Penguin Recovery Tips - a think tank thread

[bostonyear]

Since the main Penguin Update thread has 700 posts and counting, I'm hoping to start a new thread solely focused on Penguin recovery tips. I have a site that was hit by Penguin and I am trying to work my way out of it.

I think reason I was penalized was my content. I was inadvertently keyword stuffing. This is just the way I have been writing content for years. I have updated the content on my main pages where I have fixed the blatant keyword stuffing. My density levels are much more in line. My main question is:

I have over 80 blog posts that have some instances of keyword stuffing. Do I need to go back and fix all of these pages? Some of the posts are over 3 years old? I also have some really old pages that are buried in my site that may have poor content. Should fixing these old pages be a priority?

[driller41]

Has anyone any thoughts if directory links using optimised anchor text have caused any demotions - that is a theme in my demotions.

I used a directory submission tool to submit to loads (500 ish) of directories using optiised anchor text, it seems a silly thing to demote a site for in my opinion built the evidence is there.

[Hissingsid]

The site at the top of serps for the top term in my niche has 90% paid directory links, you know the kind of $250, 100 directory, optimised anchor text, package providers.

The domain name is an exact KW match.

The KW density is almost 8% for the term but interestingly there are no meta tags. So no meta KW or description only a title and the title is 40% the term.

[Robert Charlton]

I thought these had disappeared ten years ago...

I used a directory submission tool to submit to loads (500 ish) of directories using optimized anchor text

The common wisdom used to be that if you were lucky, these would simply create a lot of email spam for you. Why are people suddenly surprised that hundreds of such submissions would create problems?

[themaninthejar]

Logically, Directory submissions could always be justified by the hope of getting click-through traffic, irrespective of any incidental backlink value. Therefore, if a search engine doesn't like them maybe it should discount them completely rather than penalise them.

[driller41]

Well they are certainly low value - I would agree there, in fact they are useless, but I had a tool to add the links - Sliq submiter and thought I may as well use it, the belief being that Google would have the common sense to ignore any links it did not like - but after Penguin I am questioning that assumption.

I must stress that directory links are only a theory - I still find it hard to believe that they would downgrade sites because of these links.

Does anyone concur - are directory links an unlikely target for Penguin?

I must say that once again google has got us all scared of linking for promotion - which is one of the goals of course.

[Hissingsid]

Is the fact that they are directories just a co-incidence. Could it be same anchor text from 500 sites regardless of what sort of sites they are?

Could it be a combination of this, KW packed description meta and sitewide anchor text all just adding up to over optimisation.

I was interested to see that a site with no meta description or KW is doing well despite having bought links from a directory network.

[driller41]

Well that is what I concluded Hissingsid - I just cant get my head around Google demoting sites because of over optimised directory links - they certainly did not benefit rankings before Penguin so why kill a site now?

[Wilburforce]

Could it be same anchor text from 500 sites regardless of what sort of sites they are?

That certainly looks like what is affecting me. There are two search terms - one the main money term, the other a technical help term - that are in high quantity backlinks (none from directories, and none of them solicited), and both terms have gone from top of page 1 to pages 4 and 5 respectively, with a downward lurch of 5-10 places every week or so.

The penalty applies to exact match of anchor text (so widget is clobbered, but movable widget isn't).

Other terms (including ones that find those pages) do not seem to be affected.

As I have already mentioned, the periodical drop makes me suspect Penguin is recursive, and probably applies to exact match anchor text.

[themaninthejar]

Could it be a combination of this, KW packed description meta and sitewide anchor text all just adding up to over optimisation.

I'm focussing on the onsite keyword density primarily (now paused waiting for WMT to catch up and give me an updated Content Keyword report). I'm waiting for more evidence that backlinks are a big enough part of the pie to merit my concern, although you can't ignore the emails many people have received telling them about "possibly artificial or unnatural links pointing to your site that could be intended to manipulate PageRank". I have not received such an email, but I'm unsure if its absence means my backlinks are all safe.

My thinking is that any on-site de-optimising work I do has the added benefit of making the site slicker, faster and improving visitor experience slightly. Any off-site de-linking I do is possibly time down the drain with no added side-effect benefits.

Also onsite work can be reinstated (in part) more easily if appropriate in the future.

[Hissingsid]

Another site I've just looked at has a large block of CSS in <style> tags but not in the head. They also pull in Google jsapi and the associated default style sheet for this in the body ie not in the head.

This is not in a HTML5 document so as far as I know this is an error. There is some other HTML5 syntax in there but it is definitely not and HTML5 document.

