The final phase of SEO testing is measuring the results. Despite being the most important step, it is the biggest challenge; 51% of SEOs don’t know how to decide on a clear outcome for their SEO experiments.
Pre/Post Analysis: If you are doing manual SEO testing, you are likely to use the pre/post model. While it is definitely better than nothing and helps you validate your SEO decisions, it has its problems. First: the error band is much higher because you don’t have a control group. There is also a higher chance of having a false positive/negative, which would lead you to base your SEO strategy on inaccurate data.
CausalImpact – This is what SplitSignal uses. It compares the actual the advantage of afghanistan phone numbers in business traffic in the Control group to the traffic in the Variant group. If the difference is large enough, the test is statistically significant.
img-semblog
Perform an SEO statistical test with SplitSignal
Changes to Google’s ranking algorithm are inevitable, and there is no secret recipe for increasing organic traffic.
This is why experimentation is a necessary element of SEO – by running experiments before making a site-wide change, you avoid a potentially damaging drop in traffic/clicks, while also allowing you to test more variations.
Failed tests are fine, too. In fact, they are at the core of the scientific method. A rejected hypothesis is not a failure; it is a conclusion, something we can learn from and use to build future iterations. However, without an easy way to rationalize this experimentation, a lot of time and resources can be wasted.