Anyone use Google Website Optimizer?

[Pulse]

I realize a lot of you don't like to give Google more information that you must, but their website optimizer is pretty nice and I'm tempted to try it out. Anyone have any experiences with it? Reasons not to use it? Are there any alternatives out there?

 

[rgordon83]

I use it a lot. Best free tool out there.

Plus i don;t do anything shady so google can have whatever data they want, they'll get it one way or another anyway....

 

[Enigmabomb]

I'd split test my dick in any of those racks.

 

[blackrizo]

i find that asian chick pretty hot

 

[HairyHun]

i find that asian chick pretty hot

Great minds think alike.

 

[Bofu2U]

Dibs. Anyone that knows me knows this is the definition of my kryptonite.

 

[envision]

I realize a lot of you don't like to give Google more information that you must, but their website optimizer is pretty nice and I'm tempted to try it out. Anyone have any experiences with it? Reasons not to use it? Are there any alternatives out there?

A limitation for affiliates is that it's more designed for advertisers. If you can't place the GWO conversion code/pixel on your advertiser's thank-you page, you may define a LP click-through as a conversion. So you optimize for click-throughs. It could be that some version of your LP gets high click-throughs but low conversions and you end up picking the wrong design.

But you can track conversions with P202 via subID reports. Does anyone know a way to integrate P202 with GWO? So that I could see subID x had layout version such-and-such?

 

[nator]

Main limitation of the google website optimizer is that it uses simple multivariate testing, which means that each combination gets equally exposed, and if you have enough variations you'll need shitload of sample data before you can draw conclusions.

There're tools out there that use the taguchi method, which somehow can isolate each parameter in your tests, so that you don't need to wait too long to conclude the test.

 

[threehundred]

You can draw conclusions from GWO before its through gathering data. It will show you which "parameter" is doing better as well as which combo is doing better.

Use GWO multivariate at first to get a good idea of where to start with an LP, then set up that initial LP in p202 so that you can see if conversion rate is good, test against a second LP that has the multivariate test running. Set that page up in p202.

Now you have 2 pages in p202 so that you can compare conversion rates and page CTR. Use GWO on the second page so that you can always be testing something. If your GWO testing page is not beating your earlier page, change it up. If it is beating your other page, pick the winner, put it as static page in p202 and test against another GWO multivariate page.

I hope that makes sense. I didn't read over it before posting.