Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
I, Cringely - The Survival of the Nerdiest with Robert X. Cringely
Search I,Cringely:

The Pulpit
The Pulpit

<< [ Ego, Super-ego, and ID Theft ]   |  Taguchi Me This  |   [ Changing the Game ] >>

Weekly Column

Taguchi Me This: How Two Guys From the Gold Country Are Changing Advertising Forever

Status: [CLOSED]
By Robert X. Cringely
bob@cringely.com

Privacy, identity theft, and how information technology makes the former harder and the latter easier has been our theme for the last two weeks. And this is a story that we'll continue, probably next week, as it further unfolds. But now it is time to take a short breather and consider what I learned by selling my Bowflex home gym on eBay. It is a lesson that could be worth billions.

My reason for selling the Bowflex was simple: I am a slug. Six weeks to a fitter body only works, I learned, if you actually use the machine. Muscles are apparently involved. So when I finally faced reality, I decided to sell my Bowflex on eBay. I wrote the ad, pasted a very nice picture of it that I found on the Bowflex web site, and listed the gym only to find the next morning that eBay had rejected my ad and refunded my money. It was that picture, which was copyrighted, and I had infringed Bowflex's copyright, so they very properly demanded the ad be delisted. I have no problem with that, though it might have helped had eBay made the situation a little clearer. They wrote me a very ominous e-mail saying if I relisted the item horrible things would happen to me. But when I called Bowflex they told me to just take my own darned picture, which I did.

A crated Bowflex PowerPro XT (my crates are industrial strength -- no, make that indestructible), by the way, weighs 145 lbs.

How did Bowflex know I was using their picture? A spider program looks at every eBay picture in several categories, and if any of them match an official Bowflex shot, then a challenge is automatically generated and the axe falls. What's amazing about this is that what we are talking about is not a deterrent in the classic sense -- it is actual prevention. Bowflex has the capability of finding EVERY violator, which is why they are satisfied with just stopping us rather than whacking off our hands to discourage others.

This ability to look at every eBay ad and learn something from it intrigued me, and I began to think about what other information could be gleaned the same way. Certainly, there are outfits that keep an eye on final sale prices on eBay and other auction sites so they can tell you what you can expect your item to sell for. But the potential for this historical information goes much, much further. I believe that it could be used to easily double eBay sales.

Here's how. Not everything put up for auction on eBay actually sells. Some items never reach their reserve price, some never get bids at all. So eBay would be a more efficient marketplace if people knew in advance that what they have to sell isn't likely to find a buyer. That could be gleaned easily from historical data since hardly anything for sale on eBay is truly unique. Taking the idea several steps further, we could probably use historical data to learn how to write and place successful eBay ads -- ones that yielded higher prices. Certainly, there are books about this, but I am talking about actually learning in a scientific manner what to say and do to get the best price for your junk. This ability is just now coming available, and I think it has the prospect of revolutionizing not only eBay, but every kind of advertising.

You see advertising and marketing are usually considered to be more art than science. Sure, a fair amount of statistical analysis is used to be sure the person who is reading an ad, watching a commercial or hearing a message is from the target audience, but the message itself is largely a work of art. When marketing and advertising are taught in universities, much of what is taught is anecdotal -- what has seemed to work before. But it doesn't have to be that way, at least not according to James Kowalick and Mario Fantoni, two guys who say they can show you how to use science to design ads that cost less while being 10 or more times as effective as doing it the old way.

Their secret is the Taguchi Method, which is a technique for designing experiments that converge on an ideal product solution. Devised in the late 1940s by Dr. Genichi Taguchi, then an engineer at NTT, the Japanese phone company, the method that bears his name has long been used in designing cars and computers. If you've ever wondered why the quality of Japanese cars is so high, credit Taguchi. The first American car designed using the Taguchi Method was the original Ford Taurus, which quickly became the top-selling car in America. Now just about every car from every manufacturer is designed using the method, which literally builds customer satisfaction into the design.

The Taguchi Method was first brought to the U.S. by researchers at AT&T Bell Labs. There is a global business in teaching Taguchi, but for no reason at all it tends to be restricted to engineering groups in large companies.

James Kowalick was corporate director of engineering at Aerojet General ("Why, yes, I am a rocket scientist") when he met Dr. Taguchi in 1985, and began using the method to design better rocket engines while saving millions of development dollars. Kowalick was so taken with Taguchi that he later left Aerojet and founded the Renaissance Institute at Cal Tech, where he taught more than 300 Taguchi courses to executives from high-tech companies at $13,000 per three-day session.

