Closing thoughts There are many different
Posted: Wed Feb 12, 2025 5:15 am
RLDs
The raw root linking domain count performed significantly better than the share and control at ~20.5%. As I pointed out earlier, this type of analysis is incredibly subtle because it only detects when an element is leading and Moz Link Explorer discovered the relevant element before Google . Nevertheless, this result was statistically significant with a P value <0.0001 and a 95% confidence interval that RLDs would predict future ranking changes that were 1.5% higher than random.
Page Authority
By far, the best-performing factor was Page Authority. At 21.5%, PA correctly predicted changes in the SERPs 2.6% better than random. This is a strong indication of a significant factor, outperforming uruguay number data shares by a significant margin and outperforming the best-predicting raw metric, root linking domains. This is not surprising. Page Authority is designed to predict rankings, so we should expect it to perform better than raw metrics at identifying when a ranking change might occur. Now, this doesn’t mean that Google uses Moz Page Authority to rank sites, but rather that Moz Page Authority is a relatively good approximation of the link metrics that Google is using to determine which sites to rank.
A experimental designs that we can use to help improve our research industry, and this is just one of the ways that can help us bridge the gap between useful ranking factors and lagged correlations. Experimental designs don’t have to be elaborate and statistics don’t have to be sophisticated to determine reliability. While machine learning holds great promise for improving our predictive models, simple statistics can do the trick when we’re establishing the fundamentals.
The raw root linking domain count performed significantly better than the share and control at ~20.5%. As I pointed out earlier, this type of analysis is incredibly subtle because it only detects when an element is leading and Moz Link Explorer discovered the relevant element before Google . Nevertheless, this result was statistically significant with a P value <0.0001 and a 95% confidence interval that RLDs would predict future ranking changes that were 1.5% higher than random.
Page Authority
By far, the best-performing factor was Page Authority. At 21.5%, PA correctly predicted changes in the SERPs 2.6% better than random. This is a strong indication of a significant factor, outperforming uruguay number data shares by a significant margin and outperforming the best-predicting raw metric, root linking domains. This is not surprising. Page Authority is designed to predict rankings, so we should expect it to perform better than raw metrics at identifying when a ranking change might occur. Now, this doesn’t mean that Google uses Moz Page Authority to rank sites, but rather that Moz Page Authority is a relatively good approximation of the link metrics that Google is using to determine which sites to rank.
A experimental designs that we can use to help improve our research industry, and this is just one of the ways that can help us bridge the gap between useful ranking factors and lagged correlations. Experimental designs don’t have to be elaborate and statistics don’t have to be sophisticated to determine reliability. While machine learning holds great promise for improving our predictive models, simple statistics can do the trick when we’re establishing the fundamentals.