Exact Match Inbound Links & Penguin Worries

[sench]

---

System: The following 2 messages were cut out of thread at: https://www.webmasterworld.com/google/4819521.htm [webmasterworld.com] by goodroi - 9:46 am on Dec 14, 2016 (utc -5)

---

Anyone have a comment on the relation between exact match external links and the fact penguin is real-time?

Is it correct to assume if a page has 80% exact-match, then about 75% of all of the external links to that page will be devalued, causing it to lose ranking?

That's at least my understanding.

[sench]

Sorry, I mean inbound links, to be precise

[goodroi]

I think I know what you are asking but I want to double check. Are you asking if you have too high a percentage of exact match anchor text from external websites linking inbound to your webpage will that trigger Penguin and/or other link penalties that will lower your search rankings?

[martinibuster]

Anyone have a comment on the relation between exact match external links and the fact penguin is real-time?

Exact match anchor analysis is a statistical analysis process. In my opinion, Penguin is not a statistical analysis algorithm. So the one does not have a direct connection to the other.

The above statement does not however mean you have a free pass. It is only limited to answering your question if there's a connection between anchor text analysis and Penguin.

Good luck,
;)

Roger Montti

[aristotle]

In my opinion it's best to have a large diversity of different types of backlinks. A diversity of anchor text, a natural mix of dofollow and nofollow, a diversity of different types of sites providing the backlinks, etc.

[Adam_L]

We saw a drop in SERPS for our main term early September.
We went from 1st on desktop and mobile to 3rd / 4th on desktop and bottom of page two for mobile.
Due to the difference in the drop between desktop and mobile SERPS, we assumed google didn't like our mobile site and therefore we were being penalised because of mobile issues. Makes sense right? Wrong.

After checking our backlinks, and more specifically the anchor text, we found that we had 2 domains linking to us about 30-40 times EACH with our main search term as their anchor text. They were all DOFOLLOW....
These dofollow links with the same anchor text as our main search term were the culprit. They only accounted for roughly 1.25% of our total backlinks, yet since they were removed, we have seen an almost complete recovery for mobile SERPS within a week (from rank 17, bottom of page two, back up to 3rd position), and we are slowly recovering on Desktop.

Why we saw a larger drop for mobile SERPS is still a mystery. My best guess is the seperate indexes and varying ranking factors.

When it comes to real-time penguin, anchor text DEFINITELY matters. If you have 50 links coming from a single domain, all containing the exact same anchor text, this is a red flag for google, it screams 'PAID LINKS', especially if they are dofollow. The links won't simply get devalued either, you will get penalised. Once we corrected these links, 3-4 days later we started to recover pretty quickly, confirming for us that it was real-time penguin.

[martinibuster]

...we found that we had 2 domains linking to us about 30-40 times EACH with our main search term as their anchor text. ...yet since they were removed... Once we corrected these links, 3-4 days later we started to recover pretty quickly...

How did you get those links removed? Was there a direct or indirect relationship in between you and the site providing the links, like a broker or a link builder?

[Adam_L]

A while ago we took out an ad on the sites, so we had a contact with them. I think the ads should have been on the homepage only and nofollow, but they were dofollow, and added to any newly created pages.

At the time I guess we didn't really worry about it. Then of course penguin came and slapped us in the face and we had figure out why. We pinned it down to those links and immediately disavowed them, and requested they be removed.

I think the main issue was the fact that they were dofollow and all from the same domain. When it comes to the total % of inbound exact match links, as long as they look natural (i.e. a mix of nofollow/dofollow, and from various domains), then I wouldn't think pengiun would flag them.

[edited by: Adam_L at 7:49 pm (utc) on Dec 15, 2016]

[martinibuster]

A while ago we took out an ad on the sites...

Anddddd... that's what I thought, which is why I asked the question. :)

You had control over those links. More importantly, that site is selling links (call it advertising but if it's do-follow then it's a link buy). This is a factor that in my opinion is related to Penguin that cannot be dismissed.

IF the drop was caused by that site, I would be more inclined to pin it on other factors related to it's profile than to anchor text because statistical link analysis was how the spam fighting team sorted things out ten years ago. But Google's moved on and I don't think Statistical Analysis is part of Penguin.

Beyond that, I doubt it's not just that one link on it's own, (or twenty, thirty, or a hundred links from as many pages , as they only count as a single link anyway). If it is then there must be more issues going on with your own site's backlinks if that's all it takes to affect your rankings.

You also cannot rule out that the timing was coincidental.

[Adam_L]

I agree. Sorry, I think I misunderstood the original question. I didn't mean that the only reason we saw a drop was just because of the anchor text. You're right in that it was likely their backlink profile and the fact the links were do follow.

Are you saying that just one link from that site would have had the same effect as a link on 50 different pages then?

[martinibuster]

Are you saying that just one link from that site would have had the same effect as a link on 50 different pages then?

Strictly limited to addressing how Google calculates PageRank for ranking purposes, yes, that's how Google calculates the ranking power. Many years ago there was a thing called a ROS Link (Run of Site), and it used to transfer more ranking power than just a single link. Google considered that a "bug" so they changed their algorithm to equalize more than one link from a single site so that it passes no more link ranking ability than a single link.

ROS links are common in the blogosphere, even obtaining a ROS link from the navigation happens in the wild, with no perceptible change in ranking either way.

There was a ranking glitch some time between Thanksgiving and the first week of December. If your ranking change coincides with those dates then it could have been just that.

[sench]

thx Roger for your valuable thoughts and Adam for your case

I agree it isn't probable that the anchor text analysis is directly used by penguin, but is it possible that it's still used indirectly as one of the dimensions in the learning algo on what type of site is "a paid link site"?

as in, if I were P4: "there are 5000 links from this site to other sites on the web and 4000 of those links (grouped) point to a page that has the anchor text within most of the on-page elements => this is one of the predictors that point in the way of a paid link"

of course, that would mean the scope would be the site/page of the linking party, so the distribution on your site still wouldn't matter.

I'm not sure about what I want to say, but trying to wrap my head around this, so I'll just end here in hope for comments on this particular assumption (maybe your insights help me expand into the main point)

[martinibuster]

is it possible that it's still used indirectly as one of the dimensions in the learning algo on what type of site is "a paid link site"?

Let's imagine an HR person who has a tool that background checks job candidates to weed out those who are lying about their work history and education. Before this tool the HR person used to have to call each former employer and check up on the universities to verify the degrees.

Now that the HR person has this tool that can weed out fake bio's, is there a reason for them to spend time making phone calls to verify employment and education history?

I believe that Penguin works in a similar way. In my opinion Google is able to integrate resources to realtime Penguin because they don't have to devote resources for other less accurate statistical analyses. The reason they don't have to devote resources to statistical analyses is because they are largely redundant. That's my opinion.