Panda key algo changes summarized

[pontifex]

Folks, I have been reading a lot, thinking a lot and analyzing a lot. I am still not sure, how to get the US traffic back to pre-24th of February levels! But I think it is time to summarize the key theories of the algo change in the US:

- Internal links devalued, only external count really

- Thin pages cause substantial bigger problems for a domain

- Duplicate content snippets on your page cause substantial bigger problems

- Too many external named links "widget keyword" instead of "more..." (eg) cause penalties


are what kept me working in the past 4 weeks. Do you have some additional meme?

P!

[tedster]

Welcome to the forums, IMangel.

I don't see how either of those plug-ins would create a Panda problem. StickPost just gives your home page some static content, right? That might even help. And as long as you don't mash-up the post order too often, PostMash shouldn't create any Panda issues either - at least as far as we've identified them for now.

I also can't see how sharing on Facebook Walls would work against you. And if you ever want to turn comments back on, there are plug-ins like Akismet as well as challenges to stop automated spam that can cut down comment spam dramatically.

Are you concerned because you have already seen a traffic drop that corresponds to the two Panda roll-out dates? Or are you trying to pro-actively prevent trouble before it shows up?

[IMangel]

Hi tedster, thanks for the speedy response. My blog traffic actually took a dip around October last year (2010) and has recovered slightly. Prior to the drop in traffic, I was ranking #1 for two major keyword phrases for about 5 years.

After October 2010 and Panda, one of the keywords phrases is stuck at #2. Interestingly, I can also be found at #3 and #4. In spite of these multiple listings, I am experiencing a 25% drop in monthly traffic. There's still something to be said for a #1 position in Google. What baffles me is that the current #1 site has less authority content-wise and only 25% of the number of back links that my site has.

I've decided to disable the Stick Post plug-in to see if that helps.

Also, I already use a Wordpress anti-spam plug-in but am reluctant to turn the blog comments back on. They mostly seem to be from service providers and vendors in that industry and tend to be very promotional, adding no value to the conversation.

I see the Facebook sharing and liking plug-ins as an alternative to the blog comments and many of my Wordpress pages are widely shared and liked on FB.

[2macarena]

SEOPTI: Yeah they are cruel. They smacked my $13K exact match domain with -50 and it is still in da box. It was an affiliate redirect issue for sure so I fixed it and filed reconsideration. They did respond saying "you still violate our quality guidelines". So, hmmm If I do, it is because of backlinks then. Cutts said that manual penalties are like "time outs" And I think the maximum duration is 365 days. So, the domain should be back soon. Anyway, my point is that you gotta be careful with backlinks and interlinking strategies, especially for web1 and especially for purchased or new domains and exact match domains. But then the question is "how to be carefull". And it is a big one that I know some about but do not know even more about. In my view, while many people here consider mainly on-site variables for their Panda slaps, I believe the off-site issues are as likely. I am closely following one of my competitiors and he knows what he's doing for sure. So, recently he has added twitter "tweet" and facebook "like" buttons. Oh my its too long a post already. Thanks for your time!

[brinked]

I think I may be on to something in regards to panda 2. I took on 2 new clients that were hit by panda 2. Both were effected, but only a few pages.

In studying these pages I took a look at there google preview screenshots, in an effort to really see how google is seeing them. I can see google can see where text is placed on a page, this is evidenced by the fact that they highlight the text in the screenshot that you searched for. In both cases, all pages that were effected, all pages had text being pushed down....but heres the kicker, it wasnt always ads that are pushing the content down. 1 case theres listings pushing the content toward the bottom of the page and the other client had a flash widget pushing the content down to the bottom of the page. This content was not the actual main content but google was probably viewing it as that and viewing it as keyword stuffing since the content was not in a prime area on the page.

[crobb305]

2macarena

Is the -50 penalty a separate issue for you or do you also think you were affected by Panda?

[SEOPTI]

Thanks for this report 2macarena.

Now I know the -50 is a "multi purpose penalty". They do it in the same way with Panda. Different "violations" receive exactly the same penalty/treatment. It all becomes clear.

[2macarena]

Yeah -50 hit one of my domains and it happened way nefore panda update. As to Panda, I experienced slight drops in my EZA articles but this is because Panda slapped EZA big time in general. My Web1 domains actually gained juice after Panda. I cannot give any advice about Panda because I am as oblivious as many people are. I just got a few pieaces of the puzzle I guess that I am split testing right now.

[heisje]

I am being UNPANDALIZED - many of my pandalized sites are being gradually UNPANDALIZED!

Is it just me, or is this a more general phenomenon?


