Do Internal Links Lead To Penguin Slap / Algorithm-Based Penalties?

[Planet13]

Does anyone have any evidence (or strong suspicions) that over-aggressive INTERNAL linking can lead to either the de-indexing of a page or a serious Penguin / Algorithm slap?

After analyzing external back links pointing TO our site, it appears that the page that suffered the most damage had ONLY two back links; one questionable, one from a major US news outlet.

However, this page has the highest number of INTERNAL links pointing at it, and the INTERNAL anchor text is optimized.

[Saffron]

What do you mean? Linking to other articles within your site? If so, I'm wondering the same thing.

[rish3]

Finally, something that could be a "close to provable" case study on penguin recovery. Keep us posted.

[aakk9999]

I think you can suffer OOP with overly optimised and repeating internal linking (using the same anchor text over and over).

This may happen more in case where you try to rank for a keyword phrase (e.g. two or three word phrase) and you repeat the exact phrase over and over again in your in-content internal linking, where it would not naturally flow in the language.

Perhaps you can try to de-optimise internal anchor text and see if you get any improvements?

[Planet13]

@ Saffron:

"What do you mean? Linking to other articles within your site?"

Yes, WITHIN my site.

Since my site is both articles AND ecommerce, I mean linking from the content section of certain articles, as well as from the navigation bar of the ecommerce section to one page. The anchor text is the same (namely, "Keyword1 Keyword 2").

@ aakk9999

"I think you can suffer OOP with overly optimised and repeating internal linking (using the same anchor text over and over).

This may happen more in case where you try to rank for a keyword phrase (e.g. two or three word phrase) and you repeat the exact phrase over and over again in your in-content internal linking, where it would not naturally flow in the language."

so to summarize, you think the possibility of an OOP is higher from exact match anchor text from IN CONTENT links, as opposed to exact match anchor text from the navigation bar, correct?

"Perhaps you can try to de-optimise internal anchor text and see if you get any improvements? "

I was thinking that. But I remember a statement netmeg made a while back that she thought it was unwise to link to the same internal page on a site using different anchor text.

(Hopefully she will hear us talking in here and chime in.)

How would you deoptimize the anchor text from in content links?

suppose the navigation bar links that go to the page have the anchor text "blue widgets"

What sort of anchor text would you have for the in content links?

[taberstruths]

I would think that a navigation bar would be ignored but if you thought it was a problem you could always "no follow" those links. In content links should be viewed from a user experience standpoint. If they are being done to enhance the user experience like suggested articles or actual references to what is being talked about then I would do it. If you are just trying to use in content anchor text for seo purposes then I would remove it. Just because an article uses the word "love" on it does not mean the article is about love and that there needs to be a reference to the word "love."

[aakk9999]

so to summarize, you think the possibility of an OOP is higher from exact match anchor text from IN CONTENT links, as opposed to exact match anchor text from the navigation bar, correct?

Yes, I believe so. But I think you should still be careful not to have spammy-sounding options in navigation bar.

I was thinking that. But I remember a statement netmeg made a while back that she thought it was unwise to link to the same internal page on a site using different anchor text.

I am not sure if netmeg meant a lots of permutations of the similar anchor text phrase or linking with many completely different phrases - it depends on the context of her reply.

I tend to have anchor text I wish to optimise for from a few selected "stronger" pages on the site (but where the content of that stronger page is appropriate to have an internal link from), and more natural linking from other pages - where interlinking is done mostly because it makes sense for the visitor.

[aakk9999]

I would think that a navigation bar would be ignored but if you thought it was a problem you could always "no follow" those links.

I personally would not "nofollow" main navigation links, in fact I would not use "nofollow" on internal links at all - it can send a message "I do not trust these links".

If you are concerned, then just create main navigation that is aimed at visitors rather than at SE and you should be ok.

[Planet13]

@ aakk9999:

"If you are concerned, then just create main navigation that is aimed at visitors rather than at SE and you should be ok."

Yes, that is what I am TRYING to do. But the most appropriate anchor text for a visitor happens to be the most appropriate text for search engines, which has caused my site not to rank AT ALL for that keyword.

So i am thinking I pretty much HAVE to keep the anchor text in the navigation bar as "keyword1 keyword2" because:

1) It is a succinct, accurate description of the page to which it links, and

2) The "keyword1 keyword2" format is the SAME format as all the other links in the navigation bar.

