SEO Correlation Study
The good folks at Searchmetrics have put together a very comprehensive study based on a very large collection of data showing correlations between search engine rankings and a large number of factors. What this basically comes down to is, if you’re interested in your website showing up high in Google search results, here are the things you need to do to have the best chance. If you’re at all interested in SEO, I encourage you to download the report, though I understand that 80+ pages is a lot to read. If you don’t have the time to read 80+ pages of SEO related data, here are some things that are good to know.
Correlation vs. Causation
First, correlation does not mean causation. This came up last year when Moz released their ranking correlation study. There are elements like social media which have a high correlation with search results. However, Google has been clear that they do not crawl social media data (though it would stand to reason that they value Google+ data highly…). In this case, the opposite is likely true. Because pages rank so well in search results, they naturally have more people that visit their page and subsequently share on social media. it also stands to reason that pages shared more on social media will also accumulate more links due to a higher readership.
New this year is the study of bounce rate and time on site. This data, fed to Google straight through Google Analytics, has also been explicitly said not to be used by Google. However, they are highly correlated. Once again, content that people are reading and engaging with on a website is likely to be shared more, have more links, and relevant to the Google search terms. This isn’t necessarily causation, but there is high correlation.
Important Ranking Factors
Content is very important. The updated post from Points Group shows not only the effect of WordPress on SEO, but also new content. Content has been important and continues to be. Some content guidelines that I have found to be very important, and the study has found as well are:
- Long content ranks better. If you can write at least 900 words, you’re much more likely to rank.
- Quality is important. Quite bluntly, crappy content typically doesn’t rank well.
- Break up long content. Though keywords in <hx> tags is now not correlating as high, it does break up the content and significantly helps readability. It also can affect ranking slightly.
- Content that is easily understood by an average 12 year old is more likely to rank.
- Use semantically significant wording. In other words, use synonyms and other variations of your keywords. Google is getting much better at picking up intent.
Links are still very important too. I don’t want to get into this in great detail in this post, but if you check out Jon Cooper’s extensive link building post, you’ll get the info you need for link building. What you need to know is that:
- Quality of links is important (I still recommend using Domain Authority from Moz as the measure for this. Certainly no longer using PageRank).
- Number of links is important, but not if they’re spammy. Google has basically been destroying spammy networks.
- Anchor text is still very big. We have, however, seen that varying your anchor text with keywords surrounding links (think context) has been working very well also.
The Bottom Line
The industry can create studies all day, and we can analyze them even further. There’s value in that, but ultimately we already know exactly what Google wants. Google wants users to create good content. They want their search results to provide quality in the top spots (which draws users back to their site in the future, with a higher likelihood of clicking on ads). So, if you continue to create content on a regular basis, provide quality and value with that content, you’ll be in good shape. Content promotion helps as well, especially with building new links. Have a strategy, stick to it over the long haul, and you should be fine.
