Panda 4.1 Rolling Out

[netmeg]

Based on user (and webmaster!) feedback, we’ve been able to discover a few more signals to help Panda identify low-quality content more precisely. This results in a greater diversity of high-quality small- and medium-sized sites ranking higher, which is nice.

Depending on the locale, around 3-5% of queries are affected.

Google+:
https://plus.google.com/+PierreFar/posts/7CWs3a3yoeY [plus.google.com]

(sorry, it's https so it won't link)

Searchengineland:

http://searchengineland.com/panda-update-rolling-204313 [searchengineland.com]

[edited by: aakk9999 at 11:40 pm (utc) on Sep 25, 2014]
[edit reason] Made link clickable [/edit]

[seoskunk]

I am actively making changes to a site despite the update, I don't know if I have missed the boat with the changes. But I am carrying on as normal.

I don't understand how a "slow rollout" works

I think they got the idea from the Chinese Water Torture :)

[rish3]

Some posters in this thread said that they saw this rollout start on Sept 19, which would be the end of last week, not earlier this week. Also, I don't understand how a "slow rollout" works. Does it mean that the search results keep shuffling, and that the rankings of individual sites keep bouncing around?

I don't have citations to back it up, but from past experience...

- Google has done limited pre-rollout testing (subset of sites/niches) with large changes before. That may explain the changes that started on the 19th.

- Matt has talked about Panda (Penguin as well) updates consisting of two distinctly separate processes....the data update and the algorithm update. That's one way they can slow-roll it...roll the data, then roll the algo days later.

- They can roll out in a granular way, and have consistently done so. Not just at the "geo tld" (google.co.uk, .in, etc), but at the individual data center/server level, within a single google tld.

Between the three, you would see almost continuous shuffling, especially if you check from 2 or more devices that connect to a different data center. My past experience is that it takes a full week from their official tweet that an update has started until the dust finally settles.

That's also a good reason not to get overly excited (or dissapointed) with the SERP changes you've seen today. I have seen many people celebrate (or complain) too early, only to see a 180 degree turn of events at the end of the rollout.

edit: typo

[RedBar]

It's certainly not done me any favours so far whatever they claim to be doing:

Friday PVs v Sept average 77%
v previous week average 67%

USA PVS v Sept average 60%
v previous week average 56.5%

UK PVS v Sept average 65%
v previous week average 52%

India PVS v Sept average 66%
v previous week average 65%

Fubard for me at the moment Google, not impressed at all.

[snippet]

I'm noticing a generic info site at the top of my medical/health related searches. Kind of disappointing to see that site above gov sites and authority sites. Their content is always so generic and just seems either copied or watered down.

I tried to read their page for a few searches thinking maybe it was a good resource. But about a paragraph in I realize they are not going to be much help since I quickly lose trust with no identifiable source of their info.

In medical and health related queries, I need to see who is saying it to decide if I trust it. Or at least cite some authoritative sources if you are just re-writing something.

[tibiritabara]

Hi folks,

I have several websites testing panda since several years ago. Traffics were Ok using spam content which is on of the main target of google panda.

I've detected decreasing traffic from Google since June. Google were depreciating duplicate content every week, loosing 10% to 15% every week since June.

The decreasing of traffic have been increased since September 18th.

This kind of "discreet" roll out can be connected to genetic algorithms that google is using to evaluate content, user experience, etc.

cheers

[EditorialGuy]

In medical and health related queries, I need to see who is saying it to decide if I trust it. Or at least cite some authoritative sources if you are just re-writing something.

That's something they're working on, according to Matt Cutts. (Earlier this year, he mentioned health queries specifically when talking about how Google wanted to reward authority or subject expertise. He didn't mention an ETA but said they had "an engineer" working on it.)

[chrisv1963]

It's disappointing that after this update there's still so much Pinterest crap dominating the serps in my niche. If I want to see content copied from other websites or pictures posted by business that hope to increase sales, then I'll go straight to Pinterest. I don't need Google for that. I want to find good sites with the ORIGINAL content.

