I know I've been MIA for a really long time, but it's always good to see familiar names and faces here. 3 babies later I finally am at a point where I can focus on work and am starting to rebuild my practically extinct business. I've only just looked around a bit so I'm not sure what sorts of things you guys have been talking about but I thought I'd start by sharing some of the info that I've been finding regarding penguin.
Until now I've been running a couple of large sites, but they got hit really badly with Penguin. I did a ton of backlinks analysis (loving linkresearchtools) to compare our site, with sites that dropped, with sites that are doing well. I found a few things:
* My sites had pretty high anchor density with just a few target keywords
* All of my links were paid links
* The backlink anchors were very artificial. All of them were "kw kw" and we didn't have enough variety (click here for ___, Site.com, etc)
* We didn't have much of a presence on social media sites
* Our ratio of paid artificial links and natural looking links was just too low
The sites that are doing well have pretty good looking backlinks profiles. Everything that I mentioned above was reversed on sites doing well. One of the sites that has a gorgeous backlink profile:
www.brickhousesecurity.com
Consequently, I think this update has been about increasing the importance of social media markers and givin less importance to backlinks and anchor text. By replacing social media markers with anchor density and other easily manipulated metircs Google can finally combat paid links, link networks, google bowlinig, etc.
I don't know how many of you are still doing old-school SEO with smaller sites/brands but to me this is a pretty big game changer. I believe that to rank well now requires systems and processes for white hat link building as well as active social media campaigns.
A few ideas of the methods I was planning to use to combat this:
* Create and promote infographics
* Create a database of sites that accept guest bloggers and work on establishing a name as an author and then distribute content on these guest blog sites.
* Run contests to increase likes and tweets in social media world
* Doing old-school PR and contacting sites, establishing conversations, and trying to figure out how to get people to pick up stories and share our stuff
I am having trouble systematizing all of this. In the past things were easy because all we did was built up a large network of people willing to sell links, then we found clients, sold links, ran reports to make sure the links were still up, paid through automatic paypal subscriptions. Bam. It was a hands-off, easy business model. It was easy, made money, effort-less.
Now, it seems very labor intensive. I hired an intern at $10 an hour who is a communications major who I believe can work on all of this, but that isn't terribly scalable.
Anyways, just thought I'd throw out some thoughts/observations and see if any of this rings a bell with anyone.
Hope everyone is doing well!
Glad to see you back in the saddle, we need more discussion like this.
Personally, what stands out on my winners (or keepers) and losers are:
* My sites had pretty high anchor density with just a few target keywords
* The backlink anchors were very artificial. All of them were "kw kw" and we didn't have enough variety (click here for ___, Site.com, etc)
Pretty much sites that had a couple or few campaigns run on a few specific keys without much variation on anchors and link types seem to have suffered.
This doesn't seem to affect my stuff much:
* We didn't have much of a presence on social media sites
I have some sites with a social profile, getting likes and such regularly, but took a hit. I have sites with no social stuff that are doing better.
* Our ratio of paid artificial links and natural looking links was just too low
I don't see any difference in link type or placement as long as there is some variance. I think the paid type and natural looking can be lumped together when not supported by lesser fill-in type links. More of a natural looking link profile overall, not quantity of natural looking specific links, if that makes sense.
My guess, and I've not really looked in depth as I think it's still early, is having little anchor variance on lots of links, and having a lot of similar links with no variety can hurt you. All in-content and in-context links sounds like it would be great, but in my mind a site should also have some blogroll/sidebar/etc type links, and/or comments, and/or forum links, etc. Social can play a part in it (a link is a link), but doesn't look like a specific social requirement from where I'm sitting.
I'm not sure social is a requirement at present but I'm pretty convinced its moving that way - not that a site needs its own huge social presence, but that within its niche it should have at least an average number of social mentions, links and interactions. I'm thinking of the social graph as a bit like links used to be - the way people talk about sites online has changed totally from where it was 10 years ago, but the 'if people are talking about it then that's a recommendation' is still as valid today as it was then, if the way they measure 'talking about it' can change.
Harder to game or shortcut to this though as I'm a lot less confident of my ability to create social links which will pass a hand inspection without hard work than I am of being able to get paid links which would pass hand inspection. That's definitely possible at the moment but all of those social links which are being mapped are really hard to replicate in a high enough volume network to be useful, moving forward, imho.
From what I've read so far (it's too early to test myself) it looks like penguin is to do with anchor text variance.
It seems that if you have too many links using a keyphrase then it backfires and has a null/negative effect.
