New Panda Update - September 27

[BrodyDodes]

Google rolled out its most recent iteration of the Panda algorithm on Wednesday 9/28/11 (or Tuesday according to who you ask). Saw the first recovery of one of my punished sites since the first iteration back in February. Anyone else see recoveries?


[webpronews.com...]

When asked if an iteration of Panda was implemented this week, a Google spokesperson told us, “yes.” She also provided the following statement:

“We’re continuing to iterate on our Panda algorithm as part of our commitment to returning high-quality sites to Google users. This most recent update is one of the roughly 500 changes we make to our ranking algorithms each year.”

If you’ve followed the Google Panda update saga throughout the year, you may recall Dani Horowitz’s story. She runs an IT discussion community called Daniweb, and it was hit hard by the Panda update, but she made a lot of changes, and gradually started to build back some Google cred

[affiliation]

I think Google previously would have punished these sites, but have they have changed tact, are they now punishing us in order to stop us submitting to these article websites that are known for allowing and encouraging people to copy their content.

This may be a lot of what Panda is all about, to stop people using article and maybe even press release websites. If we get a peanality we are less likely to use this type of site for promotion.

[Hissingsid]

This may be a lot of what Panda is all about, to stop people using article and maybe even press release websites. If we get a peanality we are less likely to use this type of site for promotion.

I'm seeing too many examples of sites that have used PR and article sites being pushed up the for this to be correct.

[Hissingsid]

Just thinking aloud again.

For some two word terms for example <<widget service>> I'm seeing sites that are strong semantic webs for "widget" that include "widget service" in them have been given a boost. Those that are strong for "service" but not so strong for widget have lower rankings for this term than previous.

This does not apply for all terms.

This could be for all sorts of reasons but perhaps most likely are:

Perhaps Panda has a large element of semantic webs analysis in the algorithm and sites that are strong for the general theme of a search give a stronger signal into that analysis than ones that are strong for the specific term.

and/or

Sites that are strong for the general topic provide better user satisfaction data. Perhaps because users stick around and explore the site more than one that only answers one specific question.

Alternatively if semantic web analysis is involved I wonder if Google is using a technique that summarises semantic analysis into vectors and errors in the way the analysis is conducted and the vectors stored results in the "false positives" we are seeing in SERPS.

Cheers

Sid

[MikeNoLastName]

I think there may be some confusion by some on the definition of "bounce rate". At one time I think it WAS naively defined as going into a web site, viewing only one page and exiting (perhaps with some time variable functioned in. As I understand it, in most cases Goog can't reliably tell how many pages on your site a person views (unless you used Goog analytics for your site! or they have google toolbar). Going to another page on your site, could be just as bad as leaving, since it may mean they did not find what they were looking for and simply gave you another chance. This is too easy to manipulate by site owners by doing page refreshes or leading people on with enticements.
Therefore they can only evaluate what they CAN see and that is: whether a person RETURNS to G and SELECTS _+ANOTHER+_ SEARCH RESULT USING THE SAME SEARCH TERMS!
In response to the apparent concern by some in this thread, at least _I_ firmly believe... because it makes the most logical sense and any other treatment by G would be simply silly, if they get to your site searching on "green widgets", find the info they wanted, then backed out to g and search on "red doodads" before clicking on another search result... then it should not count as a bounce. Whereas, if they DO click on another result without changing the search terms, it should count as a bounce.
This of course, has many exceptions, as someone previously mentioned where, for instance, someone is of the type (like me) that likes to comparison shop before finally returning and buying.
My suggestion is to search on "search engine" click on G, then go back and click on the next SE. Easy to do to your competitors too, then click on your own link, wait, and go back and search for another keyword ad infinatum ;-)

[Dave_Hybrid]

Im curious as to everyone interpretation of Matt Cutts tweet, ie; Panda Flux in next few weeks but less impact this time ~2%.

Would you consider things to get worse, ie; there cranking things up so if 10% were hit last time 12% are likely to be impacted overall?

Or, they are effectively turning things down, total impact around 2% so in the big scheme of things an 8% roll back?

Anyone?

[tedster]

Small amounts of "Panda flux" have tended to follow after each iteration. And that flux has meant some sites that were hurt regain traffic (some or all) - and likewise some sites that got a boost have lost some or all of their gains.

This flux tends to affect only a small portion of all the sites impacted, whether positively or negatively, and so this "Panda flux" had seen little comment here in the past, just a few posts.

As a matter of fact (and human nature) we see many fewer posts from websites that gained in any Panda iteration than we do from those whose traffic suffers. That does tend to give a skewed sense of what's going on. So those who have access to large amounts of rankings data - not just their own niche - have an advantage in seeing the bigger picture.

I wish I had more of that kind of data myself. I only have a taste, and so far it hasn't been enough to indicate any conclusions, but it can be tantalizing.

[tedster]

The following idea is speculation only. My thoughts have been partially stimulated by the "Panda flux" phenomenon, and also by the idea of machine learning and data modeling for complex goals.

Trying to build a machine model that simulates the human idea of "quality" is a highly adventurous undertaking. We've all seen examples this year of how these early results have missed the mark.