Taguchi's objective is robust design, which means building a product, system, or process that works well even in the presence of degrading influences. That means products that deliver value without breaking and services that are enduring while being as simple as possible. Taguchi first determines the control factors that go into designing a product, their interdependencies, then generates an orthogonal array specifying the number of experiments required to find the optimal solution.

If the last paragraph reads like Esperanto to you, maybe that explains why mainly eggheads have been attracted to Taguchi. The short version is that however they work, the Taguchi Methods can take a project with thousands, even millions of combinations of variables, and quickly reduce it to a couple dozen simple experiments that can be run simultaneously and will determine the cheapest way to achieve a goal. Instead of considering one variable at a time, Taguchi is able to test many variables at once, which is why the number of tests can be so small. It's a bloody miracle. Taguchi not only shows the right way to do something, it also tells you what the cost in dollars will be of doing it the wrong way.

But until now, Taguchi has been too obscure or too abstruse to make its way out of laboratories and into real products and services from non-high-tech companies. "I taught over 300 courses for industry where we designed cars and electronic devices, but it wasn't until one day I took over my wife's kitchen and used Taguchi to perfect my recipe for vanilla wafer cookies that I realized how broadly it could be applied," Kowalick recalls. "It took 16 batches, but by the end of the afternoon I had those wafers dialed in."

That's when Kowalick turned the Taguchi Method to advertising, with the goal of significantly raising the response rate for ad campaigns. After all, advertising is just a process for sharing information and inducing responses. Like any process, advertising can be optimized if the control variables can be properly defined. Kowalick did a test case with direct mail campaigns for a local winery, then another with Internet advertising for an insurance agency. That was where his longtime friend Mario Fantoni, a former marketing executive for Oracle and management-consultant executive at AT Kearney, entered the picture.

In their original tests, they designed the Taguchi experiments by hand and conducted them using direct e-mail (yes, spam) with the result that the response rate grew by more than 22 times while the cost of advertising to achieve this result dropped by a factor of six. That is a 13,200 percent increase in advertising productivity, which is probably some kind of record.

What it proved was not just that you could sell stuff using spam advertising but that you could optimize advertising through scientific experiments. Remember, this was two weeks and not very much of an investment leading to a massive improvement -- something that simply does not happen in the real world.

Until now.

"The approach condenses the experience of investigating billions of possible combinations for an ad, including copy words, graphics, visual impression sequence, and integration with a larger sales system, into a few dozen mailings or e-mailings," says Fantoni. "It is also a psychological approach that carefully considers what potential readers want to read. Copywriters and artists still have to write the words and draw the pictures that follow the procedures derived from the experiments, but once they become accustomed to actually knowing why what they do works, it becomes a way of life."

This process is very different from the kind of testing that is presently done by major advertisers. A big advertiser may come up with several campaigns it tests in separate markets, but the approach tends to be shotgun, with little rigor in determining how one campaign is different from another and what influences are actually being measured. So Taguchi is potentially a great leveler, making it possible for smaller companies to advertise just as effectively as their larger competitors.

Now apply this idea to eBay advertising. The beauty here is that the experiments probably don't have to be conducted in the classic sense of inserting ads and comparing responses. There are enough items, enough ads, and enough buyers on eBay to probably run the Taguchi experiments against historical data. If automated, this could be done in minutes from start to finish. Clearly this is a service eBay and the other auction sites ought to offer for a little extra money, say a buck per ad.

Of course, Kowalick and Fantoni have started a company to exploit this technology. No silly, I have no financial interest in their company, though they once did buy me lunch. Their company, which is based in Oregon House, California, is called MR2, for "Maximized Response Rate." You can find a link to their web site under the "I Like It!" button on this page. THEY CLAIM THEIR WORK CAN BE APPLIED TO ANY PRODUCT OR SERVICE AND ANY ADVERTISING MEDIUM. And what presently requires sitting with Kowalick and Fantoni (at a cost of about $8,800) will soon be reduced to a $499 interactive program (written in Flash of all things) that will run on a PC, bringing all the benefits of Taguchi without requiring that a nerd be enclosed to make it work. The software is in beta test right now.

Comments from the Tribe

Status: [CLOSED] read all comments (0)