TBH, if you think about it, what's common with most backlinks that try and game the results? A high percentage of keyworded anchor text for starters.
To me this make perfect sense and I'm surprised this hasn't happened sooner.
Have you guys seen this case study? http://www.branded3.com/tweets-vs-rankings
Provides interesting reading on social signals, albeit just twitter ones :)
One thing I'd mention re penguin is that it appears to be very head-term specific. Big assumption here: The pain I am dealing with at the moment is penguin related - however the timing at least supports that.
Big, obvious head terms that have been supported with exact match anchors have dropped. Less obvious, but still valuable terms that had been targeted in the same way still OK and even up. In fact it's mostly home + 2 major category pages that have tanked for us.
> mostly home...tanked
Ditto.
A site where I've done next to nothing linkwise for the domain but plenty for the deep pages is doing fine.
Damn penguin.
For me I had a few sites and 1 client site tank. I think it's looking like anchor text because well, pre penguin that's how to get a site ranking. The sites that survived are the few I made directory links to, these are pretty good because their anchor text is either URL or Visit Site.
What I am seeing working now is:
- Sites with weird paid links - I have a debt management site doing fairly well with this.
Big Brands - Because they get a lot of "brand" links eg http://www.brand.com/category
Weird 301 things happening. I haven't tested but maybe 301ing is a way out of this. or 307?
Looking through digital point (i know...) it's interesting that a lot of the sites selling links are now showing URL anchor text in their blogrolls. For at least 1 site I follow, this is working.
EMD + 1 page of content... :(
High quality spun content on 2.0 and good article sites is working for a friend of mine, I'm trying to replicate. These are then backed up with RSS and bookmarks.
Strangely I'm not really seeing a lot of EMD's falling as harder than over optimised non-keyworded domains. Sure some have dived but overall it seems an EMD will drop 5-10-20 places but a non-EMD will drop 4-50-100.
2much back in the saddle?! Awesome, good to have you back here girl :)
And back with a great post :) I think you're on the ball here - it all boils down to anchor text imo. Think it was Dave (N) that said somewhere that;
there is no anchor text
I kind of believe that as it would kill a lot of the text link spam. If Google would switch to looking a lot more at neighborhoods, context, authority and social media - they could more or less discrad anchor text - and even penialize for it as they know it's seo's that work the anchors. Everyone else is just happy with a "click here" or "read more".
It's about trying to fake a natural link profile - which is hard for an seo as when presented with a link opp, you simply can't pass the option to have a keyword rather than just the URL etc :)
yeah - we are seeing the same.
All our client sites survived OK - we take care with article spam, all keywords are rotated and a realistic mix is used.
OTOH - one of our joint venture sites got whacked - down from #3 to #57 (approx), then climbed back up to page 3 after a few days.
On that one we got lazy and tried a load of spammy links that were sitewide, non-relevant to the site content, and too heavy on the main keyword.
I will be have a go at reviving it by buffering some links in through relevant on-topic feeder sites - and doing some social stuff on it.
Thanks guys it's great to be back here and I'm hoping with more valuable info than just this ;-)
I just bit the bullet and paid for a full membership of linkresearchtools and it hurt but I don't regret it. I've been running reports on sites in my client's industry that are doing better, and sites that tanked. The picture looks different now.
I'm almost certainly we ARE talking about an anchor over-optimization penalty. One of my client's sites lost it's long tail and went to the 2nd page for it's main keyword (which is also it's name). Looking at the anchor density, it's quite high. I think if you go over 20% you trigger an alarm.
Now here's where I think it gets tricky. I did find some sites that are doing better after penguin that have even higher anchor density than 20%. The difference between their site, and my client's site, is that, upon human review, their anchors look more natural.
My hypothesis is that when the anchor density for a money term is too high, it triggers an alarm that gets a human to review at their web spam center in India. Then the human determines if those links are natural or not. The data I was seeing was not discernable to a computer but easy for a human to determine what looks real and what doesn't.
If it's just an Anchor over-optimization penalty, we should be able to recover by swtiching around the anchors on all the paid links that we have, while simultaenously building social media / web 2.0 links.
Of course the problem is doing this in a scaleable manner. If anybody knows of a service that offers an affordable/scaleable social media service, please let me know. In the meantime I am trying to build it myself.
Other metrics that I used was deep link ratio, percentage of total anchors by total links, percentage of links by different categories, etc. The only commonaliteis I found was anchor text density and anchor text profile.
Hopefully it's just an algorithmic change and the sites will improve as we switch around the anchor density.