[samwest]

@chris - same goes for Houzz...I don't have to search for it anymore, their millions got the buzz out there, we all know about it, "ad nauseam"...just like eBay, Facebook, Amazon...they don't even need listings.

I am seeing better user interaction today and the traffic faucet seems to have been opened a bit more. Hope I can start breathing again.

[RedBar]

Add another one to that image list roomzaar ... could we have a new www please, one for originals and one for thieves and plagiarists?

Google will obviously be in the plagiarist one along with all the other snake oil people.

[samwest]

@Red - yup, same as a gazillion other wanna be sites that jump on the bandwagon of the first successful site. Eventually they'll all cancel each other out. We can only hope.

[deeper]

John Müller in his hangouts often encourages to give him (for example G+mail to him) concrete examples of bad rankings and he will pass them to engeneers to check if algo should be improved.

If here are several guys, each with several relevant examples, you could together create one big list...

Worst case: nothing changes, but may be it gives at least some fire... so why not trying it?

[deeper]

To be more precice: "bad" rankings, i.e., concrete examples of good rankinks which look suspicious, not very useful for the user and therefore hard to understand.

[aristotle]

deeper --
For years Google has had a "Send feedback' link at the bottom of their search results page. I used to use it to report hacked sites and parked pages that had high rankings, but when I would check back a week or two later, nothing was ever done to correct the problem. So I decided it was a waste of time and stopped doing it.

So next time I see that type of thing, maybe I should try reporting it directly to John Muller.

[Saffron]

I would really like to talk to John about the domain clustering I am still seeing. One site having 40% of the listings on page one does not seem right at all.

In other news, my site does seem to be picking up a little which is good. I've worked my backside off improving older articles.

[dethfire]

Hard to say for me. Weekends are always low, but today is low even for a saturday

[deeper]

@aristotle, Saffron:
We all know how many factors influence the rankings and Google may not have always the same definition of "useful" for a user. Further you don't get feedback for your feedback, so certainly it often may be frustrating. Is it worth spending time? I don't know.

But I'm rather sure they really ckeck what they get and want to improve their SERPs for the user. Therefore having a longer list with blatant examples... collected by the power of a forum... may be concentrated on one or two "issues"...it's really worth thinking about it. They probably need some evidence, time, data and importance in order to change the algo. Not surprising to me, so hopes should target not to next week. Their primary goal is to improve the algo on the long run with stable data, not to correct some sites immediately.

If a site clearly violates guidelines their reaction CAN be very fast, eben with single cases. I know one example.

Google makes mistakes and I don't want to defend them, but their wish to get feedback in order to improve their work is not only "public relation".

Just my thoughts.

[netmeg]

I am the beneficiary of domain crowding on multiple sites (and I'm not a big brand by any means) so it gets no complaints from me.

[menntarra 34]

Hi, i made changes to my site on the 4th of September. I changed a lot: fewer ads, a whole new design, which is now responsive. It is much much faster. The whole system changed, content can be reached much easier. And despite all this i suffered another 10% loss. Do you think that this update measured my changes, or may i hope that they did not take my changes into account. So maybe renewed site can come out as a "winner" with the next PANDA ? I see the massive crawl came to my site at about 8-9th of September, which should mean that they sadly saw my changed... What do you think?

[samwest]

@menntarra.. I did the same on Aug 4, switched from static html to CMS, 301'd everything, same content with more added. Better navigation, faster 14 year old site and now I'm looking at either zero visitors or 1...all day. No sign of any recovery yet...only traffic is coming from previous customers. Nothing new except new lows.

[Awarn]

The issue we keeping what should logically be rewarded - not what actually is rewarded. We think technically. In our eyes clean code, design, speed etc. They could care less. What you see is what they want so adjust accordingly.

[deeper]

@samwest:
When did you loose your rankings? Before panda?

