare headlines in copywriting or pay-per-click versus organic traffic methods, Taguchi can help pinpoint what works and doesn't work.

I recommend: TaguchiNow uses multivariate and Taguchi Design of Experiment Methods in marketing and advertising. Clients range in size from small businesses to corporate entities. SiteSpect offers turnkey Taguchi solutions with metric and web analytic applications for websites.

Train your employees in Taguchi Design of Experiment

Send your employees to seminars and workshops to learn basic and advanced skills in Taguchi methods. Training is easy and convenient to find.

I recommend: ASI puts on workshops and trains in DOE for project managements and lean manufacturing. Learn how to design DOE experiments with Launsby. Look at the list of upcoming classes and seminars at their site.

Tips & Tactics

Helpful advice for making the most of this Guide

• •  Get some basic training in Taguchi Methods and Design of Experiments before spending a lot of money. Learn enough so you can hire the right professionals to help you measure, test and improve your bottom line. The method depends on alternatives, testing and choices. Look at your product or service ahead of time to decide what you think your choices are.

Taguchi Methods and Design of Experiments (DOE) are an oft-utilized and sometimes controversial way of looking at how to improve your company's profit margin. To put it simply, Taguchi methodology focuses on applying engineering techniques to your company's processes rather than more-traditional statistics. By focusing on adjusting your processes' inputs as far up-stream as possible, you're able to significantly improve your company's quality engineering.

While people can spend their entire professional careers studying Taguchi DOE and how to apply it, most business owners don't have the time to do so. Instead, they rely on Design of Experiments software (DOE) as well as consultants who provide them with an appropriate Taguchi Design of Experiments solution for their business. There are a wide variety of such options available in the market today; the only difficulty can be in choosing the right one.

When looking at utilizing outside help for Taguchi experimental design in your business, among the key points to consider are:

1. The ability of your company to apply findings from Design of Experiments software;

2. The possibility of training one or more of your employees in the Taguchi method;

3. The opportunity to bring in an outside Design of Experiments method consultant.

Action Steps
The best contacts and resources to help you get it done

Find Taguchi experiments software that fits your business

If you or one of your employees already has rudimentary knowledge in Taguchi Design of Experiments, it's possible that all you will need is the right software to turn that knowledge into higher profits. Whether you prefer a software module that works with Microsoft Excel or a standalone program, there are options available for your company.

I recommend: Minitab is a standalone piece of software that offers a variety of statistical analysis options, including Taguchi DOE. KnowWare offers a series of QI Macros that integrate with Microsoft Excel, including some for the Taguchi Method.

Train yourself or your employees in Taguchi Methods and Design of Experiments

If you truly want the best application of Taguchi methodology on a day-in, day-out basis, you should seek expert training for yourself or some of your employees. By learning as much about the Design of Experiments method as possible, you're more likely to correctly apply your knowledge continuously in your business.

I recommend: Integral Concepts offers both Design of Experiments and Advanced Design of Experiments training courses to the public.

Hire a consultant to implement a Taguchi Design of Experiments solution

If you don't have the knowledge to appropriately apply software or the time to become trained on the Taguchi method, your best solution would be to hire a consultant with the knowledge you need. By utilizing an outside consultant, you can help increase your profit margin through quality engineering.

I recommend: Stat-Ease has several Design of Experiments experts who can consult with your company on your quality engineering needs.

Tips & Tactics

Helpful advice for making the most of this Guide

• •  While some statisticians discount Taguchi Methods and Design of Experiments, it's considered strong enough to be a critical part of the renowned Six Sigma certification. By utilizing the methods appropriately, you can help your company to achieve this prestigious level.

Genichi Taguchi used statistics and engineering in his design of experiments methods in the 1950s for manufacturing processes. Today Taguchi methods are in Internet marketing, advertising, biotechnology and manufacturing. These innovative methods work by methods of testing and measurement. For example, you can try alternative designs or parts in a manufactured product to see which one customers like the best.

Traditional Taguchi methods and design of experiments (DOE) basics today fit other industries besides manufacturing. Taguchi is proving popular in Internet marketing and advertising. Direct mail, for instance, likes to test copy to see what works the best whether it is an alternate headline, design of packaging, or sales-letter style. When a lot of different options and choices are available, math and statistics of Taguchi methods bring a level of accuracy and clarity to the process. Consider the following for Taguchi basics.

1. Look at Taguchi methods in industry and the marketplace.

2. Read online articles on Taguchi methods.

3. Attend a workshop in the design of experiments method.

Action Steps
The best contacts and resources to help you get it done

Look at Taguchi methods in industry and in the marketplace

Taguchi methods apply to a broad spectrum of industry and markets from manufacturing to Internet marketing and advertising.

I recommend: Verster applies the DOE method to improve website rankings with search engines. TaguchiNow applies this multivariate testing method to advertising and marketing—and the Internet. KnowWare International uses the Taguchi methodology in manufacturing processes and services. They work at all levels of corporations with Taguchi and six sigma methods.

Read helpful online articles on the Taguchi method

These websites and articles give basic concepts and ideas on DOE that business owners may find helpful in improving processes and methods of testing and measurement.

I recommend: This article at iSixSigma reveals little known facts about Taguchi. Optimost gives the history and background of the method.

Get trained in Taguchi design of experiments to better understand the process

Training is the first step in the application of a new method to your business process. Go to a workshop, seminar or take a class in DOE methods. At first, the whole idea can seem confusing and mysterious--after all, it's about statistics and engineering. There are beginning (and advanced) courses available, however, that can help you get a handle on the subject.

I recommend: Stat-Ease trains in the use of DOE for design in six-sigma improvement. Attend one of their workshops and learn the method. Training-Classes.com also offers workshops in Taguchi.

Tips & Tactics

Helpful advice for making the most of this Guide

• •  Train yourself in the basics of Taguchi DOE before spending a lot of money on the process. This way you will be better equipped to hire the right professionals to help you measure, test and improve your bottom line.

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 them 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.

Action Steps
The best contacts and resources to help you get it done

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.

I recommend: FreeQuality.org offers an introduction to the orthogonal array selector and provides a link to it. 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.

I recommend: iSixSigma explains the steps in robust parameter design.

Robustification

Robustification is used to adjust the design variables. Doing this 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.

I recommend: AbsoluteAstronomy.com defines robustification and explains its principles.

Optimization

Optimization means to make 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.

I recommend: 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 all 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.

I recommend: 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.

I recommend: John M. Browne explains probabilistic design, and his website has links to information on the process as well as research based on it. Clint Steele discusses putting these techniques into practice.