Google and nofollow Links: What's New in 2019

Ing. Philipp Doblhofer

Ing. Philipp Doblhofer

04. Oct. 2019

In September 2019, Google announced news about the handling of nofollow links. But to understand these better, first a small digression: Why are website links important for search engines? Why can they influence rankings positively, as well as negatively?

PageRank: The Basis of the Search Engine

There are over a billion websites around the world. To be able to find desired content at all, search engines were developed early on. Upon a search request, these list suitable results - sorted by relevance. But the question arises, how can it be decided what is relevant? One simple way would be, for example, to rank according to the frequency of occurrence of the search terms. However, this creates great potential for manipulation. It would already be enough to mention the search terms on the page several times for which one wants to be found (so-called Keyword Stuffing).

To circumvent this problem, Larry Page (the namesake of PageRank) and Sergei Brin had an idea, from which Google soon came to life. The idea is based on scientific papers: The more often a scientific work is cited, the more important it seems to be.

Applied to the PageRank process, this means: The more often a page is linked, the more important or relevant it is likely to be. The ranking also considers how high the PageRank of the linking page is: That is, if a page has a high PageRank, it passes on a lot of PageRank through its links.

But this is not the ultimate wisdom. Here too, the results can be very easily manipulated. Whoever generates the most links to their own pages would simply move to position 1. Guest books and forums were massively flooded with spam posts due to this loophole. Because the only thing that counted at that time was to get many links to one’s page - no matter where they came from.

An Attempt at Remedy: “nofollow“

To regain control of the situation, a link attribute was introduced in 2005 that allows individual links to be “devalued”. Instead of marking a link as usual in HTML:

<a href=“https://www.codeaware.at/“>codeaware GmbH</a>

from this point on, the option rel=“nofollow”

<a href=“https://www.codeaware.at/“ rel=“nofollow“>codeaware GmbH</a>

can be used to indicate that these should not be considered for the PageRank calculation. Thus, it can automatically be communicated to search engines that this is not a “recommendation” according to PageRank for all links from comments, guest book entries, forum posts, etc.

In Google’s Webmaster Guidelines¹, for example, it is explicitly stated that links that could lead to manipulation of rankings should be marked as nofollow links. This includes, among other things, paid links (e.g., advertising) as well as excessive link exchange.

Unnatural Links: Penalties through the Google Penguin

But even “nofollow” attributes could not contain the manipulations of the rankings. The next step were the so-called Penguin updates between 2012 and 2013. Through these, an attempt was made to recognize pages that have an unnatural link profile.

That means: A website that has, for example, 4,739 dofollow links (i.e., links without a nofollow attribute), but not a single nofollow link, seems suspicious.

Since then, it has also been examined what link texts (anchor texts) are used (are “hard” keywords always used?). How quickly does the number of incoming links increase? Are the links from topic-relevant pages? A link from a gardening forum to an inn is in most cases not natural.

Individual links are not critical here. However, if there are numerous indications of an unnatural link profile, this can lead to a penalty or even de-indexing of the website.

The Next Attempt: Medic Update

However, it was time for a major update again in 2018. To achieve even better quality of search results, E-A-T has been playing a major role since this update.

E-A-T in this case means: Expertise, Authority, and Trust.

It is therefore important to be an expert for your own content, to be an authority, as well as to be trustworthy.

One approach here could be, for example, the so-called TrustRank. Here, pages with 100% trustworthiness are manually selected. Each outgoing link from these pages only receives a part of this trustworthiness. Thus, the further a webpage is in the link graph from these “trustworthy pages”, the less TrustRank is granted to this webpage.

What are considered pages with high E-A-T values for Google can be read in the Search Quality Evaluator Guidelines². These are the guidelines for people who manually judge pages for Google according to their quality.

What’s Next?

Out of fear of “unjustifiably” passing on PageRank, nofollow is now often marked on all links to external pages. A well-known example is Wikipedia. All references are linked with nofollow there - to keep spam attempts low. But isn’t a webpage that is mentioned in Wikipedia - provided it is not spam - also relevant for search results?

To solve this dilemma, Google has recently introduced further link attributes:

The biggest change here is that from now on, nofollow links will no longer be definitively excluded from the ranking. The nofollow link attribute is now only a hint for Google. Google therefore decides for itself which links will be included in the site ranking from now on, and which will not.

How these changes will affect the search results will be shown over time. However, it is clear that Google will in the future have further parameters to help identify unnatural link profiles through these changes.

The original Google article on the new link attributes can be found here: https://webmasters.googleblog.com/…/evolving-nofollow-new-ways-to-identify.html