[samwest]

@deeper - that's the odd part, we really didn't loose rankings, we just lost traffic as if they put the site in the sandbox. We did lose long tail, plurals and semantic search, but that was slowly building 4 years ago already.

Daily traffic used to be a nice, natural sine wave pattern, up and down, now it's as flat as a pancake...even since the switch.

I have 50k members, so the traffic I am seeing is paid return customers and maybe a few PPC hits and a 1 or 0 hit trickle of organic taffic.

If I run a keyword rank report, my positions seem to be static, all the same, a few new ones and +/- 1 or 2 deltas, but nothing drastic.

We were never hard hit by Panda, only my changing the site from html to CMS. It's not a canonical issue, as far as I can tell.

Looks like they installed a soft clamping diode on my site. It's as if most pages lost all their juice.

[Lorel]

My traffic has increased over 8% since Sept 19 and still growing and ranking has improved on a lot of keywords. At first I though the increase in traffic was due to colleges starting up as I get a lot of traffic from students of all ages.

My traffic dropped last May and suspecting Panda I removed several pages and added more text to others, etc. I thought the drop in May might also have been due to colleges closing down for the summer but now I suspect it was Panda, or maybe both.

[samwest]

@lorel - sounds like you have less than a full years worth of data. In my niche, we see (used to see) the same pattern. Highest day of the year was always Dec 26, lowest was around July 4 and from there it always went slowly up toward Christmas. 14 years, same pattern...except this year, which is like the trajectory of a lead balloon, on fire.

[deeper]

@samwest:
Same rankings but
a) Google-user don't click? or
b) they click but don't reach yout site? Almost no traffic at all?

Sounds like a technical issue. Measuring your traffic works fine?

[Broadway]

I would think this is the most important thing I could tell anyone suffering from Panda is this:
Read the articles that have been published on SearchEngineWatch about Panda recovery.

They tell you how to use Google Analytics (mostly landing page data) to see how Google interprets the worth of your individual pages.

Bottom line, if you have any page Google doesn't send traffic to, it's a sign that they don't respect it
(either because of duplicate/thin content, poor user engagement, whatever...).

So either no-index it, combine it with other pages and 301 redirect, do the Prev/Next thing, or else 404 it.

I would think that all websites are affected by the 20/80 rule.
80% of all of the good stuff about your site (traffic, ad revenue, whatever...) comes from just 20% of your pages.

Slash and burn the non-engaging 80% (via one of the methods above) and you'll never miss them.
And there's a good chance you'll loose your Panda penalty in the process because to Google the metrics of your site will look stronger.

[samwest]

@Broadway...that is probably the best advice and direction I have ever received from WebmasterWorld, even from those (except a few) who claim to be experts.

My old html site was 70 pages, combining my 70 pages in with all my old blog articles and some real clinkers from as far back as 2007 has obviously diluted my relevancy.
Old site = 70 pages
New site = 300 pages.

I only include pages and posts in my sitemap. No attachments, galleries or nuthin' All pages and posts that I intended to keep are run through Bing's SEO check tool. None were flagged.

At least now why and where to start chopping. Thanks

[aok88]

I have two questions:

1. Why did they about Panda 4.1 'a “slow rollout” began earlier this week and will continue into next week, before being complete", when it seems like it has Already rolled out? Searchmetrics already has a set of winners and losers! (if you do a search for 'panda 4.1) [blog.searchmetrics.com...]

#2 Why did MC say they won't be announcing any more Panda's, then they went ahead and announced 4 and 4.1? Does this mean they will start announcing them again and does this mean there were NO other Panda's since May's Panda 4 till now?

[Sand]

Saw big gains this week, after being hit by Panda 4. I changed nothing on the site between updates.

[aristotle]

Saw big gains this week, after being hit by Panda 4.

Did you basically gain back what you lost with Panda 4, or did you only regain part of what you lost, or did you gain more than you lost thereby reaching new heights?