My days of making free money are over, now I'm actually having to work for it ;-)
More of a natural looking link profile overall,
-- I totally agree! Sites that have natural looking profiles to the human eye are the ones that are ranking well.
I'm thinking of the social graph as a bit like links used to be
-- Agree to this too. In the past blogs didn't exist and social media didn't exist so links were votes. Now everyone has a social media footprint so I think they are shifting the emphasis from links as votes to likes/mentions/blogs as votes. That's where/how they are incorporating the social media markers. Perhaps at this time they are not turned up so high, but I think in the future it'll become more important. This is where authorrank will prbably become important too.
That's definitely possible at the moment but all of those social links which are being mapped are really hard to replicate in a high enough volume network to be useful, moving forward, imho
-- I know!! Whoever figures out an inexpensive and scaleable way to do this will make a lot of money!
EMD + 1 page of content...
-- What's EMD?
It's about trying to fake a natural link profile - which is hard for an seo as when presented with a link opp, you simply can't pass the option to have a keyword rather than just the URL
-- Yeah it hurts to pay for a link and have hte anchor say "Click here". Ouch! But I think it's essential now.
according to this this article, it's not just link over optimisation but it's also the relevance of the content surrounding the link
http://www.micrositemasters.com/blog/penguin-analysis-seo-isnt-dead-but-you-need-to-act-smarter-and-5-easy-ways-to-do-so/
TBH it makes sense really.
Before hand people were slapping links all over the pace with the anchor text of their chosen phrase they wanted to rank for.
People were also stuffing their links on pages so diluted in relevance even humans couldn't work out what the site was about.
From G's perspective this is sign of link spamming and something their algos can easily detect now.
I think what we will see more of now is dedicated themed link sources e.g. motor, law, travel etc which will allow a page relevant to rank for a phrase that the on page SEO is optimised for not for what the link anchor text says.
> relevance of the content surrounding the link
Yep I reckon this has been turned up. One dinged site is still ranking fine for terms containing a certain word. The ODP listing for the site goes something like:
BRAND - UK *Word* blah blah blah
So there's plenty of non-anchor links closely followed by that word spread around.
http://insidesearch.blogspot.com/2012/05/search-quality-highlights-53-changes.html
"Anchors bug fix. [launch codename "Organochloride", project codename "Anchors"] This change fixed a bug related to our handling of anchors."
The data on this one is pretty confusing. As much as I'd like to think it was just an anchor text over-optimization problem, I keep finding exceptions to this. Check out this site:
chicago bankruptcy attorney 604 42.4%
chicago bankruptcy 476 33.5%
bankruptcy chicago 70 4.9%
[img] chicago bankruptcy attorney 43 3%
[img] chicago bankruptcy 37 2.6%
naperville bankruptcy attorney 20 1.4%
The site ranks #1 for chicago bankruptcy attorney.
The quality of hte links sucks too. Here are some links as an example:
http://1-800badcredit.com/Our_Partners-Business.html
http://1-800badcredit.com/Our_Partners-Legal_Services.html
http://1business-insurance.co.za/links-l0-0.html
http://1st-coating.com/
http://1st-coating.com/index.htm
http://1st-coating.com/link_partners.htm
http://22centurytech.com/business_and_economy/law?p=11
http://312.net/bankruptcy_lawyers.htm
http://3dinternships.com/component/content/article.html?id=2
http://3draf.com/links-l0-7.html
http://5starstour.com/links/bussiness_marketing
http://5starstour.com/links/bussiness_marketing/page-1.html?s=A
http://5starstour.com/links/bussiness_marketing/page-1.html?s=H
http://5starstour.com/links/bussiness_marketing/page-1.html?s=P
http://5startravelcenter.com/resources/resources_8_-_other_useful_links_webmaster_services_2.html
http://6neat.net/links-l0-1.html
http://815.net/bankruptcy_lawyers.htm
http://aastudentguide.com/sites.php
