Google Updates and SERP Changes - June 2011

[dataguy]

< continued from [webmasterworld.com...] >

FWIW, my early change indicators are ringing, starting about 5 am CST this morning, and it seems to be Panda related. Possibly reverse-Panda related.

[edited by: tedster at 3:31 pm (utc) on Jun 1, 2011]

[superclown2]

Has anyone with very small sites been affected by all this? I mean something like 10-20 or less pages? Or is it only large e-commerce and UGC sites being affected?

I have some fairly mundane small sites that greatly outrank my larg-ish (500 pages) specialist site for that site's main keyphrases. The big site has far more unique, highly relevant content but languishes below sites which were only designed as feeders in the first place. Crazy.

[arikgub]

reloading google search results page 3 times in 5 minutes and getting 3 different sets of results.

absolutely crazy.

[indyank]

seeing some strange behavior.Not the ones as experienced by arikgub.

1) Right now, as I am adding this, I am seeing a domain (.nl TLD) that frames a page in a popular site ranking high (#3) in google SERPS for a key phrase (two words).The original site is nowhere to be found while it used to rank in positions #3 or #4 before.The .nl domain is framing a page from the popular site whose content in English language.

2) 20 hours back, while doing logged in searches, I saw several pages on sites I control rank in the same positions as they did before panda for their keywords.But when I tried it now, they all seem to have fallen into their post panda positions.

[eyeinthesky]

One of my sites that went AWOL a few weeks ago came back from the dead today. Sites can come and go like a yoyo.

Form hero to zero and back to hero in a few days?

[danijelzi]

Today I've searched for a couple of keywords via Ubuntu's Google custom search (homepage in Firefox), Web Search option selected. What I have noticed that scrapers who are ranking for queries via Google.com are removed on Ubuntu search. My keyword positions for a pandalized site are the same as on Google.com, so the only difference is absence of scraper pages.

On another custom search search.icq.com is the same situation.

Tried also AOL's custom search, but the results there are the same as via Google.com.

[marketingmagic]

Traffic level back to pre May 17th traffic levels yesterday - hoping it's a sign that the dials have been re-adjusted again at the Plex.

[muejl1]

At least in my niche, I've noticed that the Panda update, and especially the latest tweaks, have really changed the SERP results between Yahoo/Bing and Google. I wonder if the latest algo change had more to do with Google seperating themselves from the crowd?

[zeus]

I think they have just started something with the rankings, I have followed a few pages and seen them move about 3 spots back and forth which is normal but now one of the pages has gone from page 6 on google to 12 which of cause is bad but maybe its just the start and it will be on page one or 2 again as pre panda 2.0

[ANielsen]

Have seen major G. SERP changes in some two-words keywords since yesterday, June 13, 11 pm. North Carolina time.

SERPS are the same today. My traffic has not changed significantly, so maybe the changes is only in my area (my sites got up in SERPS for those words).

I suspect that G. does some shuffling of sites to study bounce rates and time on site for different types of keywords.

Maybe they are checking three-word keywords for this Panda version (2.2) implementation.

[HuskyPup]

I'm seeing a lot of movement for my widgets tonight UK time, pages that had been demoted or missing since 2.1 suddenly back on the front page. My logs are showing a substantial increase in traffic...we'll see for how long!

[dazzlindonna]

Fingers crossed for you, HuskyPup

[dickbaker]

I'm seeing a return of some phrases to the first page, although they're not the phrases that brought the greatest traffic.

I made changes to the site a few weeks ago, including no-indexing half the site, adding content, etc, but I haven't had time to work on the site for about two to three weeks. Too much else going on with the site (orders, changing hosts, etc).

I may be getting a little bit of Google love, but I don't think I'm going to get as much as I had without doing much more than I already have.

[AlyssaS]

Maybe they are checking three-word keywords for this Panda version (2.2) implementation.

Yes! I think Panda 1 (Feb 24th) and Panda 2 (April 11th) were about assessing those sites that ranked for one and two word keywords. Panda 2.1 and now Panda 2.2 seem to be about those sites that rank primarily for three word keywords. And I expect they will continue down the chain in a months time looking to nail smaller sites that rank for longer keywords.

Regarding the split testing going on, someone on the WF posted up this hypothetical schedule:

6th June

Group A (site a,b,c) Back to Page 1,2
Group B (site x,y,z) Gone

10th June

Group A (site a,b,c) Gone
Group B (site x,y,z) Back to Page 1,2

14th June

Group A (site a,b,c) Back to Page 1,2
Group B (site x,y,z) Gone

18th June

Group A (site a,b,c) ?
Group B (site x,y,z) ?

If you are in Group B you probably have a couple of days to improve your site and get it re-cached so that you shine when you are flipped back onto the first page. It's too late for Group A, they are being assessed as we speak.

This is like being at school taking exams isn't it?

[ANielsen]

Alyssa,

For me May 5/6, also evening, was a disaster, although my sites are very professional (honestly). Tomorrow I will see what really went on in Analytics.

I think it is wise of G if G does not to derank pages with relatively high bounce rates, or short durations of visits, if high bounce rates and short visits are common across multiple sites for that keyword (KW).

I can imagine they "scan" keywords at different sites, gather data, and act accordingly.

Since there is an apparant big move for two-word KW, they are probably switching to other KW's now since something seem to be going on.

If that is the approach taken by G., they will probably make the internet better - even though it sucks when your site is deranked.

But if site A is found on the same keyword as site B, and people like staying there much longer and bounce less on site B, site B is probably more relevant to their clients (the average surfer).

