Talk:Multivariate testing in marketing/Archives/2012

Very narrow article

To represent multivariate testing as something particular to internet marketing seems very strange to me. Just because the guy who started the article was familiar with this one application doesn't means that's what it is. Look at some book sources and expand the article. I think our persistent IP vandal is actually trying to contribute something meaningful about design of experiments, but has no idea how to go about writing a wikipedia article section. Maybe someone could look at what he's doing and help him out, instead of just edit warring. Dicklyon 23:22, 27 July 2007 (UTC)

Added IDDEA:

Hi Team,

I had to add another DOE - IDDEA. It is an approach that involves evolutionary algorithms. Patent number: 7016882. Filing date: Apr 26, 2004. Issue date: Mar 21, 2006 Thanks. —Preceding unsigned comment added by Auka (talk • contribs) 13:31, 20 June 2008 (UTC)

Page Tidyup Oct 09

I've begun the process of a page tidyup, the coverage of tagging vs proxy DNS / on premise implementation approaches was very partisan and skewed in favour of the latter approach. I hope you'll agree that the new revision is more worthy of being called encyclopaedic content.

I can see that the design of experiment approach and many other areas also need improvement. I removed the sections that were repetitively re-describing multivariate testing for email and for mobile web, I think it's sufficient just to note that both are possible?

More help appreciated in bringing the quality of this article up to scratch please! AlasdairBailey (talk) 14:12, 20 October 2009 (UTC)

On a small point, I see that article MVT gives different meanings/links for multivariate testing and multi-variable testing but both lead to this article. I don't think it seems sensible to try to include discussion of experiment design in general here. It may be best to rename the article to something like "multivariate testing (web sites)" or "multivariate testing (market research)" and then to concentrate on the one situation. I would limit discussion of experiment design to just a few words to outline appropropriate links. "Multivariate testing" seems not to have a specific meaning in statistics or to be a specific topix, as it doesn't appear in stats dictionaries: of couse it is used some places, see eg Student's t-test#Multivariate testing. Melcombe (talk) 15:13, 20 October 2009 (UTC)

Very narrow article.

The use of multivariate testing in web design is irrelevant. This page should cover its use in the sciences (+ other fields). —Preceding unsigned comment added by 94.213.97.235 (talk) 17:41, 23 July 2010 (UTC)