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?

[topstar]

major change to your site's content, content management system, or server architecture.

After losing traffic since last October (Panda) and now Penguin I redesigned the whole site and within hours of launching the new site it lost all top rankings. 301s and 404s are set up correctly. Is this sandbox or a new filter and how long would it take to come back? Anyone else had a major change (layout, urls, internal linking) during the last months?

[Planet13]

@Gmorgan I've also seen the direct opposite where homepages are not affected but deep pages are.

I see that too.

[Gmorgan]

There goes the homepage theory then.

The internal pages you guys saw drop, were they ones that had been an SEO priority (i.e. you'd keyworded them and pointed links at them)?

[topstar]

I am seeing an interesting recovery in one niche I follow. A spammer who lost his network of sites in Penguin is coming back with brand new domains and - guess what - similar automatically generated backlinks as before. The links appear on pages with dates Apr 23 and 24 - the Penguin date. So it took less than a month for him to get back to page 1 for a big money keyword. Spammy backlinks still pass link juice.

[Planet13]

The internal pages you guys saw drop, were they ones that had been an SEO priority (i.e. you'd keyworded them and pointed links at them)?

the target page dropped for the PLURAL version of the keyword (blue widgets) and was "replaced" in the SERPs by the home page at the same position. For the singular version of the keyword (blue widget), it still ranks in the same position.

Not necessarily an SEO priority. However, one of the keywords (blue) appears in the domain name as well as in the page name (w w w.exampleblue.com/blue-widgets.html).

There were no external links built to that page.

I did do some link begging to the home page, or to some article pages on my site which in turn link to the target page. Only got a handful of links in total (maybe around 10). But no EXTERNAL links directly to that page.

The internal links DO have exact match PLURAL anchor text to that page.

The keyword phrase [blue widgets] is repeated numerous times on the page, but it is a LONG page (like 12 paragraphs of text, each paragraph having between four and 6 sentences), and it would sound quite amateurish to use "it" and "them" repeatedly in lieu of "blue widgets"

Despite having fallen out of the SERPs for that keyword and losing its place as a LANDING page, it is still the second most popular page on the site, due to people following internal navigation / clicking internal links to the page, because it covers a subject that people who visit my site are interested in.

[dougwilson]

I'm probably not suffering penguin as bad as others. Overall, it's like other slow downs (Springtime Seasonal) but I have noticed that google bot crawling has flat-lined.

This could be due to Penguin's algorithm being written to "Find The Crap". No crap, on to the next site.

Of course when it finds all the crap it could end up in search results as an unforeseen byproduct event. Maybe an, "Oh, we didn't think of that", or maybe "The crappiest of the crap goes to the top so it can be easily seen and removed permanently" and "we did think of it, and we don't care if people suffer in the interim".

I thought that this could be one way to explain Google's recent search, search result, behavior.

[affiliation]

I dropped from #7 to page 3, here is what I have done so far in order to aid recovery. This is a site that I can take a chance with as it is not my bread and butter.

1. Over the past few weeks I have been de-optimising external links from KEYWORD to SITENAME com, Click Here, etc. I am lucky that I can get access to most. (Keyword density in my links are now below 50%)

This rose the site 5 places. after a week of doing this.

2. I made the harsh decision to de-optimise the home page, completely rewrote with better content, removed keywords from alt and title tags. The harshest decission I made was to rewrite the title and make it softer.

I had a bad experience of title rewritting in another site before, but I thought it was worth the try.

I checked today the keywords I normally have, dropped from page 3 to page 14. I guess this has not worked.

I apprecaite this is very early to see any changes, so I will play the wait and see game. I have sites that have not been hit, but what I cannot understand is the sites that were not hit is the sort of sites Google says Penguin would affect. Does this mean I should begin to build sites as I was building them pre 2008, bad content and not easy to read?

[tedster]

Remember also that Penguin (just like Panda) is not a real-time algorithm for now - it will refresh only periodically. So you won't know if your changes helped or hurt until Penguin re-runs. I'm not sure if Google will announce that or not, so we may need to watch this thread for any recovery reports - and until then, any changes will just be banging into the regular, real-time algorithm.

[Jez123]

Remember also that Penguin (just like Panda) is not a real-time algorithm for now - it will refresh only periodically.

How often has the Panda algo been refreshed? Is there any patterns to the times refreshed and so on? Just trying to make some sort of guess when Penguin might get refreshed. I was hoping that it would be today (being the 24th after they ran it on the 24th of last month).

[timwilliams]

Here's what I'm doing to recover from penguin, Nothing. It's been a refreshing few weeks working on our sites without the slightest concern for google.
I look at it like this, penguin is one of two things, it's either an adwords money grab and we're all screwed unless we advertise, or they're just a mess right now and they'll get it fixed. If that's the case and you truly have a good site then it should rise back to the top once google gets it right.

[Jez123]

@ timwilliams.