[ANielsen]

The last sentence should have been:

If site A is found on the same keyword as site B, and people like staying on site A much longer and bounce less than on site B, site A is probably more relevant to the average surfer.

Sorry :)

[walkman]

For me May 5/6, also evening, was a disaster, although my sites are very professional (honestly). Tomorrow I will see what really went on in Analytics.

If they are using time on site and bounce rate as the defining, make or break point, then content and designing for the user might very easily backfire on you. That's truer in certain niches and it's not very hard to see why.

I really hope they aren't, I might need to change my site design.

[ANielsen]

Wish I knew how to copy-paste text from your post, Walkman.

If my suspicions are correct, (they may not be) design that encourage visitors to stay longer and bounce less is king.

That should be kept up against loading speed of course.

Wish I knew some designs that makes people bounce less and stay longer, for the same content.

Any research available?

[walkman]

If my suspicions are correct, (they may not be) design that encourage visitors to stay longer and bounce less is king.

All depends on the niche, what you say can even work against users in plenty of cases, especially for those coming from search engines. I'd be shocked if Google used this as a hero/zero factor. It's deeply flawed.

[outland88]

They’re definitely shaking things up in my areas starting Tuesday night. Branding became even stronger. Some of the bigger names hit by Panda seem to be returning. Commerce sites are becoming fewer and fewer in top results.

[HuskyPup]

For me May 5/6, also evening, was a disaster

That was Panda 2.1

From my post last evening:

I'm seeing a lot of movement for my widgets tonight UK time,

It seems to have reverted but it did give me my highest page impressions day and visitors this month so far. Something's certainly going on but does anyone, including Google, know what?

[tranquilito]

Another big algo update seems to be in progress right now. Anyone else noticing changes in SERPS ?

[Pilsen]

I see the major changes today.

Noticed a huge traffic drop 10-12 hours ago at 2 websites (9k & 30k daily traffic), the 2nd one has already recovered, the first one got 50% traffic drop

[dickbaker]

I ran a ranking program late last night, and there's been some movement, but nothing significant.

On one particular page, though, it went from #108 to #28. This was for a phrase for which I ranked in the top ten pre-Panda.

Last month I did a video of the widget that this page is about, and put the video and the video transcript on the page. Other pages on the site are holding this one back (if we all still agree on that concept), so this would probably be back on page one.

Taking a couple of days to do a video for one widget when I have a couple thousand widgets just isn't feasible, though. That's even assuming I could get my hands on all of the models of widgets.

[ohno]

Can anyone explain why any tweaks these days seem to result in a total nightmare? In years gone by we would not notice an algo tweak. These days we know one is coming...

The pattern :-

Traffic goes UP massivly, sales normal
Lots of foreign traffic starts to appear, sales decline
Traffic then becomes normal, sales dead

Then up to two weeks of anything between normality & total mess.

Checking the SERP's gives no clues. Although I started a thread as to why the results no longer feature Google shopping results for lots of search terms that once did.

[zeus]

tweaks, yes i also saw some movements yesterday, but absolutely not to the better.

[coachm]

The traffic on our previously untouched by Panda site was just cut by about 65% on the 14th. Oddly we had our biggest traffic day in ten years on that site on the 13th.

Just adding in an additional data point. I'm tempted to change lots of things, but I have to keep reminding myself to wait it out and at least hope it stabilizes so I can figure out a strategy.

[dazzlindonna]

At this point, after seeing it mentioned time and time again, I think we have to conclude that if your site gets a huge boost in traffic one day, for unknown reasons, you'd better get ready for the ship to sink the next day.

Now, if we could find patterns in the analytics from everyone who had this boost before fall, we could perhaps determine a common denominator. For example, let's assume all this traffic is because Google is sending a hoard of quality raters to the site (go with me here, even if this is untrue). Where is all this traffic going once they reach the site? What's their navigation path? When do they bail? Is there a pattern we can see based on what they looked at? If so, maybe that pattern can tell us what features of our sites we need to be focusing on as the kill-points for Panda.

[ohno]

Does anyone else see foreign traffic during these blips? It happens every time with us, when I see it my heart sinks as I know what is coming........

Strangely soemthing happened around 3pm for us on one site, like a tap was switched back on.

[ohno]

Although thinking about is, I can't see any way to avoid this? It's like what used to be the monthly Google dance. We never change anything & the sites have always recovered?

[AlyssaS]

At this point, after seeing it mentioned time and time again, I think we have to conclude that if your site gets a huge boost in traffic one day, for unknown reasons, you'd better get ready for the ship to sink the next day.

Now, if we could find patterns in the analytics from everyone who had this boost before fall, we could perhaps determine a common denominator.

I think it's as simple as them putting a site high up on page 1 to assess it.

Loads of people report seeing "junky" sites on the top, and loads of people are reporting sharp swings in SERPs, were stuff is suddenly bumped right up.

I think this is deliberate - they've been bumped up to see what the user reaction is.

This Panda machine-learning thing probably needs sufficient data to confirm whether a site is good enough - otherwise, whatever they are seeing may be statistically insignificant. Perhaps they are targeting a certain number of SERPs page impressions - so you are up there till they get the data they want, and for big sites this will be achieved quickly. Once they've got that, they compare to their control for whatever it is they are assessing, and if you fail, down you go.

I think this thing is being run only on the top two pages. But that's just me guessing, like everyone else! I'm assuming that as the machine accumulates more and more data, the swings won't be as violent, perhaps only new sites will get suddenly bumped up for assessment.