.

[pontifex]

Another thing I see the longer I look: Google is MUCH faster than it used to be pre-23rd of Feb. New domains show backlinks in matter of days. Link discovery got so much quicker that I think it means something... but what?

[Shatner]

>>>I am being UNPANDALIZED - many of my pandalized sites are being gradually UNPANDALIZED!

How much recovery are you seeing? How many pages is your site? When did this start? What changes have you made?

[heisje]

250 commercial sites : no changes - took a wait & see attitude. I see a much increased spidering activity by Google.

.

[c41lum]

Im not seeing any recovery from my sites....has anybody seen any recovery from the UK or is it just US sites.

[agent_x]

Newbie 1st post.

In reply to brinked, this prompts me to mention something I've observed.

One of my sites that was hit by Panda 2 I decided to change radically as it was what you would definitely call a thin affiliate site. I thickened it up by adding more pages with original content. After 10 days this had no effect, so I decided to change the way it looked too. This involved changes to the CSS and then a couple of days later, some code changes and a different logo graphic.

After that, when looking at the search results and viewing the preview thumbnail, I saw that Google had immediately indexed the code changes and the new logo, but NOT the CSS changes from two days before. Today, three days later, it still has not downloaded the new CSS style sheet.

If CSS has become important in Panda's algorithm, this might be why making changes to a site doesn't seem to bring any better results for a while. Does anyone know how long, or at what intervals, Google fetches style sheets?

[enigma1]

After that, when looking at the search results and viewing the preview thumbnail, I saw that Google had immediately indexed the code changes and the new logo, but NOT the CSS changes from two days before.

Hi agent_x welcome to the forums,

why google would fetch the css for its index? It fetches the html part that has the references to the side scripts like css, js etc. And so when you view the content with your browser the later pulls-in the css from the site or any other resource that is listed in html head section.

It's good practice to setup the html so the fully qualified paths to the various resources is used. I am saying this in case your html base tag or some path reference is not resolved correctly and viewing the google's cache or preview etc, doesn't render parts of the page (after the recent mods you made)

[agent_x]

If there was a problem with the way the url was resolved then the preview would never have looked right in the first place. So you're saying it's the browser that downloads the CSS even if the user never visits? What would be a reliable way of testing if that's the case?

[patc]

I would be surprised if Google didn't crawl CSS because of understanding the layout of a site properly, and therefore apply any quality metrics against it. Plus in the site preview (in the SERPs) my site shows it rendered correctly, which must mean they have a copy of the CSS server side.

[enigma1]

So you're saying it's the browser that downloads the CSS even if the user never visits? What would be a reliable way of testing if that's the case?

With the cache yes changing the stylesheet you should noticed it immediately, with the preview they take screenshots but when whatever browser is in use for rendering it will use the side scripts.

If you check the cached version does the page renders correctly?

PS: I would think is more complicated to store the various stylesheets. Instead they forward the request to a browser like system it takes a screenshot of the page.

[agent_x]

patc, that's what I think - I think they must have the technology to render the page to determine what elements are above the fold and what are below and so on, and I think they also use this technology to render the preview thumbnails you see on the search page, which are downloaded post-load. But if someone knows for definite that is not what they do I'd be glad to be corrected.

[agent_x]

enigma1, yes the cahed version renders correctly.

[ciol]

not only this, it has caused reduction in number of Google crawled pages of websites in US and outside US as well.

[Bewenched]

@2Macarena

They did respond saying "you still violate our quality guidelines"

I've never ever gotten a response from them when I filed a reconsideration.

[crobb305]

I am being UNPANDALIZED - many of my pandalized sites are being gradually UNPANDALIZED!

Is it just me, or is this a more general phenomenon?

A lot of people are reporting improvements in the past 24 hours, especially in this thread [webmasterworld.com...] . Yesterday I reported a 45% improvement. Today that number is up to 60% improvement. So as of today, I have recovered about 60% of my traffic. It is NOT full recovery as of yet. Rather, it is a very phrase-specific recovery. I am back to number 1 with sitelinks for key phrases; but, modifying the phrase to put the words in reverse order often throws me down to page 3.

Monday, my recovery was 10%. Yesterday 45%. Today 60%. This is definitely an upward trend, as more phrases pop in. But until I see those key phrases ranking in reverse order, I won't declare victory. We have seen traffic surges prior to a spanking, so I don't know if it will stick.

I will also say that my bounce rate is at its lowest point in 9 months. It would seem that the quality of traffic has improved. It may be due to the fact that I have not recovered my trophy phrases (not in the top 100), but those have never converted for me.