I guess I am going to have to remove some of the in content links from various inner pages.
~~~~~

On another note:

anybody remember that Matt Cutts video from a few years back about what links to put on the HOME PAGE of a site, and he said something to the effect of having a link to your most profitable products?

I wonder if he would change that suggestion now? I wonder if in the age of Penguin it is a BAD idea to link from your home page to your most profitable product?

[EditorialGuy]

But the most appropriate anchor text for a visitor happens to be the most appropriate text for search engines, which has caused my site not to rank AT ALL for that keyword.

How do you know that your internal links have caused your site not to rank at all for the [anchor text] keyword? I haven't had that problem, and I'm sure there are many other Webmaster World members who use descriptive anchor text for internal links without having their rankings for [anchor text] disappear.

You said that the page most affected has two inbound links, one legit and one questionable. Maybe that's where the problem lies. Have you considered disavowing the questionable link and, ideally, getting a few additional legitimate backlinks for that page?

[Planet13]

"How do you know that your internal links have caused your site not to rank at all for the [anchor text] keyword?"

I admit that I really don't "know" anything. Everything with google nowadays seems to be a shot in the dark.

I do know that my "blue widgets" page stopped ranking for blue widgets some weeks after Penguin 1 (around late May 2012), and the home page started ranking for that term instead. Later on (fall of 2012) none of the pages on my site would rank for "blue widgets".

I went from number 9 to nothing. Oddly enough, it ranks for things like "porcelain blue widgets" even though I didn't have that keyword anywhere on the page.

"Have you considered disavowing the questionable link and, ideally, getting a few additional legitimate backlinks for that page?"

i was able to get the questionable link removed about two weeks ago.

However, back in about February of this year, I changed the blue widgets page from blue-widgets.html to blue_widgets.html, without doing a 301 redirect.

So even before I got the questionable link removed, it had been pointing for the last several months at a 404 page (of course, that meant the GOOD link was pointing to a 404 page, too).

Still, no pages on my site could be found for blue widgets.

At this point, I am pretty much "betting" without really knowing that the algorithm-based elimination from the SERPs was either:

1) from approximately four internal, in content, TEXT links on some of the most popular page on my site, or

2) From the site-wide navigation bar link that contains the anchor text "blue widgets", or

3) From the internal IMAGE-based link from approximately 40 of my pages to the blue widgets page, or

4) A combination of the above, or

5) some combination of the above in conjunction with some external backlink (either with the questionable link I knew about OR some other external link that I have not found).

[graeme_p]

I have been wondering about this too.

My biggest site is a reference work. It is natural to have an internal link to an explanation whenever I use a term that needs it - much as Wikipedia does.

However, how does Google decide whether this is natural or not? Lots of in-content links with the same anchor text...

[teokolo]

My top website (which is under a manual penalty for unnatural links) has been hit recently with a OOP or Panda penalty for the most competitive keywords, dropping from page 10 to nowhere.
I decided to analyze my INTERNAL links (about 5.000 pages) in a scientific manner, cleaning up exact match keywords and using descriptive links: site has recovered coming back around page 10 in few days.

[graeme_p]

@teokolo, were your exact match links all to a small number of pages, or to lots of different pages? What criteria did you use to decide what to clean up?

[chrisv1963]

cleaning up exact match keywords and using descriptive links: site has recovered coming back around page 10 in few days.

Google is becoming sickening. It looks like these days you can get penalized for linking to a page about blue widget by using the anchor text "blue widgets".

[teokolo]

@graeme_p, I had hundreds of pages linking to homepage with the same (or very similar) anchor text. I removed most of these links and edited remaining ones to be more descriptive.

For my analysis, I made a csv report of every internal link (using some linux magic: wget,sed,awk,grep) and I found some pages receiving too many links and very similar anchor text. I removed links from less relevant pages till internal backlinks profile seemed natural.

[aristotle]

The search volume for a particular keyword combination could also be a factor. If you direct most of your internal "link juice" to pages that target high volume search terms, that would be a form of over-optimization.

[Martin Ice Web]

I think this depends on the "quality "( after Googles Terms of low Quality sites ) of the site. In my niche most sites do heavy interlinking between internal pages. Breadcrumps links to the same site. Links to "Next item", "previous item" are all follow. They do very well with this.
MOst of this ecom Shops are full of links to "just visited"," best to buy", "other bought this", Bestseller links to internal widget pages. ALL are follow. There are more links then Information on the pages.
