Originally Posted by jact

To facilitate proper testing you need to be doing A/B testing during the same timeframe with the same traffic sources. Pulling trailers for a week will tell you something, but will not tell you anything scientifically. Too many variables can change hour to hour, day to day, week to week.