It looks like they have made a mistake, it still looks OK in browsers but the effect of all of the text in the css is to dilute the KW density.

[themaninthejar]

dilute the KW density

I'm hoping for WMT Content Keyword update very soon. It's averaged once every 4 days since mid-April and the 24th was last update I saw.

[diberry]

I must say that once again google has got us all scared of linking for promotion - which is one of the goals of course.

Interesting that you mention this, because a couple of months ago I had a brunch with some people from a CPM ad network I use. I was talking about how one gets afraid to promote one's site because Google might see it as "linkbait" or "overoptimization." I've avoided not only promotional linking but "guest posts" and all sorts of other forms of promotion that existed (or had equivalents that existed) long before Google. I joked, "It's as if Google doesn't want anyone but themselves making money online."

To my surprise, they nodded somberly and said that was exactly it. Now, they're a CPM company competing directly with Adsense, and that's their perspective.

That really solidified my choice to start doing whatever makes business sense, and whatever Google makes of it I'll just have to live with. Because at the end of the day, if some form of promotion is bringing you more traffic from sources other than Google, then if Google dumps you at least you still have that traffic.

[My_Media]

I had some thin content - about 8% of my total site content. There were long articles which I broke up into several small articles and published under slightly different titles. At the time (2010) I thought it would be better to have more content. Now I am seeing that these short articles are not ranking or bringing in traffic.

In an attempt to undo this, I am merging these articles together again where necessary. For example, for 3 articles of 300 words each, I now have one article of 900 words. I have done 301 redirects from the old articles to the combined article.

Within days Google removed the links of the old articles. I am wondering if this type of restructuring will ultimately hurt my site? It has been badly hit by Panda and Penguin updates anyway ...

[textex]

We are down about 40%. Our link profile is our biggest issue. We've done some pruning...but, for links that we think are of value and are not ones we want to prune, is diluting the anchor text and switching it to our company name or something like 'click here' going to be effective? Or is the link already shot since we used KWs in our link and therefore changing it will not do anything...and its time to move on?

[Wilburforce]

Or is the link already shot since we used KWs in our link

Keywords in backlinks is definitely an issue (not sure about keywords in internal links).

It seems very clear to me that there is relationship between exact match anchor-text backlinks and search term. If masses of backlinks are for Widget Sales (which is what my business is), and the search term is Widget Sales, my page is a long way down the results.

This doesn't happen if the search term is Widget Sales Place, or Special Widget Sales.

One thought for a possible mechanism is this:

What if Google now counts exact-match backlinks towards keyword density in some way? My on-page density for Widget Sales may be reasonable (with an on-page count of 12). If there are 2,000 Widget Sales backlinks - OK, perhaps these are offset by some factor - it crosses the threshold, and there is nothing I can do about it on-page.

[textex]

I know its an issue. But, is it worth changing the existing link from kws to something generic? Will changing an existing link help? Or is the link shot and its time to move on?

[Wilburforce]

If there are many links and they are all in your control, then changing some of them might help. If those under your control are a small fraction of the total, it probably won't make any difference.

The issue I have identified is with exact matches, so including more or different words in the anchor-text might be effective.

[Hissingsid]

It seems very clear to me that there is relationship between exact match anchor-text backlinks and search term. If masses of backlinks are for Widget Sales (which is what my business is), and the search term is Widget Sales, my page is a long way down the results.

If this is true then it is definitely on a term by term basis. Using your example for <<Widget sales>> I have exact kw anchor text and my site is suffering, just a few places. For <<big widget sales>> I have a similar level of exact anchor text but I'm at #1 and #2.

[textex]

It definitely is term by term.

[gouri]

From what you guys have seen, is Penguin phrase specific or is it a penalty given to a site?

I see some posts here about how it is phrase based, but I am thinking that it applies to a site.

[Wilburforce]

Yes, term by term.

Using your example for <<Widget sales>> I have exact kw anchor text and my site is suffering, just a few places. For <<big widget sales>> I have a similar level of exact anchor text but I'm at #1 and #2.

Are these linking to the same landing page?

[gouri]

From what you guys have seen, is Penguin phrase specific or is it a penalty given to a site?

I see some posts here about how it is phrase based, but I am thinking that it applies to a site.

I am seeing less visitors to a site for certain phrases that I have not included as anchor text in external backlinks. That is why I was wondering if it is sitewide, and not just for phrases that might be used as anchor text too often?

[Wilburforce]

From what you guys have seen, is Penguin phrase specific or is it a penalty given to a site?

