Ben joins Rand at the Whiteboard this week to talk correlation data and how to use it.
How do we collect it?
We take a huge list of queries and test them on the search engines.
What have we seen?
Most of what is gathered from the collected data is what you would expect.
H1s – there isn’t a great correlation between having header tags and ranking better. However, this could be due to H1 tags being used inappropriately.
Alt text – good correlation with better ranking.
Title tag positioning – getting your keywords at the front of your title tag is much better than later.
URL formats – the longer the path, the worse you will rank.
Link data – external links seem better than internal links, and having a diversity of linking domains is always good.
How should we apply?
Remember that just because H1’s don’t appear to rank better from the correlational data, that doesn’t mean we shouldn’t use them. They are still excellent for usability, for example. This is the same for any correlational finding – question as to how you can best apply it, but without hurting your site.
Being careful with the correlations
Always keep in mind that these findings are CORRELATIONS, not CAUSATIONS. Therefore, consider how you can apply them to your site, but make sure it makes sense for your site.