Building Better Spam

henbane writes "Cringely is plugging a new method of advertising from Dr. Jim Kowalick and Mario Fantoni. Their book entitled 'E-Mailing Your Way to Sales With the Taguchi Approach' is out in the autumn. What could be worse than a method which increases the returns on spam?"

What does this have to do with spam?

[NineNine]

Did anybody RTFA? What does this have to do with spam? This is a originally a way of improving processes, primarily in engineering and/or manufacturing. Now, it's been applied to marketing. Since when is all spam considered marketing? I give this article a -1, Troll.

Whoa

[HanClinto]

From the article:

If you've ever wondered why the quality of Japanese cars is so high, credit Taguchi.

Okay, has anyone heard of a guy named W. Edwards Deming [deming.org]?

To paraphrase Tommy Boy, "I could take a dump in a box and mark it guaranteed". I.E. *marketing* does not make quality. I never heard of this Taguchi guy.

not spam, just the Toguchi method

[Anonymous Coward]

Cringley isnt' advocated better spam, but the Toguchi method in general. As an example of its effectiveness, he states that the two "e-mail marketers" improved effectiveness by 13,600% (that's 136x).

Taguchi Method

[robbyjo]

For those of you interested on learning Taguchi method. Here's a good intro [isixsigma.com].

Re:What does this have to do with spam?

[jacksonyee]

I definitely second your comment there. This process looks like it's built for efficiency and a good design process, and could be applied to almost any field of engineering or problem solving.

Marketing here in the U.S. has definitely become an evil, throat-cutting industry, but if someone can prevail with simple, effective ads that get to the point and don't leave consumers confused and disgusted, I think that would be a win for us all.

Of course, as human nature would have it, the disgusting, false ads will probably be the ones who generate the most revenue... most consumers don't bother taking the time to research, and thus they are left as sheep to the slaughter of their money.

Re:Whoa

[pavon]

Taguchi did not develop this method for advertizing. He developed for engineering, and it was never used for advertizing till extremely recently. It is a general method for determining a handfull of tests which can be used to find the empericall tradeoffs between a whole slew of various factors. Taguchi used it to improve the quality of japanese cars while keeping costs down.

Giving him full credit for the quality of japanese cars, as Cringly did, was definately an overstatement though.

Re:He doesn't plug spamming

[SarekOfVulcan]

This sounds like it shares some basic principles with Pairwise Testing. Basically, the theory here is that a large percentage of bugs come in through the combination of two inputs. So, if there are 10 bits in a flag, you need to only make sure that each pair shows up once: with various constraints, I got down from 1024 to 12 values to check on a recent test.

Check out http://www.developsense.com/testing/PairwiseTestin g.html [developsense.com] for a better explanation, or anything else Google brings up for you.

Re:Whoa

[Anonymous Coward]

My understanding of Deming's approach was to ignore the "outliers" or noise - and focus on the majority of results in stream-lining processes... theory essentially being to focus on low-hanging fruit...

Taguchi appears to be focussed on designing systems with greater tolerance for noise without increasing cost disproportionately...

Core difference appears to be Deming's assumption that statistical outliers are costly to capture - whereas Taguchi encourages efficient and focussed experimentation to determine which noise (outliers) can be accommodated economically...

A former employer was big on Deming (and another on Kaizen) - elements of both are valid, but Taguchi seems to be a value add over and above... (from my PHB perspective)...

Re:What could be worse...

[azav]

If you read the article, this is not about spam. It is about using an engineering methodology to identify and reduce variables that affect the successful outcome of a product. This approach is then being turned to advertising and marketing. If anything, it will produce a less annoying advertising message. Something I'm sure we all would not mind seeing.

The title of the Slashdot article is misleading.

Microsoft Accelerator for Six Sigma

[cpeterso]

Microsoft has a new product called the "Microsoft Accelerator for Six Sigma" [microsoft.com]:

Microsoft Accelerator for Six Sigma is an integrated set of products and services customized for Six Sigma practitioners. The accelerator can help Six Sigma project teams more effectively manage a large number of projects, more easily track their financial impact, optimize and track resources, and electronically share knowledge gathered across the enterprise.

Orthogonal == perpendicular

[yerricde]

"Vectors" are a construct from linear algebra; if you're not familiar, please review an introductory text on linear algebra. "Orthogonal vectors" are vectors that are perpendicular in vector space. You can test whether two nonzero vectors in an inner product space are orthogonal by taking their dot product. (The dot product of two vectors is a scalar; to compute it, multiply them element-wise and then sum all the products.) If the sum is close to zero, the vectors are close to orthogonal.

You can learn more on Wikipedia [wikipedia.org].