This sites have only one thin in common, they are big adsense buyers ( very big ). Most of them have overtaken page #1 with ads above, Shopping ads, and adsense ads on left bar. No, no good at all.

As i did everything to my site ( and now i thin i treated it not very well because it was going through several process of restructuring ) and nothing helps and every Little angry Panda and penguin and what the heck Little animals they have harmed my site and pulled the crap up. I did now come to the conclusion that UE may have to do soemthing with coding, type of leading the user by showing him drop boxes with filter options withhin the page.

Therefor I started yesterday night to recode one of my sites to multichoice user dropbox navigation (MUDN TRADEMARK ). I will tell you what I see. Unfortunately the site got hit today with -80%. So we will see.

[Planet13]

@ aristotle:

The search volume for a particular keyword combination could also be a factor. If you direct most of your internal "link juice" to pages that target high volume search terms, that would be a form of over-optimization.

Thanks for mentioning that. I think that is a good point.

In MY particular case, the keywords "blue widgets" is NOT a particularly high volume keyword, and in MOST cases, google assumes that searches for that keyword are for information more than purchases (at least as far as I can tell).

So the problem might be that I am linking FROM informational pages to my (same site) ecommerce page.

[indyank]

Why do you think it has been affected by the algorithms? Was it ranking earlier and has dropped now? If yes, for how long it had been ranking on top?

I don't see any issues with internal links or with informational pages linking to e-commerce pages as long as they are related and if they are in content links. If they are navigational links, it shouldn't really matter.

[Sand]

I think it's a definite possibility.

A friend of mine had a site that was decimated by the first Penguin. Prior to that, he had been at the top of the rankings for hundreds of terms for many years.

Analysis showed that he had very few backlinks pointing to his site, and the ones that he did have were editorially given from trustworthy sources. Nothing fit the pattern of what we know about Penguin, so it was a bit of a mystery.

The only thing we could find is that he was fairly spammy with his internal links. That said, cleaning them up never resulted in a recovery.

[Planet13]

@ indyank:

Why do you think it has been affected by the algorithms? Was it ranking earlier and has dropped now? If yes, for how long it had been ranking on top?

I am not sure whom you are asking.

If you are asking me, then please kindly re-read message #4589445

(In addition, I lost 35% of impressions and clicks from Penguin 2.0, so there are obviously external factors outside of any internal factors, too.)

If, on the other hand, you are not asking myself for clarification, please let us know to whom your question is directed.

[indyank]

Sorry i didn't notice that message earlier.

I changed the blue widgets page from blue-widgets.html to blue_widgets.html, without doing a 301 redirect.

Did you change the internal in-content links and navigational links to point to the new page?

Are the in-content links on pages which are topically closely related? How are those pages performing? Are they affected by any of these algorithmic animals?

Panda is a site wide algorithm. For penguin, Matt Cutts has addressed this question and I am sure he is right.

[webpronews.com...]

There was even a discussion thread here on the same topic.

[webmasterworld.com...]

[Robert Charlton]

So the problem might be that I am linking FROM informational pages to my (same site) ecommerce page.

It might help to think about such links, particularly contextual links, from the user's point of view.

If the user is reading an informational page and, say, your first use of the "blue widgets" phrase on this page is a contextual link, I think the reader might naturally expect that clicking it would lead to more information about blue widgets. The NYT and Wikipedia, eg, use such links essentially as footnotes. So, if it's not an informational link, you've perhaps misled the user. If I were so misled, I might leave the site... or certainly back out of the link.

Ditto with the image links... Was it clear where they'd be taking the user? Were they for Google or were they for the user?

Did you use analytics to get a sense of when users left those pages? Sounds like it might be too late now, but I'm guessing if you'd tracked your visitors, you'd see some dissatisfaction because of those links.

It's very possible to set up clearly marked "buy" buttons, or text links at the end of an article, that indicate to the visitor that, if they're interested, they can buy blue widgets from you as well as read about them.

While you may get fewer information page visitors clicking on those ecommerce links, it's more likely the visitors are going to the ecommerce blue widget page for the right reason. I think you're too fixated on that anchor text... and from previous concerns you've expressed about whether Google might not like your page if it was an ecommerce page, I think you've been trying to fuzz the distinction for years... when in fact clarifying the distinction might help you.