I almost wish I had done nothing too. If my site ever does come back it's not going to have any links pointing to it anyway - well just a few scrapers. I have removed or nofollowed everything that I control. I did nothing at all during Florida and it came back. I can't help but wonder if I have done the right thing. It /has/ forced me to look at every aspect of the site and make some improvements to content and code as well as having some time to add useful fun things like video tutorial on how to use the tools on the site. Just wish there was some visitors to see it!

[affiliation]

it will refresh only periodically.

There has been something happening today.

I have been looking at a few sites today that have been on the move in a search phrase I have been following for over a year. The old #1 dropped to page two at the introduction of Penguin, but has now moved up to #5.

I have seen the onsite work carried out for this, and this is why I decided to overhaul my homepage. His Meta Title was always over optimised, so he has slightly modified his title. Content, he has introduced another keyword to the content and made the article just short of 500 words.

I have not looked into the backlinks too much, but he had been pretty reckless with them in the past. I have noticed with the links I have looked into he has introduced topic / site related links rather than the URL and Click Here approach that many people advise. A pretty nifty approach when you think about it.

What I have seen today also is a site come from nowhere, straight to #8. 61 words of content and only a partial keyword related phrase. This has really threw me, what is someone doing there who is not trying for this keyword related term?

[affiliation]

Here is something else I notice since Penguin on a site I own which has never been positioned well and probably running a penality.

1. do a search & click page 5, 6 or 7 - my site does not show
2. go to page 8, then click back to page 5, 6 or 7 my site shows.

Is this another type of penality only showing if you back track or for every other search?

Anyone else see this?

[serenoo]

On 23 May one of my website recovered from Penguin. But to be sure I need to wait for next days to check if it continues to get clicks from google as it used to do before 24 Apr.

[Gemini23]

affiliation - I think this is likely to be a caching matter.... or due to 'personal settings'

[diberry]

Okay, this is bizarre. For the past three days, Google.com has sent me the exact same number of searches. And once it hits that number, it stops sending me searches for the rest of that day. This sure sounds like throttling, which I thought in the past had been reserved for bigger sites than mine (which is relatively tiny compared to the throttled ones I've read about).

Has anyone seen throttling in relation to Panda? Or did we ever figure out what throttling was supposed to achieve? And yes, I know no one's proved it happens, but it's awfully strange to have exactly the same number of searches from Google.com three days running. Nothing even close to that has ever happened in the history of any of my sites. Sometimes it's close, but never exact.

ETA: I just searched myself and clicked a time or two, just to mess with it, LOL. So today's number is now slightly up, but still? Too strange. I'll see if I get anymore natural clicks tonight.

[backdraft7]

@diberry - for the past 30 days, I have had the same number of sales every day +/- 1. Back in the good old days, we'd at least see some great days mixed in. Not sure how they figure this, since we have no conversion goals setup in analytics, but somehow they seem to know. They now have us throttled to just below poverty level, right were they want us.

[RedCardinal]

@backdraft7 if you seriously believe that then I think you need help for more than just your website...

[tedster]

@RedCardinal, I've seen stats like that. It can sound unbelievable until you actually see it - I know. But the phenomenon is real, if relatively rare.

[RedCardinal]

@tedster I don't doubt the stats or the reality of what's being described. I do however suspect that Google controlling for conversions is far fetched in the extreme. Controlling/throttling referrals might be trivial for them, but just think about what would be involved with them somehow tracking actual conversion on third party sites. In real-time (or near-real-time). Then applying this as a filter to their SERPs.

I'm afraid that sounds extremely implausible to me. Far, far more plausible that what you're seeing is an actual result of the Ux than anything Google controls.

Lastly the question would be why would they do this? What possible motive could they have for applying the required resources to do this, and what would be their return on this investment?

Perhaps I should have a more open mind?

[theskunk]

@backdraft7 OK - for sure I have considered the same in the past, but I doubt very much that this is possible when you consider the resources that would be required to make it work. Maybe tedster can offer more insight with some stats. Maybe this is a very good reason not to use google cart.

The controlling of traffic on the other hand, they seem to do very well.

[timwilliams]

I'm not completely convinced of the existence of throttling but it would not be that difficult given google's data. They can categorize the type of search a person is doing (nav vs trans) and make a statistical analysis as to what the percentage chance of each individual search converting is without much effort at all.

Now for the WHY. I know we are talking about the serps but this makes sense for both the serps and the ads. Google needs multiple advertisers for adwords to work in their benefit. If there was only 1 dominate advertiser in a niche that advertiser would not have to bid up their ad. Same goes for the serps, if a serp is dominating the sales then the advertisers go away. It's in google's best interest for everyone to make a little money and no one to dominate. Keep us all hungry but nourished...

[theskunk]

@diberry Yes i see the same on my 12 year old ECOM. After penguin i lost 20% of searches and the same 20% of visitors. But throttling was in place before Penguin. I guess its all about sharing the traffic, being fair to everyone, its a VERY easy way for google to ensure that its not just those that GAME them best that get the most customers.... it seems reasonably fair to me.

