05 November 2006

Keep on Testing, Igor! (repost)

Regular readers may be bored by my continuous emphasis on testing, but if you don't know whether the latest piece of finely honed copy is better or worse than the one that went before, then what drives you to change it again other than the calendar and some organic reaction deep in your gut?

I've been testing an AdWord on our champagne site against our base ad. The test has been running for about 5 weeks now which is quite a long time. Why so long? The answer is that the data is confusing and it depends which timebase you look at in deciding which is the better ad.

Over the whole of the month of October, the new ad underperformed the base ad by about 8%. I was a little surprised by that as I thought that the new ad was structurally better and should have performed better. I decided to keep it on.

In the first 4 days of November the new ad has outperformed the base ad by more than 100% and overall it is now a better ad than our base ad, despite underperforming for the whole of October. Is this a blip or a real indication of the relative merits of the 2 ads? I don't know.

If anyone has any suggestions then I would be happy to entertain them.

Labels: , ,


Post a Comment

Links to this post:

Create a Link

<< Home