The main effect I am now seeing is phrase-specific. It seemed more general at the beginning, but I have made a number of on-site changes (principally, removing external links), and results are now generally OK apart from a couple of key terms, #1 money term being one of them!

Even at the beginning there were a few pages and terms that were not affected, so it wasn't site-wide in my case, but I wouldn't like to generalise.

[diberry]

What I'm seeing is I lost rank on my most competitive phrases, and "my domain", but I'm still ranking well on some less competitive terms, plus a lot of longtail ones that keep shifting around.

But I just got the "no manual spam action" email in WMT, so there's a third possibility. What if Penguin involves both penalizing certain shifts and affecting certain phrases in the algo? That could explain why some webmasters are seeing huge drops overall that very much look like a sitewide penality, while others are just losing out on certain phrases we used to rank well for.

Fourth possibility: Penguin is strictly about changing the SERPs on certain phrases, but Google also handed out a bunch of manual penalties at the same time without publicizing that side of it.

I've been monitoring the competitive search terms I used to rank highly for, and I just don't understand how the results are good for users. In one example, it's:

1. TV channel's website
2. Another result from same website
3. A link list page from About.com
4. Big brand site
5. Images
6. Same big brand site as #4
7. Another link list page from an established site
8. Another big brand site
9. Yet another big brand site
10. YouTube video

There are only actually 6 resources in that top 10 list, unless you count YouTube or images as a resource, which most people wouldn't because this query is hunting for some simple, printable text. As a user, I feel like the whole first page is just thinly disguised ads when I get a SERP like that. I know non-techie users may see it differently, but who doesn't feel their time is being wasted when you only really get 6 resources where you're expecting 10?

I never understood Google ranking my page so highly on this term, but nor do I understand how this new result is any better. Except, of course, for the big brands.

[tedster]

no manual spam action

Right - Penguin is an algorithm change, not a manual action.

A standard reconsideration request will not help you, but if you feel your site has been impacted by a false positive, you can submit Google's special form just for Penquin: [docs.google.com...]

[Hissingsid]

I'm pretty sure that what I'm seeing for my own site is a phrase by phrase issue. Having said that, in my serps there are some whole sites that have suffered for all or at least many search terms.

I'm just hypothesising that there is something related to anchor text on and off site and keyword stuffing in certain areas (types of tags like image alts for example) of pages that cause the term by term effect. Anchor text in sitewide links like the big footer on every page with links to "key" pages on the site may also be involved in this.

Then there's a different ssue that affects a whole site that is caused by link buying, hidden back links, links from certain networks etc.

If you have both then Ker-Bam! you're gone.

[Wilburforce]

Google's special form just for Penquin: [docs.google.com...]

This is a strong clue: they are asking for _affected page_ and _specific query_.

[Hissingsid]

I'm interested to hear from others on how quickly changes resulted in improvement.

I did a bunch of changes on 27th and today 30th May I've moved up 4 places to #3.

Changes made:

Changed anchor text in some backlinks I have easy access to reduce keyword in anchor.

On the home page only:
Reduced KW density overall
Reduced KW density in image alts, anchor, description
Removed <meta name="Abstract"
Removed <meta name="Keywords"
Remover title="keywords" from links

On the sitewide links:
Reduced KW usage in sitewide links.

I have no evidence that there is a cuase and effect ie the 4 place move might be coincidental to these changes but I'd like to think that the changes have caused the effect. Is 3 days too quick for this to be the case?

[themaninthejar]

I used google.co.uk to look up the word "Bisous", which a rather lovely young lady from Paris used at the bottom of an email to me. Turns out it means "kisses for a friend".

As with most searches I do these days I had a good close look at the top ten. All the top positions are definitions of the word in one form or another, but there at #8 was a British site selling Belgian chocolates made to look like a human @nus... Well it certainly caught my attention... Why was this site anywhere near this serp?

Looking at the source code it transpires that the word "bisous" is included in the meta-keywords. There's no inclusion of the word in the text on the page and nothing in backlinks that I can find.

Although I'm loathe to draw any firm conclusions, it does suggest that Google is paying some form of attention to the meta-keyword tag. As this tag is an obvious place to keyword-stuff it might merit another look at your tags. I actually went through my site a couple of weeks ago and reduced the content of the keyword tag to one word or a two-word phrase that best described the page at hand. I think this example serp justifies the effort that took.

Incidentally, that serp still stands at the time of writing if anyone wants to do any further research...

[crobb305]

Hissingsid, I made some changes on/around the 27th and saw a 20% to 30% increase in traffic yesterday. I am like you in that I can't say the two events are related. Maybe there was an update to one of the algorithms.