[marvinmacs]

I just came from this thread, which I find absolutely useful:
How long should my marketing funnel be?

Using that thread as my context, I'd like to seek help regarding split tests, per se (a search on "split test" in this forum didn't lead me to a thread specifically on this topic, or maybe I didn't look hard enough). I understand what a split test is. But I'm able to implement this, so far, in aweber -- specifically to check which email catcher design works best.

May I know how you do split tests? Do you use a third party program to do this? If so, what are they? Are there cost-effective ways of doing split tests that won't harm the budget of a start-up?

Thank you in advance for your help.

[MikeVillar]

I won't go into specifics because this service is VERY straightforward (to the point that it almost holds your hand): Google Website Optimizer http://www.google.com/websiteoptimizer

If you're looking for more advanced, enterprise features: Omniture's test and target Website Optimization & Tracking | Multivariate A/B Testing | Website Testing | Adobe Powered by Omniture

Kamusta pre?

[_jc]

I have a problem with Google Website Optimizer - and a sort of work-around that I use.

The problem is that, like any good split tester, GWO doesn't give results until you have enough sales (or other visitor actions) to give a high statistical significance. But when a site is new, or otherwise has low traffic, such A/B split tests can take months to complete.

My work-around in these cases is to assign a "proxy sale" action item, instead of a sale. That is, some visitor action, which only visitors with a lot of motivation or interest will take, but something short of a sale - something that can happen much more often than a sale would.

Of course this compromises the meaning of your result somewhat - but still provides a good learning opportunity about site improvements.

Hope this helps...
_jim coe