Google Analytics Content Experiments don't deliver 50/50?

[VistageSEO]

Our A/B test is actually delivering at about a 70/30 page view rate. 70% in favor of the original version and only 30% of the new. We are sending 100% of our traffic to this homepage test.

Has anyone else experienced this? There seems to be a lot of folks experiencing this.....anyone know why?

[Mark_Jay_Apsey_Jr.]

This happens all the time. We have had 90/10 page view rates with just the addition of a call to action or new graphics or a simpler page. If you really want to define what it is you should do some multivariate testing within the A/B testing situation. So for example, if you added a call to action and a new graphic design, then just add the call to action as A1 and then call to action and graphic as A2 and then call to action and a different graphic as A3. Then a different call to action and same graphic as version 1 as A4.  Then you can start to hone on specifically what is casuing the differences.

Also, please make sure you have enough traffic to be able to get solid data. Sometimes if you are driving a small amount of traffic, you might get spikes of difference in a short period of time, that will level out with more traffic....we like to see a minimum of 1,000 clicks before we start looking at the results.

Hope this helps. If so, please let us know.

Mark