Page 25 - SEO Mad Scientist 2020 Report
P. 25

Update: April 3, 2020

               After adding our table of contents, we saw no additional ranking movement. We believe the way
               the jump links are setup could play a part in effectiveness, but simply having them may not make
               a difference. We hypothesize you must combine multiple SEO tactics for these to be impactful.

               We checked back April 10, 2020, but the tests had some mixed data. It was hard to ensure our
               findings were due to the jump links and not content-related, so we restructured these tests to
               reduce content variables and re-ran them.

               Test Results: May 1, 2020


               After reindexing the jump links, we began to see an increase in site impressions! At the time the
               data was limited and we couldn't draw too many conclusions so we set out to get some answers.

               The first thing we found was that although jump links do show in the SERPs, these links are not
               indexable on their own. They are always accompanied by their root URL. Meaning, if you put a
               jump link on a webpage with a keyword in it, the jump link URL itself is not a direct ranking signal.
               If your keyword is ONLY in the jump link URL, the page will not show in SERPs for the term.

               While we now know these links are NOT a direct ranking signal, we wondered if they are indirect
               ranking signals? In the previous test, we saw a correlation between impressions increasing when
               placing the jump links on our website. The fact is though, there are many things that could have
               influenced this.

               The table of contents was pulled from the headings meaning - we added all of our heading terms
               on the page, increasing our term frequency. This could have caused the increased relevance.
               The jump link URL could also have been an indirect signal as mentioned or possibly the links
               themselves could have been counted as internal links on the site.

























                                                                                                          25
   20   21   22   23   24   25   26   27   28   29   30