[walkman]

I am being UNPANDALIZED - many of my pandalized sites are being gradually UNPANDALIZED!

Is it just me, or is this a more general phenomenon?

250 commercial sites : no changes - took a wait & see attitude. I see a much increased spidering activity by Google.

Moved up or what?

[heisje]

Moved up or what?

Badly hit by Panda, gradually recovering traffic levels, such as prior to Panda, but not a full recovery as yet. Also, some previously "ignored" sites are increasingly getting "healthy" traffic.

Total traffic was at 40% of pre-Panda level, (30% during Easter) - now stands at 60%. This is a 50% increase within just 2 - 3 days, disregarding low Easter levels. Trend seems to hold at present.

.

[Bewenched]

I've seen a bit of a traffic increase since Easter sunday. I have made site modifications to make the site better, but nothing drastic, it was mostly benefits for my customers.

[crobb305]

Total traffic was at 40% of pre-Panda level, (30% during Easter) - now stands at 60%. This is a 50% increase within just 2 - 3 days, disregarding low Easter levels. Trend seems to hold at present.

heisje,

It sounds like you and I are seeing the exact same recovery rates. Are you observing phrase-specific improvements, with sensitivity to wording (i.e., reversing the order of the words will change your ranking drastically)?

I know we should probably be talking about this in the Update thread. I'm just wondering if there is a phrase-based connection from Pandalization and restoration. I am also seeing a gradual update to my site: search (gradually displaying new title tags that I added last week). It could be that ranking is being restored as the internal pages get updated cache. In other words, maybe this is the first update that is allowing for real ranking improvements based on our quality modifications to internal pages, and we are seeing a gradual traffic increase each day as those internal changes/quality scores/etc get calculated.

[dibbern2]

I'm seeing Panda effects split across two very different groups of searches. Primary terms, very short 1 or 2 words, got hit hardest. This is as should be expected.

Long tails, even just one more word added to a primary term, hardly moved at all. I'm a little surprised at the apparent lack-of-connection between deranking a page for "blue widgets" and keeping it the same, or even boosting it, for "blue widgets sales"

[walkman]

Long tails, even just one more word added to a primary term, hardly moved at all. I'm a little surprised at the apparent lack-of-connection between deranking a page for "blue widgets" and keeping it the same, or even boosting it, for "blue widgets sales"

Guessing: probably Goog sees it as a second tier site. Of course, Goog has screwed this one up so it's work in progress.

I see a lot more referrals from Goog internationals and branded sites (like G powered ISP portals) today

[2macarena]

Bewenched: I was surprised too. Here is exact message:

"Dear site owner or webmaster of http://www.example.com/, We received a request from a site owner to reconsider http://www.example.com/ for compliance with Google's Webmaster Guidelines. We've reviewed your site and we believe that some or all of your pages still violate our quality guidelines. In order to preserve the quality of our search engine, pages from http://www.example.com/ may not appear or may not rank as highly in Google's search results, or may otherwise be considered to be less trustworthy than sites which follow the quality guidelines. If you wish to be reconsidered again, please correct or remove all pages that are outside our quality guidelines. When such changes have been made, please visit [google.com...] resubmit your site for reconsideration. If you have additional questions about how to resolve this issue, please see our Webmaster Help Forum for support. Sincerely, Google Search Quality Team"

Related to Panda: Did you see this at all?
[wired.com...]

I took a look at a few patents and now it's getting all clear:

1. They are always testing something so there is no stabilty in results for the most part.

2. It is no wonder when some people's long standing sites dropped. One patent says that they now collecting a database of mouse movements (hovers, etc) and clicks, on keyword and site level. They started this collection from 2007. Now they collected enough data to see patterns and they launched this Panda thing. One patented indicator for example, compares average time users spent on the site say in previous year with the new average time, say for a few past months and if users spend much less time, then this document is considered "stale" and lowered in ranking. And there are many other signals including click through rates, back button clicks, query modifications after back button, etc.

The point is that even if you did not change anything about the backlinks or your site, and it dropped it is because of the whole array of these new signals that you can def. control but it is more trickier now than, say backlinks, etc.

I am making sure I am reading up on all the Google patents now. Boy, I already learnt a huge deal of what I was doing wrong. Domains like yada-yada-bla-bla.com are no good and a lot of other stuff. This is my 5 cents, hope it helps.

[edited by: Robert_Charlton at 6:47 pm (utc) on Apr 27, 2011]
[edit reason] examplified site domain name [/edit]

[2macarena]

pontifex: in that wired article there is the answer to your question btw.