IMO

For anyone to seriously think that considering the billions of websites out there that their site is REALLY the best, and they are the only one deserving of a good search result ... c'mon

[tedster]

@RedCardinal, I agree. Throttling by using a conversion metric seems extremely unlikely to me, too. I think that it more likely to be just some kind of a coincidence that occurs at lower volumes of sales.

At the same time, conversion optimization is something every e-commerce site should be doing. If you can increase revenue by 20% from the same traffic and ranking levels - that's awesome. Then if traffic goes DOWN while the conversion percentage increases, you have a better reason for suspicions. I have never seen anyone with that situation so far.

[themaninthejar]

When doing a site:domain search I have always expected to see the site's pages linked using the title tag of the page. Last week I did the same search and discovered pages being linked in the following format:

Internal anchor text for the page - Site's main two-word keyphrase target - Brand name

I posted this elsewhere on here at the time. I'm reposting now as I just went though the search again and counted 72% of my pages listed in this way (last week I estimated just under half were in this format). I can only speculate about what this means (or even whether it's a recent development as I don't conduct this search very often) and would be interested to know if anyone else can see this with their sites. Is this Google making "notes to self" for the evaluation of anchor text on incoming links? Feedback appreciated.

Closer to the subject of this thread, I've conducted an extensive cull of keyword use across all elements of my site in an attempt to "flatten" out the overall usage to a more "natural" profile. I started out with the following profile for my five most used words info from WMT Content Keywords):

Word #2 of targeted keyphrase - 21% incidence;
Word #1 of targeted keyphrase - 19%;
Untargeted word - 3%;
Untargeted word - 3%;
Variation of targeted word #1 - 3%.

When presented in that format it's undeniably keyword stuffing (though I never set out to make it so).

I've seen a closing of the gap since I started the edit process on 19th April. The profile is now 16/11/4/4/3 and this has coincided with an improvement in ranking from page 11 (at the worst point) to page 5 (although, as ever, there's nothing to prove the connection).

I'll update as (and if) I see further improvements.

[Gemini23]

Themaninthejar..

Not sure what you mean by this...
"When doing a site:domain search I have always expected to see the site's pages linked using the title tag of the page. Last week I did the same search and discovered pages being linked in the following format: "

you search on Google site:exampledomain.com and you look at site links on there? how exactly are you DISCOVERING site links by Google SERPS?

[backdraft7]

My God! - You all think I'm hallucinating this? we've been experiencing throttling or traffic shaping since May Day hit in 2010. It' been discussed and observed between many other members. Not to sound even more "insane" but just because you can't prove UFO's exist, (while they are observed regularly), doesn't mean they don't exist. I'd love to share my data that shows the natural variation we used to experience vs. the tight, limited data that has evolved since May Day.

Google has been collecting data on our site for 10+ years and it remains a small 78 page site with tracking code on each page. I just don't set goal pages, but I'm sure Googles algo can figure this out. It may also be that the years my feedback between GAW and their help forums perhaps they have set me up as a test case, a sort of "landmark" in their data set. This also works in with the Zombie traffic issue, another well observed, (but due to Google's lack of transparency), poorly understood shaping topic.

At fifty years old and semi retired, my experience as a former systems engineer on everything from early image processing mainframes to modern industrial automation controls systems, I guess I'd have s slight understanding of what 'is' and what 'is not' within the capabilities of a computing system. Your assumption that Google is incapable of this type of data manipulation is what's implausible.

Nobody, but NOBODY can lay claim that they know exactly how Google operates. It's nothing but a big black box with a poorly written manual. That's the way they want it. All we can do is observe it's output and try to dissect hints from that data.

So, since you're new RC, I'll let your comment slide, but you need to read up on this subject before suggesting I need "mental" help. And, yes, you should have a more open mind.

Thanks for having my back (on the initial reply at least) on this one tedster!

[themaninthejar]

@Gemini - Search www.domain.co.uk in google and then click "More results from www.domain.co.uk"

[diberry]

Hey Backdraft7, a few people have come in and backed you up, for what it's worth. I agree that RC's phrasing was inflammatory - one can express healthy skepticism without suggesting someone else needs mental help.

For anyone who's interested, here's the old thread which provides lots of stats and IMO makes it extremely hard to doubt that at least some kind of throttling goes on: [webmasterworld.com...]

@timwilliams makes a good case as to why Google might engage in throttling conversions and not just traffic.

This, btw, is why I got rid of Google Analytics last year. The more chance we give Google to know who's visiting, what they're doing when they get there, etc., the more *opportunity* they have to make sure the algo profits them at our expense. I don't claim to know that they ARE doing that, only that they COULD, so why even make it possible?

@backdraft7, I don't know if getting rid of Analytics is practical for you, but I don't see how they could continue throttling anything but the number of searches they send you if they can't track what else is happening on your site.

[tedster]

@backdraft7, you know I've been a believer in traffic throttling for years - just not that Google uses your conversions to create it. On that idea, I have yet to see data that convinces me, as I described above.