Welcome to WebmasterWorld Guest from

Forum Moderators: open

Message Too Old, No Replies

Multi-variate testing

Can it be done without messing up SEs?



4:40 pm on May 3, 2007 (gmt 0)

10+ Year Member

Hi, been awhile since I've been on WebmasterWorld :o)

I'm very interested in 2 different brands of software that allow split and multi-variate testing. I like what I've seen but am unsure of the SE aspect?

I've no intention of fooling the SEs with keyword stuffing or anything like that, I just want to optimise the site to get the best possible response from visitors.

The SW I'm looking at can change elements of a static html page, different font, different headline etc.

Ironically what I like about it is the fact Adsense is not affordable due to a current poor conversion rate compare to the price of the product and the price of clicks ($30 product, 50-60c clicks, without an extremely high conversion rate it's not worth throwing money at Google).

So improving my conversion rate is my No1 goal - but as my traffic is organic and only around 30 uniques a day, I daren't mess with what little traffic I get.

Multi-variate would be perfect for me but what happens with SEs?

Will it confuse the heck out of them, will lack of cookies means they'll always see the control page or what?

Would it be penalised?




4:41 pm on May 3, 2007 (gmt 0)

10+ Year Member

And I forgot to mention, please don't remind me that Google offer a free version, for various reasons I am totally disinterested in using it, thanks!



4:52 pm on May 4, 2007 (gmt 0)

10+ Year Member

A popular topic then?




8:15 pm on May 4, 2007 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member

Although I've never done a test myself to prove whether A/B testing is frowned on by the SE's, I've never heard any anecdotal evidence that it is.

This topic has come up here recently [webmasterworld.com], and didn't get many replies then either. However, following a link in that thread to this one [webmasterworld.com] will lead you to a short thread on the topic which advises people to exclude bots from the B version.

This technique would seem to require a redirect to a B page that would be off limits via robots.txt. Strictly speaking, that's not really cloaking. It might be safer, though.


10:11 pm on May 4, 2007 (gmt 0)

10+ Year Member

Thanks for that, though not sure how to do them re-direct or robot things.

I know how to upload a r' text file, just not sure how to write it or what to put.

Wrote to the maker of the software and got the reply that yes it could be a problem but will be fixed in the next version, which is more vague than I'm willing to play with right now.

The site has 2 main sales-letter pages (2 different products), about 12 articles, and the usual collection of FAQ, privacy, contact etc. Of those the sales pages and 2 of the articles are generating virtually all of the traffic. I know the page I'd most like to tweak but if I direct SEs away from it I'd lose roughly 15 of my 30 to 50 visitors...


One might hope that if google have introduced such testing they'd give guidelines how use it without screwing the serps...



Featured Threads

Hot Threads This Week

Hot Threads This Month