Moz has released 2013 Search Engine Ranking Factors, a survey supplemented with real-world correlation data from scientific examination of more than 17,000 keywords search results. The data has been examined by Dr. Matt Peters and his data science team. Moz surveys more than 100 top industry professionals every two years to compile its biennial Search Engine Ranking Factors.
Some of the 2013 data was previously released by Moz but the complete set is now presented to everyone. Here's an in-depth data analysis of the 2013 Search Engine Ranking Factors.
Why it is Called Ranking Factors?
Google claims that it uses 200 signals in search algorithm. While nobody is aware of what these signals exactly are, Moz examines the high-ranking pages to understand the characteristics of pages that search engines like to reward. This provides a hint of the ranking factors being used by the search engines.
In case the correlation data shows that high-ranking pages are somewhere associated with large number of external backlinks, Moz makes a guess that backlinks are still important for Google.
However, it is not always true that correlation data will point towards a direct ranking factor (causation). It might point towards the characteristics of high ranking pages.
Image Courtesy: Rank Fishkin
Comparing the Correlation Metrics
The 2013 Ranking Factors contain an interactive chart that will let you compare different categories of correlations with each other. In the image below, page-level link metrics (in yellow) is compared to domain-level link metrics (blue).
On comparing the data, Moz team found out that high-ranking URLs are highly correlated with page-level link metrics instead of general domain-level link metrics.
In this year's study, Page Authority is the most highly correlated SEO metrics at 0.39.
On-page Keyword Usage vs. Page-level Anchor Text
The correlation with page-level anchor text remains strong, whereas, the correlation with on-page keyword usage has declined. According to the 2013 report, the number of root domains with partial match anchor text linking to the page has 0.29 correlation.
Social Correlations vs. Link Metric Correlations
The social metric correlations are almost equivalent to link metric correlations. Therefore, Moz advices not to jump into the conclusion with correlation data. It suspects that Google might use social data in search algorithm. However, it is not true that pages with higher number of social shares have higher number of links.
Moz invited 128 industry experts to discuss on the successful tactics according to them, which work for search engines and the future of search. Here's the chart of what experts think:
To see the complete 2013 Search Engine Ranking Factors, click here.