http://ac21portability.com/userinfo.php?uid=2912
http://accurate-miniatures.com.hypestat.com/
http://adatmentes.info.hu/link5.html
http://adventureindochina.com.vn/index.php?option=com_content&view=article&id=915:other-links-exchange-2&catid=296:links-exchange&Itemid=463
http://adventureindochina.com.vn/index.php?option=com_content&view=section&layout=blog&id=36&Itemid=460&limitstart=15
http://advertising-in-india.com/legal/chicago-bankruptcy-attorney.htm
http://affichescinema.com/partenaires.html
http://alaska-fly-fishing.net/resources/attorneys.asp
http://alaskatrophyadventures.com/resources/attorneys.asp
http://alaskawestair.com/resources/page3.html
http://amiralpalacehotel.com/links/link7.php
http://ammarg.com.np/links.php
http://analysis.heatkeys.com/Attorney+Chicago+Bankruptcy
http://angkorcambodiaguide.com/?opt=6&hlineid=68&langid=18
http://angkortransports.com/?opt=1&hlineid=68&langid=18
http://angkortransports.com/index.php?opt=1&hlineid=68&langid=18
http://antonioesquinca.com.hypestat.com/
http://anuvids.com.hypestat.com/
http://aquapur-wasserbetten.com/matratze-stinkt.htm
http://aquariumhomecare.com/modules/profile/userinfo.php?uid=5412
http://armandstrunks.net/link12.html
http://art-dialer.com/en/categ-Bankruptcy_Law-233_5.html
http://art-dialer.com/en/search-Affordable-3.html
http://art-dialer.com/en/search-Bad_Cars-2.html
http://art-dialer.com/en/search-Bankruptcy-3.html
http://art-dialer.com/en/search-Bankruptcy_Dupage_Illinois_Lawyer.html
http://art-dialer.com/en/search-Bankruptcy_Lawyer_Illinois.html
http://art-dialer.com/en/search-Bankruptcy_Lawyer_In_Chicago.html
http://art-dialer.com/en/search-Bills.html
http://art-dialer.com/en/search-Cars-2.html
http://art-dialer.com/en/search-Chicagoland.html
http://art-dialer.com/en/search-Families.html
http://art-dialer.com/en/search-Illinois_Bankruptcy_Lawyer_Attorney.html
http://art-dialer.com/en/search-Ny_Bankruptcy_Lawyer-2.html
http://art-dialer.com/en/search-Start_up_New_Businesses.html
http://articlenexus.com/Article/Fishing-in-Alaska/836641
http://arzopark.com.hypestat.com/
http://arzosoft.com.hypestat.com/
http://aseantravelandtours.com/links/index.php?c=1
http://aseantravelandtours.com/links/index.php?c=1&p=21
http://aseantravelandtours.com/links/index.php?c=1&p=41
http://aseantravelandtours.com/links/index.php?s=A&c=1&p=1
http://aseantravelandtours.com/links/index.php?s=H&c=1&p=1
http://asia-ceo.org/links.htm
http://asiasolarpanels.com/resources/helper.html
http://aspaceplacestorage.com/ourfavoritepartner1b.html
http://attorney-america.com/db/Illinois/Bankruptcy%20--%20Debtors/All
http://avsnewsonline.com.hypestat.com/
http://awla.org.hypestat.com/
http://baggernation.com.hypestat.com/
http://bahiasantander.com/enlaces/enlaces.htm
http://bankruptcy-pages.com/main.php?keywords=bankruptcy%20lawyer%20chicago
http://bankruptcy-pages.com/main.php?keywords=bankruptcy%20lawyer%20chicago
http://bankruptcy-services.regionaldirectory.us/illinois.htm
http://bankruptcy.deans-knowledgebase.com/Links/Small_Business_Bankruptcy_Page_1.php
http://bankruptcyattorneydir.com/chicago-bankruptcy-attorney.html
http://bankruptcyattorneyfortworth.com/other_bankruptcy_resources.php
http://bankruptcydistrictcourt.com/illinois-bankruptcy-lawyers/
http://bearcreekbears.com.hypestat.com/
http://beckspeedster.com.whohost.me/
http://bestbusinesssearch.info/index.php?s=A&c=44&p=7
http://bestbusinesssearch.info/index.php?s=A&c=44&p=7
http://bestcase.com/bkattys_eh.htm
http://bestcase.com/bkattysg_flky.htm
http://bestofamericabyhorseback.com.hypestat.com/
http://bigbaddivas.com.hypestat.com/
http://biowatchmed.net/business_and_economy/law?s=A&p=481
http://biowatchmed.net/business_and_economy/lawpage-25.html?s=A&p=481
http://blackbirdhotel.com/links/link5.php
http://blanknyc.com.hypestat.com/
http://blanknyc.com.whohost.me/
http://blog.dein-tor.com/roseroth/feed/
http://blog.samuraibeatradio.com/2010/06/09/hideo-kojima-event-at-uniqlo-on-june-11.aspx?ref=rss
http://blogaboutmoneyonline.blogspot.se/2008/05/monetize-your-blog-with-oiopublisher.html
http://blogs.ugidotnet.org/Allino/archive/2006/12/28/63423.aspx
http://boatsshopping.com/
I see link exchanges, blog commenting spam, paid links. Most of the links are non-contextual and non-relevant.
