View Single Post
Old 08-19-2006, 02:08 AM  
Randomonium
Confirmed User
 
Join Date: Aug 2004
Posts: 152
Quote:
Originally Posted by Paul Markham
Next week I will make the tour and updates very much about video. showing the top videos, updating more videoa.

Let's see what difference it make to conversions over a week.
Instead of switching it in for a week, consider running a test segment or segments.

Comparing at different times is not comparing apples to apples.

Everything from end of the month, to season, to varying amounts of traffic per source can affect conversions and not be related to the performance of the tour you are trying.

In order to allow for that if you are just going to be switching them out and comparing them you need to allow a big sample size so that fluctuations normalize.

By running test segments concurrently you are able to get better information from a smaller sample size.

It's the same idea that I described here
http://www.gofuckyourself.com/showth...8#post10654318

just applied to a paysite with conversions being how you track performance.
Randomonium is offline   Share thread on Digg Share thread on Twitter Share thread on Reddit Share thread on Facebook Reply With Quote