As a certified control freak, I loves me some split tests. So clear-cut, so reliable! But the world is the world, and sometimes you find yourself with questions that can’t be answered with a split test. Shocking, I know! But it happens.
Say, for example, that you want to figure out whether it makes sense to bid on your brand terms in PPC when they already rank well in SEO. This isn’t an easy thing to test with a split. You could control your ads so that they only show to 50% of search traffic, but it will be hard to parse out which SEO traffic wasn’t served against a search ad.
What is a marketing nerd to do?
The same principle applies to this test design as to your split tests– control what you can to minimize outside factors. For something like the PPC question, you might run your ads every other day so you know which days didn’t have ads running and can split out those SEO results.
(Some of these slides truly had about 15 seconds of material behind them… this was one. We are done for today, class.)
Next up: Why marketing is not like rotisserie chicken.
(Photo credit: pixyliao via Flickr)