The only way I can understand this, is a manual tweak? With sites that are still doing well with this sort of spam simply not having been manually penalized?
I checked to see if maybe there was a social media component to this, but they don't seem to have a presence on web 2.0 or social media sites.
Any other thoughts/experiences/theories?
If any one has some sites you'd like me to run so we can compare data let me know.
Marcela, how old are the links that link to that site? As in, how long has that site had those particular sucky links? If those links have been linking to that site for a long time, as in years, that might account for that site being dumped in Penguin.
I have seen that before...I've also seen that sites that spend a lot of money with adwords get preferencial treatment.
Maybe there's thresholds before certain algos kick in.
From a processing point of view it would keep costs down and increase speed (e.g. panda is ran manually probably because of this), while still keeping competitive niches under control.
It would also explain why some sites like above defy the logic being applied.
Looking at it from a pure quantity of links POV is probably wrong and we need to look at it from a PR perspective. From their 100% of PR what % is keyworded compared to branded, what % are on semantically relevant pages etc
So far as I can tell, the main target of Penguin has been to penalise use of spun links where these dominate a site's link profile.
A URL with minimal spun links as little affected, but where spun links dominate, keyword rankings crash.
I say URL and not domain because I've tested this with subfolders and these can be hit much easier than a homepage.
The scary thing is that normally Google simply devalue what they perceive as low quality links. Now it appears that low quality links carry an active penalty.
From what I've seen so far, it should be pretty easy to take out general SME websites, or at the very least major categories on them, with nothing more complex than a $40 pligg blast.
Pandora's box has opened.
I'm not seeing a correlation to spun content links myself.
Happened across this article while looking at something else:
http://www.micrositemasters.com/blog/penguin-analysis-seo-isnt-dead-but-you-need-to-act-smarter-and-5-easy-ways-to-do-so/
Interesting data from a rank tracking site/company. (never heard of them before)
Basically says money keys at 50% or higher in your anchors means penalty.
They also note that non-relevant links draw a penalty. Looking at the graphs it looks like a site with 10% or less links from relevant sites likely got penalized. (Graph starts rising at 30% and shoots up at 10%.)
There are always exceptions, but it looks like a site with both high density on ranking anchors and low relevant links is a likely target. This seems to line up with what I'm seeing. Might play into the spun content links idea that the links are on non-relevant sites instead of due to being on spun content.
yeah - its not spun - otherwise I'd be dead in the water.
Only one site hit, and it fits the description in Drastic's post - high percentage of same anchor links, and high percentage of irrelevant sites
Not sure whether it is relevant 'sites' or relevant 'pages' - as my example site fails on both counts.
>Not sure whether it is relevant 'sites' or relevant 'pages' - as my example site fails on both counts.
That's what I've been wondering. Usually spun content is on topic though, right?
so basically, penguin has just done something which google claim they've been doing since about 2007?
I know we all know they weren't, but the scary thing, if that's the case, is that I didn't reallise quite how far behind the FUD they actually were. I wonder how much easier I could make my life now and still be safe?!
> Usually spun content is on topic though, right?
Not necessarily if the spinning is automated rather than done properly. Or outsourced and not checked come to that.
I would have thought it has to be relevant pages rather than sites, or links from major news sites or wide-ranging tech blogs etc would be useless or even harmful?
Yeah, page level makes more sense, at least in some aspects. Maybe we should be looking at context level as well?
[added]
The relevance issue could by a byproduct of just lack of anchor text variance. Sites with more natural links are going to have variance and are going to have more links from on topic urls. SEO links are more likely to have little variance and less likely to be on topic in general.
I'm also thinking that rather than penalize, they are simply devaluing these links causing rank to drop.
>> Usually spun content is on topic though, right?
>
>Not necessarily if the spinning is automated rather than done properly. Or outsourced
Relevance is many things - what about links between languages? Eg. links from Italian to French? I wonder how much translation is going on in calculating relevance and topics.
I really think that the site on which the link is doesn't matter if it is on topic or not. I think it has to do with the age of the link (how long that links has been pointing to the site).
I'd show you an example, but this conversation is on the outside...
You could start a thread in the IC if you want...
:(
ic :'(
I'm not sure about spun either because a friend of mine is selling a "get out of penguin" service for $150 which is basically only spun content and so far it's working. It's all the nasty stuff but with pretty good quality stun content and directories.
