In our own experience with CTR testing, two weeks seems to be a minimum duration of time to run SEO tests. Although, if you are able to execute testing for a longer duration of time, that is always welcomed as it allows you to gather a larger sample size of SEO metrics to influence your results. For our tests, we decided to run them for a 26-day period. With saying that and SEO best practices in mind, it is important to compare your test CTR to a control CTR of the same time frame. Therefore, we compared our test to the same amount of time (26 days) before we began testing.
SEOTesting.com allows you to setup SEO tests quickly and makes it simple to pick control groups in order to get as close to split testing as possible. Now, if our test in our testing tool comes back and says that we didn’t see any significant improvements, we go back to the drawing board and iterate until we do see significant improvements. Then, once we know what works for our client and target market, we can implement these tactics on a larger list of content pieces.