There is another area of human endeavor where machine learning and data modeling is trying to achieve a complex goal - weather forecasting. And that area has several decades of experience built up. If your experience goes back a few decades, you probably realize that short term weather forecasting has improved rather remarkably over that time.

Weather prediction relies on using more than one algorithm - each model takes its data model through to an end result, but then those various predictions get integrated (often by human analysts). You see this especially around predicting hurricane tracks. My point being that a single algorithm is often not enough to accomplish complex goals.

And so I wonder if the "Panda flux" we notice is the result of other supportive algorithms kicking in to balance the main algorithm - perhaps being integrated after some human analysis of the shortfalls and edges cases that surfaced.

The "main Panda algorithm" itself may really be a complex combination of several different models. If that's so, then it would be almost impossible to reverse engineer the thing - and it certainly does seem impossible to take apart right now. In the frustrating absence of some basic understanding, the only guesses we can make are based on very localized and (relatively) small data sets that we each have access to.

[Freedom]

Wonderful post tedter. It would explain why so many webmasters see "one thing," and other see something different.

[Dave_Hybrid]

Yep, thanks for that, interesting reading.

[Hissingsid]

I remember off forum discussions with various people after the Florida update. It ended up that it didn't really matter what you thought had caused the problem. What mattered was finding something that you really believed in and doing something about it in a concerted way. Some people thought it was an over optimisation penalty that caused their problem, others that semantics was the root issue and that they needed to broaden their page and site semantics.

In the end doing something about one of these probably did something about the other so it didn't matter if you were right.

I think that Panda is probably the same in terms of find something you believe in and implement it. If you work on the assumption that you need to improve the user experience, find reasons for people to find satisfaction in your site and make your site more semantically rich there's a fair chance that you will do stuff that gives the right signals to Google and you'll start the recovery.

That's what I think I'll do anyway.

Sid

[Matrix]

I have a question?

is all you guys using Google analytic on site that are punish ?

[dunivan]

There are ways te recover from panda, just the majority of large-site owners don't feel like the rigorous review that must be done to identify problems and attempt to remedy them. Look for content that has a high bounce rate pre-panda. Scrutinize it with the 22 questions google put out, re write it to be THE SOURCE of information on that topic. If you are unsure about some content, put it on an island, a subdomain for example, and see what kind of search impact it has for those keywords.

For those concerned that the google crawl spike is indicative of getting nailed by panda - it is not, i have a few sites that were crawled at double rates for a day and have had no panda issue.

Also, if you think you have been hit, check your organic search traffic, if this seems to be the same, it is probably not panda.

[Matrix]

For those concerned that the google crawl spike is indicative of getting nailed by panda - it is not, i have a few sites that were crawled at double rates for a day and have had no panda issue.


Crawling rate have been accelerate for a reason, to find the sites that have bad and good content or use allots of long tail keyword to rank better then the sites that rank good for short key word and in fact should also rank top for long tail key phrase...

[dunivan]

Crawling rate have been accelerate for a reason, to find the sites that have bad and good content or use allots of long tail keyword to rank better then the sites that rank good for short key word and in fact should also rank top for long tail key phrase...

Right, but a consistent increased crawl would be indicative of a higher ranking long-tail search in that example. The single crawl jump is google trying to get a complete index prior to dropping the hammer on sites across the index.

[Rasputin]

Hi Dunivan, are you saying that you actually have a site that has completely recovered from panda (on one of the panda run dates) where the only change made to the site was to identify poor quality articles and to rewrite them, or were other changes made to the site as well?

If so are you 100% sure it was panda issue (daft question perhaps but a lot of changes seem to be attributed to panda nowadays even outside the dates when panda is being run)?

I am aware that google suggest this, and it's pretty much what I'm doing myself (except I'm using bounce rate along with time on site and article size to identify low quality articles rather than just bounce rate), but I haven't heard of it actually working for anyone...

[zeus]

I would never use Google analytic thats just to much site info for one SE

[tedster]

all you guys using Google analytic on site that are punish ?

A definite "no" on that idea. I have had several sites with Panda problems approach me who don't use Google Analytics.

[dunivan]

Hi Dunivan, are you saying that you actually have a site that has completely recovered from panda (on one of the panda run dates) where the only change made to the site was to identify poor quality articles and to rewrite them, or were other changes made to the site as well?

None of my client sites have had panda issues. I have however helped other seo's recover using those methods. All that was done was refinement of site architecture, and a rewriting of low quality content which was id'd using bounce-rate/visitors ratio (e.g., a page that got 100 visits but had a 93% bounce rate is probably low quality)

If so are you 100% sure it was panda issue (daft question perhaps but a lot of changes seem to be attributed to panda nowadays even outside the dates when panda is being run)?

It was panda, organic google search was down 80-85%

I am aware that google suggest this, and it's pretty much what I'm doing myself (except I'm using bounce rate along with time on site and article size to identify low quality articles rather than just bounce rate), but I haven't heard of it actually working for anyone...