BUT... If Penguin is the same as Panda as in it's an offline algo run once per month it will be VERY interesting to see what happens to these when it runs again. And if they do drop would it work again and his client would be locked into paying him every month to get out of the sandbox every month.
Quote from: Drastic on May 08, 2012, 04:31:51 PM
I'm not seeing a correlation to spun content links myself.
It's probably because the only areas I see an effect is where spun content is involved, though this is more social media links than articles.
Quote from: Drastic on May 08, 2012, 06:50:43 PM
I'm also thinking that rather than penalize, they are simply devaluing these links causing rank to drop.
I'm seeing a negative effect - presumably based on link profile. Positions gained through normal links have disappeared.
However, I only have a small set of sites I monitor, so my base is quite limited.
Seeing a few things that correlate with the anchor text observation in the micrositemasters post, but only contradictions to the relevance suggestion. Not sure their analysis was all that strong on the relevance factor.
I find it hard to think they will be acting based on relevance unless it is heavily modified based on other 'fuzzy logic' factors.
Then there is the question of how they define relevance... could be :
* site relevance (seems unlikely to me)
* page relevance (more likely perhaps)
* surrounding text relevance (eg. paragraph or a specific number of words around the link)
* other relevance factors - incoming links, page title etc
... or even some hybrid combination of the above.
Not suggesting it is, just that those factors need to be at least considered.
Having said that, past experience suggests it is less complicated than we give them credit for.
Quote from: 4Eyes on May 09, 2012, 11:33:35 AM
I find it hard to think they will be acting based on relevance unless it is heavily modified based on other 'fuzzy logic' factors.
Then there is the question of how they define relevance... could be :
* site relevance (seems unlikely to me)
* page relevance (more likely perhaps)
* surrounding text relevance (eg. paragraph or a specific number of words around the link)
* other relevance factors - incoming links, page title etc
... or even some hybrid combination of the above.
I think the page level is the easiest. Are the page titles relevant, is URL keyworded, are the h1etc tags related etc This would allow big generic sites to still have good links, eg a directory.
Yea have to agree with hungrygoose.
Page level makes much more sense and will allow sites totally none relevant to pass a "vote" meaningfully.
E.g. Bill Gate talking about a restaurant he visited on his personal blog
Quote from: Adam C on May 09, 2012, 09:36:05 AM
Seeing a few things that correlate with the anchor text observation in the micrositemasters post, but only contradictions to the relevance suggestion. Not sure their analysis was all that strong on the relevance factor.
Agree, and the more I think about it, the more I think it's only anchor text based and the relevance bit is just a side effect.
Drastic but if it was just anchor, then how come examples like the one above didn't get nuked?
I have a couple of sites that lost their long tails all together. The commonalities:
- Relatively high anchor density
- No social profile
- Obvious paid link profile
The issue I'm having is in identifiying commonalities, for every theory I come up with, I find a handful of sites that defy it.
I am beginning to think it is a matter of ratios.
For example, it is ok to have high anchor density if you also have a social profile and a percentage of high quality links.
Or, you can have a large amount of low quality links if your anchor density is low and you have a social presence.
I may be totally wrong but in the past I could almost always find theories that stood up to the test of various sites, but not this time.
I don't think it's a specific "hit 42.5% of density and remove anchor value," but a bit more complex where something like a few image links, site name and url anchors may keep you out of hot water when you are pushing over the limit.
I could be wrong, but that's what it looks like from here.
My experience based on sites that were actually useful, aged, PR 3+, and in most cases loved/liked by users has been-
The sites where we pushed one or two terms via forum profiles and blog comments (which really did work for my test group in days gone by BTW), have really been hit. The main anchor text terms have been removed from the G. However, after a week or so, the secondary terms for deep pages have gone up (perhaps to fill the space vacated by the other profile/comment hoggers in those terms? Maybe because they were actually really relevant pages) This should be no surprise since the holy G warned us of this.
There have been two other upsides I have found, one which I wont talk about in public here (IC), but the other is that by adding bing webmaster tools it has increased the ranking in both Bing and Yahoo. That is the single change I made to two sites, I didn't remove G's webmaster tools, just added bing's. This did not work previously (last test April 2011). Might just be a fluke, but I'll test on a few more sites. Interesting that I have Bing/Yahoo sending traffic which then clicks on G ads. I might run a split test on G/Adcenter/?? to see which brings in the money. Certainly in the phone apps tests I have been running so far there is little difference between Android and Windows Phone in terms of income (I guess it is a big fish/small fish situation).
