The following article presents the first results of an experiment we are carrying out to test the effect of three types of anchor text when ranking a webpage.
We want to know what is the positive or negative effect in ranking of a keyword located on the same page as the link, depending on whether the keyword is placed:
- In the anchor of a link as optimised anchor text.
- In the anchor of a link to an image with the alt optimised.
- In the text surrounding a link without an optimised anchor (contextual link).
We made up a keyword to rank. This keyword did not generate any result at the beginning of the experiment. We explain here the reason for using invented keywords that do not deliver results.
Although this is a very simple experiment, we wanted to propose it in two steps to isolate the results: in the fisrt step we analysed only the internal links; the second step analyses the external links.
1. We take the invented keyword.
2. 15 days before the publication of this article we created the 6 pages (3 for origin and 3 for destination).
3. We send them for indexation through Webmaster Tools. The following day they were all indexed, so we just have to wait for the results:
After 15 days, the results are quite illustrative: none of the destination pages ranks for the keyword we made up. When we do the search only the original pages appear:
NOTE: Although we tried to make a neutral environment, it is hard to completely isolate an experiment in the search engine, as there are many factors at stake that we do not control and may be influencing it. Therefore, the following conclusiones shall not be taken as categoric statements, but as an explanation of what happened in this specific situation.
- Google did not take into account the anchor of the internal link of optimised context when ranking the destination page.
- Google did not take into account the anchor of the internal link in the image with optimised alt when ranking the destination page.
- Google did not take into account the optimised text for the keyword that was besides the contextual link when ranking the destination page.
- The “vote” of the internal links was not enough to convince Google that the destination pages are relevant for the invented word.
- The only pages that have ranked are those which include the word in the content (the original pages).
As a curious data, the result of the experiment gives room for some additional conclusion that has not a lot to do with the aim of the experiment. We mean the order in which the three pages are ranked:
1ª. The webpage with the word in the text.
2ª. The webpage with the word in the text as anchor text.
3ª. The webpage that has the word in the alt.
This makes us think that, all other things being equal, Google gives higher priority to the invented word that is within the body, without link, than to a word in an anchor text. It gives both more importance than a word located in an alternative text.
What comes next
We will check this experiment to see if results change over time. This is not very probable, because as we checked in previous experiments, the ranking for invented keywords is immediate, but we will keep an eye on it (just in case).
We still have to start the second part of the experiment. In this part we will add a new element to the equation: external links. With this we will figure out if Google is not ranking the pages for the made up keyword because the internal votes are not enough, and it prefers to have something external reassuring the matter.
Image via Balakov.