01 Aug 2013

4 Things SEOs Need To Be Testing, But Aren’t

During my talk this week at Search Exchange. I mentioned that a lot of SEOs spend their time speculating about what Google is actually doing. The truth is speculation is actually a big part of our job. The problem I see often though is most SEOs are speculating with little to no testing. What is speculation with out research/testing? Assuming. Our clients don’t pay for assumptions, they need educated speculations! Now don’t get me wrong many SEO’s do run test, and some do so in the complete veil of privacy. But it feels like most don’t. I know I don’t test as much as I would like. If I had more time I would. :) and I think I would start with the following areas. The below areas have been talked about a great deal in the SEO community, but rarely has anyone done any solid research.


User Behavior

Folks from Google are always telling web masters to build sites for the user in mind. And while many of us have brushed that off as fluffy jargon, Google does have several patents that focus on understanding user behavior. Which behaviors is Google targeting? How do these behaviors pay into ranking results? And does this type of behavior analysis leave lasting results? These are all things that need to be further understood through testing. Things like bounce rates, click through rates, and time on site can easily be tracked with in Google Analytics. Then other advanced metrics such as eye tracking, heat maps, and scroll activity can easily be analyzed with the right tools.


Freshness in the Rankings

A SEO prodigy of yesteryear, ViperChill, recently wrote a really great blog post about how fresh content is dominating the results. Glen claims that you can significantly increases the likely hood of a page or article ranking higher by simply changing the date on the page to something recent. He cited a few examples that he found in the results, and claims to have hundreds of others documented. Either way, I think this type of ranking signal deserves more testing on a larger scale to find if there is a high correlation. If so it could really help SEO’s and marketers in general make the case for things like content marketing, and blogging.


Transition Rank

Last year it was discovered that Google had filed for a patent for a new way to identify aggressive SEO behavior by decreasing the rankings of a site, and then monitoring to see if the site’s markup changed to rebound from the decrease. The idea here was that if a site gets hit by a rankings decrease the SEO will notice and make changes to rebound. Thus catching the SEO at well basically doing their job. even though I sort of hate this technique of punishing web masters just because they are trying to rank, I also think it is a very smart tactic for Google to use. And some SEOs are starting to claim that Google is already starting to do this. This could be one of the hardest things to test because it would require monitoring the rankings of lots of sites, and then trying to rebound from any changes to see what happens.


Brand Signals

I know it, you know it, and Google knows it! Google loves brands. But the one thing that we maybe don’t know a lot about is what types of signals tell Google a domain has a brand attached to it? There is Goolge+, branded anchor text, navigational queries, and other things that we all assume Google is using as brand signals, but very few are testing them. If we could lock down a better understanding of these factors, it is likely that we would have an easier time with all the other aspects of our SEO efforts because we would have those “brand signals” backing us up.

Can you think of any other areas of SEO that deserve more testing? Have you tested any of the areas above? Share with us your thoughts in the comments, and until next time, happy testing! :)