Anyway, I am rambling, Massive anchor text spam is what Penguin has killed for me (and I did go to town!). Onsite spun pages are ranking better as a result (IMO).
On a side note, the G seem to have a swell of bad blood directed towards them from people who have been fans up until now. These are people who make money for adsense, in the most part. It will be interesting to see where it goes. There might be some fun left in the web yet - Someone press the reset button!
I'm thinking maybe this guy hit the nail on the head. What do you guys think?
http://www.seomoz.org/ugc/the-penguin-update-how-google-identifies-spam
Has anyone done or know of an analysis of the % of nofollow links?? I'd like to see that spread with the anchor text spread.
Overall, I don't strongly disagree with this analysis.
https://plus.google.com/u/1/111294201325870406922/posts/MjLCYnwMdBB
Quote from: 2much on May 07, 2012, 04:31:12 AM
The site ranks #1 for chicago bankruptcy attorney.
For me the site ranks 1st for their places page not their actual website in the naturals.
That's unless I'm seeing the wrong site.
I had a site get killed then it popped back up.
I thought Ras even mentioned this? There were issues that they fixed? Which means its going to be tremendously difficult to troubleshoot this one...
Some sites got popped and came back, but many didn't. A few of mine included. A few other sites I watch too also got dinged.
Perhaps I'm wrong here but this algo seems like a pretty monumental shift to me, most of the SERP's I watch changed.
I'm starting to believe that it's not as easy as before to figure it out exactly. I remember when all we had to do was look at the age a domain was registered and immediately we could see the trend and started getting links on aged sites.
I'm thinking Penguin is a combination of human reviewers, anchor density, analytics metrics, and general quality of the backlinks. This article gives some decent ideas on how to visualize the different ratios:
http://searchenginewatch.com/article/2173658/Google-Penguin-Unnatural-Links-How-to-Protect-Your-Site-Moving-Forward
I'm starting to get a system down now. Lots of links in articles going out to guest blogs. Lots of contextual links. Instead of paying an intern to write the articles I'm paying the intern to manage it and buying articles from textbrokers. It's starting to come together.
Yup 2much that's working but also add some twiddling of thumbs for 13 - 40 days for the link juice to flow fully. ;)
Quote from: dogboy on May 11, 2012, 06:21:06 PM
I had a site get killed then it popped back up.
I thought Ras even mentioned this? There were issues that they fixed? Which means its going to be tremendously difficult to troubleshoot this one...
Google cocked-up in the first place, so it's natural to expect they will be slowly rolling back the dial for weeks to come.
This all looks very much like we used to see happen between the end of 2003 with Florida, and other major updates for a year after, such as Allegra (easter 2004?).
Point being, biggest update I've seen for years, unsettled and irrelevant results on a level I haven't seen for years, biggest c##k-up I've seen for years. So back to the old Google style of keeping quiet, pretending everything's normal, while rolling back the dial.
Am expecting to see SERP's keeping changing for weeks and more sites pop back in - lesser "offenders".
ADDED:
Quote
http://www.seomoz.org/ugc/the-penguin-update-how-google-identifies-spam
We've seen Gmail blamed for previous updates, but I think this guy is just floating speculation. There's no suggestion that he's seen anything, and is just throwing an idea into the wild for discussion. Personally, I didn't read anything in there that struck as useful or relevant.
Soo many wp sites with 2-3 spun posts and the default themes ranking now. Seems to be getting worse!
lol, probably goes to show how many people were actually gaming Google's system.
It will get back to normal once the masses adapt to the new ways and the insane negativeness turned down
BTW here's a result for sh*ts and giggles - http://www.google.com/#hl=en&q=redding+motorcycle+lawyer
F*** Me!
I like the disclaimer at the bottom...
QuoteIn response to a complaint we received under the US Digital Millennium Copyright Act, we have removed 1 result(s) from this page. If you wish, you may read the DMCA complaint that caused the removal(s) at ChillingEffects.org.
...never seen that one before.
I see lots and lots and lots of them.
It's standard Google fair to link to them at ChillingEffects when there is a DMCA removal.
I'ver always thought that as the DMCA sender you could.
DMCA Google about 3rd party site breeching rights
Google would then send the notice to Chilling Effects
Chilling Effects will put the notice online
Google will link to it
Which means that Google are breeching my copyrighted letter to them and passing it to a 3rd party who is making it public, whereupon Google link to it.
so....
DMCA Chilling Effects
and Google for linking to it....
leading to a very large DMCA circle jerk....
Seeing a new tweak from yesterday.
Some winners and some losers.
