Over the past several months I’ve been attempting to go deeper than simple A-B testing with sites… I’ve been applying Sabermetrics to Affiliate Marketing. For the non-baseball fan, Sabermetrics is the invention of Bill James. His utilization of standard reference observations of the game and logical analysis of real results has turned what everyone thought about baseball on it’s head.
For example, different stadiums have profound effects on an individual ballplayers production. Stealing bases is unproductive. Much of what we credit to pitching is actually defense. His approach is to listen to what the game itself is telling us objectively. Which is exactly the kind of analysis we need to be doing as Affiliate Marketers.
Here are some questions that the application of Sabermetrics can answer for us. With identical websites does the use of one CMS over another affect SERPS? Is there a difference in profitability in using your own dedicated server over shared hosting? Does the selection of a particular offer show greater click-thru by using one network over another network that has the identical offer? What elements (i.e. cart, banner, newsletter list, pop-up on abandonment, etc) are the most to least effective?
We all think we know the answers to these questions. But do we really know? How much is being left to random chance without a true understanding of what is really happening by measuring and analysing the results of more than a simple A-B test? I think that we’re leaving alot of money on the table here. I think we don’t really know what we’re doing.. We just think we know. And I think I’ve discovered the way to find out for sure… Sabermetrics.
By studying Bill James work with baseball statistics, and creating test scenarios, a baseline of information about affiliate marketing can be established. In my own initial testing I took one of the eyeglasses sites created for an experiment last fall and started a series of tests. Although not a large enough sample to be truly scientific there are some trends that I feel need further study.
Just as in baseball, the stadium affects individual player performance, the type of hosting affects an affiliate marketers site performance. The more restrictive a host, the less money is produced from a site. With the best production being on a dedicated server. In other words, if the hosting (and the hosting tools) enable an affiliate marketer to make changes quickly, perform maintenance easily, has enough resources allocated that processor or database limits are never reached… The site will perform better financially.
For every site you have, for every offer you run, you have a wealth of statistical data that you should be analysing. From server logs, to your click report at Shareasale, banner displays from OpenX, commissions received, Google Analytics, and on and on and on. And you should be boning up on Bill James and Sabermetrics so you can use all the details at your disposal and distill the basic truths available to you instead of relying on intuition or whatever some guru says.
But just remember my favorite quote from Bill James, “Information is not to be held accountable for every misleading claim that somebody can derive from it”,