Ultimately, the success of your e-commerce pages, though, will be determined by whether your offerings are attractive in the market place. That's another whole issue. But they're definitely not going to be determined by the number of exact matches of anchor text you manage to squeeze into your site.

Regarding the issues of contextual links and site-wide links, these two threads might be of interest....

Introducing heavy internal linking to well ranking site
http://www.webmasterworld.com/google/4119868.htm [webmasterworld.com]

Site wide links and Penalties
http://www.webmasterworld.com/google/4513723.htm [webmasterworld.com]

[graeme_p]

I have stopped believing the "users point of view" bull#*$!. The Google algo is not human or an AI.

So how do we determine how many internal links with the same anchor text are too many for Google?

[Planet13]

@ indyank:

Did you change the internal in-content links and navigational links to point to the new page?

Yes, I changed those links to point to the new file name (blue_widgets.html)

However, fearing a penalty to the OLD PAGE, I did NOT do 301 redirects to the new page

Are the in-content links on pages which are topically closely related?

I would say yes, they are. It is hard to give an example here, but I would say so. However, as Robert Charlton brought up in the post after yours (read below for further discussion), it is a commerce page and people coming FORM an informational page might expect the link to go to an informational page.

One thing I should clarify; I only had about four of what we might call in-content text links pointing to that page.

However, I also had graphic "ads" to that page on about 35 pages. They were graphics that basically said, "Buy a blue widget today, save 10% off" I did those ads to drive traffic to that page.

However, I believe I did that recently, after I had falling out of the search results for the term "blue widgets"

EDIT:

How are those pages performing?... Are they affected by any of these algorithmic animals?

They were hit hard by Penguin 2 (down about 30%) They still are my better performing pages, and were previously ranked #1 or 2 for their key words. They did and still do get the lion's share of landing page traffic. Just that traffic is down 30% now.

END EDIT

according to google WMT, my whole site lost 35% of impressions and clicks on Penguin 2, and traffic is down about 35% as well (according to google analytics and statcounter).

[edited by: Planet13 at 3:03 pm (utc) on Jul 3, 2013]

[indyank]

If the user is reading an informational page and, say, your first use of the "blue widgets" phrase on this page is a contextual link, I think the reader might naturally expect that clicking it would lead to more information about blue widgets.

This is interesting. Assuming "blue widgets" is a product, linking to an e-commerce page does sound good enough as clicking on it would lead the user to the product page that might have pictures, videos (if any) specifications and its price. Isn't this good enough information for a product?

Wikipedia is a pure information site and hence all their in-content links lead to more informational pages. So IMHO, it might not be a good baseline to use in the Google algos , for these scenarios.

[EditorialGuy]

So how do we determine how many internal links with the same anchor text are too many for Google?

You probably can't, because it isn't that simple.

Ask yourself how you'd make such a judgment if you were Google.

Would you set an arbitrary number like "two internal links with the same anchor text?"

Or would that number vary according to things such as:

- The nature of the anchor text

- Whether other aspects of the site pass the "smell test"

- The site's authority

- The number of links on the page

- [Anything else that seems reasonable]

If you were clever, you might try to differentiate between clearly legitimate keywords in anchor text ("home" on any site, or "puppies" on a site about dog breeding) and and SEO-driven keywords in anchor text. And you might respond to seemingly SEO-driven keywords in different ways (count them, ignore them, penalize them), depending on the score generated by the aforementioned smell test.

[Planet13]

@ Robert Charlton:

It might help to think about such links, particularly contextual links, from the user's point of view. If the user is reading an informational page and, say, your first use of the "blue widgets" phrase on this page is a contextual link, I think the reader might naturally expect that clicking it would lead to more information about blue widgets... If I were so misled, I might leave the site... or certainly back out of the link.

Yes, I can certainly understand that.

But the metrics don't really confirm (nor deny) either way. It has one of the lowest bounce rates on my site. And while the time on page is lower than the site average, I feel that is partly because it is a category level page that links out to all the products in that category, so people who come to that page often quickly click on a link to the a product detail page.

The blue_widgets.html page does have one of the higher ecommerce conversion ratings, and since the vast majority of the traffic to that page is from people clicking links in other pages (since that page is not in google serps at all) at least some of the visitors who are clicking links in the pages are buying things.

However, I do think your idea of setting up clearly distinguished buy buttons is better.

[indyank]

Planet13, how does your e-commerce page look? doesn't it have information on the product in the form of specs, features, pictures, prices etc?