Hum how interesting. http://www.seroundtable.com/google-penguin-check-15165.html
Query -amazon.com
Seems to return results very similar to ore penguin results.
yup, definitely interesting, as normal the comments are better than the article ::) Does the -something.com string actually kick the search to a different database, does anyone know? Sounds a lot like it.
If so, and since its almost a month since Penguin, isn't it a bit strange they've not synced everything up yet? Perhaps they're expecting more adjustments
Quote from: Drastic on May 17, 2012, 02:48:21 PM
Seeing a new tweak from yesterday.
Some winners and some losers.
I've got some losers. Stuff that most people here would consider to be pretty innocent too. Nothing worse than low quality directories on the backlink front and reasonable other stuff to balance them.
I didn't see any changes my end.... probably cause a lot of my stuff has already dived :(
Idly checking GA today for my main sites, and found one dropped from 1100's to 700's daily uniques after Penguin - though none of the others affected.
Ironically, it's probably my most authorative site and not one I do link building for - an interfaith site I run for personal interest, not a money site.
I find that result somewhat bizarre, considering the anti-spam focus being touted. Maybe it has too many universities and government sites using the same anchor text. :P
Maybe it's internal links penguin looks at also?
Standard Wordpress and vbulletin install, just like my other sites. Generally same site structure and nav links on each.
This was a decent write up on Penguin...might be good to read, but much of what it says you all said...
http://www.ppcassociates.com/blog/experience/when-penguins-attack/
Have you guys seen this?
http://www.seonitro.com/easyblog/entry/pandapenguin-case-study-analysis
Pretty impressive and summarizes everything I've been seeing too.
I like that 2much, its a good writeup and pretty much says what I've been thinking especially about brands :)
Interesting link - so if I'm reading this right, then links that disappear in volume are working against a link profile? In which case, would the removal of low quality links not compound the issue?
Good read, I'm glad to see someone else thinking like I am about links being counted while not indexed.
>In which case, would the removal of low quality links not compound the issue?
Correct.
Quote
>In which case, would the removal of low quality links not compound the issue?
I remember some time ago someone on this board telling me that was a negative SEO trick. Put up 27,000 links, make sure they are all indexed by Google, remove them all.
> Put up * links, make sure they are all indexed by Google, remove them all.
Negative link growth is a funny thing
I was wondering if that would catch your attention Jason. ;D
Quote from: Drastic on May 24, 2012, 02:37:44 PM
>In which case, would the removal of low quality links not compound the issue?
Correct.
Very, very, interesting. I was going through the spammy links and found a big chunk of them already gone, through direct removal and site churn.
I had presumed the spun nature of the links was responsible for the Penguin hit, as we've seen semantic processing covered with Pandora, and Matt Cutt's highlighted spun text.
I was even wrestling with the idea of maybe even trying to get someone to DDoS the remaining junk sites!
But sounds like, instead, the best way to beat Penguin is to churn out stable links fast?
well yeah, but I was having this exact conversation with someone the other day and he was absolutely sure that his switching off of a link network had recovered one of this sites from penguin. I was wondering if switching off effectively just reconfirmed the penalty, but apparently not.
I'd like to see someone test whether you can use link network we know google are penalising for to run the link removal trick still? I have been wondering, especially considering the volume of stuff which still seems to be ranking off the back of crap inbounds, whether Penguin actually only acted on one or two link networks, but the panic induced is letting G identify others as people pull links? in which case we might be in a rather surreal scenario in which some links hurt you going in, and some don't until you pull them out.
A few months passed from Penguin. We should have some tests and results.
Since our sites that were hit were money sites we were very careful in testing. We lowered the on-page optimization, build some new "brand" anchor links and removed a few sidewide links.
The Googlers seem to be fairly confident that they didn't open widely the doors to "outside your control factors" hurting your site. One way of doing that is by combining/confirming outside data with data that is under your control (like on-page optimization and internal link structure).
Have someone been able to recover from Penguin?
Hi BST, not me. Panda, Penguin whatever.
Have you? Have the changes you made made any difference?
Income has recovered because I built new sites and spammed them to the top :)
It's said that it's an offline algo so they only run it once in a while and upload it to their main servers, so I'm not sure any real recoveries can be seen until panda 3 comes.
Hi Rupert,
There is some strange stuff going on with the site I experiment but no real recovery.
Most of the sites do really well in Bing/Yahoo and it seems stupid to let the bird in your hand and go chase another one.
Tomorrow I'll make another batch of changes to more sites.
Hi hungrygoose,
What methods are currently working for you (feel free to go deeper)?