This is where site architecture comes in, google believes panda is right all of the time, and won't recrawl what they hit until the next iteration at least. Moving the content to a sub-domain gets it recrawled and you can see your work if its better or worse, this is based off hub-page theory. I don't think that is why they show recovery, but it helps seperate the wheat from the chaff so to speak.

[dunivan]

Also, not using analytics (or another tracking software) is shooting yourself in the foot. Way to rule out new content that can get searches without working too hard.

[walkman]

(e.g., a page that got 100 visits but had a 93% bounce rate is probably low quality)

Fine in theory, but in reality bounce rate is dependent on how /with what keywords brought the user there. The page can be great but if Google sends people there with 3rd tier long tails, people will bounce. Obviously you can't change the page based on what type of keywords you get this week, as some suggest. That's why Panda is irresponsible and designed to literally kill certain sites, unless they do a specific thing. Shall we guess what it is :) ? Panda started by calling good certain sites that have skewed stats based on things most small stores and companies can't afford.

Any discussion of Panda that leaves out Google's corporate goals and long-ago stated intentions is just wasting time IMO, eventually they will knock on your door. And, I'd venture and say that most sites hurt, by far, don't even have more than a few secondary articles in their sites. Most are small shops and mom-and-pop stores but Google set the debate tone by calling the penalized sites "low quality," "shallow," giving guidelines that apply only to content farms and many followed that line like lemmings. Of course they un-penalized the high profile ones to blunt criticisms leaving others penalized, trying to keep the idea that 'panda is right so your site must suck' alive.

[dunivan]

right walkman, but a site that does tailor to the query should not have 93% bounce, maybe 70 - 80%. Couple this with a time on site metric and you get my drift.

I am starting to believe this is all about showing higher google rankings (youtube's growth in 2.5) but we will see

[walkman]

right walkman, but a site that does tailor to the query should not have 93% bounce, maybe 70 - 80%. Couple this with a time on site metric and you get my drift.

93% or 78% we can argue all day but I'll use another example I used before: if you continuously send "Nike shoes" type of keywords to my competitor and "Nike shoes size 13 in black color" to my /nike/shoes how is it fair to compare bounce rates or time on site? It's not but Google doesn't care, the rich get richer and the poor die off since Panda gives you a sitewide penalty.

That Google doesn't care about "user experience" when balanced with $$, it's obvious: just look at all the crap they fill the page with: spammy Youtube videos, image searches totally out of place, totally unrelated book results in SERPS, and the famous 7 pack filled with 100% Google, but inferior, local search pages. How stupid or brainwashed does one have to be not to see this?

[dunivan]

how is it fair to compare bounce rates or time on site?

High bounce coupled with a high time on site is indicative of a user who reads the article and is not merely skimming it, the article provides a value that the user wants to continue interaction with. High bounce with low time on site shows users are just skimming or shallowly interacting with content. Shallow interaction can be a descriptor of low-quality content.

I doubt the money interest on the algo update side of things, but the rise in certain google properties through panda updates is definitely to be scrutinized.

[Matrix]

I also notice that i still rank same place for allots of key words phrase but i don't get the traffic i use to for the same positions in search result...

[dunivan]

I also notice that i still rank same place for allots of key words phrase but i don't get the traffic i use to for the same positions in search result...

without personalization?

[Matrix]

i when from Alexa position 163,000 to now 508,000

[walkman]

High bounce coupled with a high time on site is indicative of a user who reads the article and is not merely skimming it, the article provides a value that the user wants to continue interaction with. High bounce with low time on site shows users are just skimming or shallowly interacting with content. Shallow interaction can be a descriptor of low-quality content.

I have a feeling you are missing the point, entirely. People are much more likely to be a good match, browse and stay when coming in through a good keyword than a long, long tail one. Brands already rank much higher for "good keywords" and now non-brands are now getting a sitewide penalty (maybe) because people are less likely to stay around and browse. You can have a great article but I thought it was about something else or tangentially mentions what I had in mind, I'll leave. Or if I came via a search kwd that suggested that shoes were at a certain price (maybe the price was for another model on that page) I'll leave if the price is different. But if the keyword was general I'd browse.

I doubt the money interest on the algo update side of things, but the rise in certain google properties through panda updates is definitely to be scrutinized.

Search bias, it's hard to prove but start looking around draw your own conclusions. Start with how does Google make money. But anyway, some take it very personally when others suggest that Google is a ruthless corporation racking in some $40 Billion this year, so we should drop the money talk. They're just a few super-smart and caring nerds trying to have fun and change the world. Money just falls in their lap.

[Matrix]

@ walkman

I see allots of post talking about site with article and information s about product and other , but what about site that offer a service that don't have article and product but only a service ...

is panda supposed to kill them too?

[potentialgeek]

I cut bounce rate down to 8% on one section of my site and doubled average pageviews. It still went down in rankings after Panda 2.5.

[Hissingsid]

Anyone else notice that searching for plural of a term where a site has been boosted for the singular removes the boost?

For example <<widget service>> gives a site a boost <<widgets service>> and <<widgets services>> removes the boost.