Recovery from Google Changes on 16 Nov 2012

[Sgt_Kickaxe]

On Nov 16th Google made a change that impacted a good number of sites. Some asked Google if it was a Panda update and Google said it was one of 500 changes Google made per year but was not a Panda/Penguin update, one of those is due out this week. You can read about some of the affected sites in the monthly seo thread here - [webmasterworld.com...]

Specifics
- The change was not sitewide, it impacted individual pages while leaving others alone
- Unique and on topic content on each page
- Very little else is definitive but we did have pages to compare within the same domains

While comparing affected pages nothing about the page, from an SEO perspective, stood out. Backlink profiles were varied but similar to the non-affected pages. On my own site the pages affected seemed to be more brand/transactional in nature but some of these were untouched as well.

Recovery
- The one common denominator on all affected pages was medium quality content. The pages I "beefed up" from Nov 17th to 21st have already recovered(as of yesterday).

What struck me is that this isn't the low quality type content you'd expect to see receive a downgrade. 100% unique, accurate, on topic and useful content received the downgrade but in all cases there was either not enough useful information or it was worded a little too generically and would likely get a borderline rating by a human review. Definitely not spam but only of marginal usefulness, too narrow in scope or not comprehensive enough to solve a problem.

The speed with which these pages are bouncing back, and the fact that the pages I haven't "fixed" yet aren't, suggests it's not a penalty but rather a low quality downgrade that can be resolved easily.

I thought I'd share this since some are getting drastic with their changes. "thin" content took a hit a long time ago, I'd say medium content with only minimal usefulness just did as well. Simply being "unique", "original" and "on topic" doesn't mean anything anymore.

Broad quality downgrades are not exclusive to Panda/Penguin!

[nrep]

Nope, not even a little - it was a big kick down on the 16/17th Nov and then a tiny decline until a few weeks ago, now there's a steady slow decline each day that passes.

[hitchhiker]

@nrep - sorry to hear that.. we're not seeing a daily decline currently, but we did have exactly that during jan & mar. Good news is that they tended to correct given time.

G'luck!

[akishore]

What exactly did you do to "beef up" your posts?

[1script]

What exactly did you do to "beef up" your posts?

This question is especially intriguing given that forums and other UGC-heavy sites seem to be caught up in this more than other types of sites. How do you "beef up" UGC?

[hitchhiker]

@akishore - we did absolutely nothing. Our thinking is any changes beyond what we feel is 'right' is the wrong move.

We didn't have anything wrong with our site, it does its job, serves the community - hence we didn't care that Google didn't like it.

Scary, but 'long term'.

[morpheus83]

So anyone else recovered from the update?

We were hit bad and still struggling.

[1script]

Nope, no recovery at all. Even Penguin 2.0 didn't change anything in terms of traffic amount. I am seeing a strange phenomenon in the last two days (not sure if it has anything to do with Penguin or not) - an almost complete change in the set of KWs that bring traffic. It happened once before, so we'll see if it holds this time. The overall amount of traffic is still the same except that people are now coming on different KWs. Better KWs, I might add, because all conversion parameters are up. But, again, the overall amount of traffic measured in UVs is the same as it's been since Nov 16th 2012.

[nrep]

No recovery here either, even after making a few changes to our sites. It seems to be forum related from what I can tell.

[luckystrike]

I didn't see any recovery - thoroughly cleaned up the site and link profile - disavowed many links - an authority website more than 10 years old!

I've been hit in April 2012 and since then it was month after month of decline. Google is putting money before us and anyone thinking otherwise is a fool. The proof of this is in youtube where it is a little more blatant in its preference for youtube monetised partners. This s all about money.

[hitchhiker]

Google is putting money before us and anyone thinking otherwise is a fool.

That's absolutely true, it's something we've sadly (as a 11 year old authority site) have recently had to accept. We've just dropped again (lost about 10k a day), down 70k/day since nov 19th 2012.

Our competition is clearly using every trick in the book, we've just being churning out quality UGC and it's no longer enough.

Ranking is now paid for, that's all. The last of the big independants are dying out. Long live the corporate brands and temporary fads.

[edited by: tedster at 3:27 am (utc) on May 28, 2013]

[hitchhiker]

Followup:

I wrote a thread on another 'group' - apparently I can't link to that pertinent information here. *Sigh*

A google team member replied to me, and pointed out that I have quite a 'user profile/picture' pages that are pointless (my words not his). He suggested I de-index them.

I'll do that, but I totally disagree. It's not our job to do that, I would expect them to be simply ignored. De-index should be a 'spider trap' exception, not a requirement.

Either way, it was cool of them to respond.

[morpheus83]

You contacted them via the reconsideration form?

[hitchhiker]

No, Goog Group. John kindly replied.

Mods, I shortened it, is this now ok? [bit.ly...]

[Bigwebmaster]

I doubt the profile pages are causing this specific penalty as we have thousands and thousands of profile pages as well, except we have had them as noindex,nofollow long before this happened to us on Nov 16, 2012.

Having them as index,follow could trigger another problem though, such as with Panda, but in my opinion this is not the cause for this update. I think your rankings would have dove on one of the Panda updates if it had to do with your "pointless" profile pages, not this one.

It is hard to tell, but I still believe this specific update may be related to user generated content that falls into a category that isn't bad, but not absolutely fantastic either. Content that is useful and isn't thin, but of medium quality like Sgt_Kickaxe first mentioned. That is the only theory that makes sense to me looking at the type of sites that were affected on this specific date.

[tedster]

Mods, I shortened it, is this now ok?

I didn't realize that John Mueller responded at first. He's an authoritative Google source, so it's very OK. And actually, it's better to post the original Google link so that other members know it's safe to click on.

[productforums.google.com...]

[morpheus83]

Great going Hitchhiker and a very insightful you have there. I hope the changes suggested by John Mueller help you out.

One thing I would want to point out is that - Your pages are missing canonical tags. You should look into that asap.

[Robert Charlton]

I think the reason is because on every page repeats the same text and just change some data (name of city dwellers, zip code, etc.).
Content is of poor quality but the same thing is repeated on every page.
Also I have the same site in another language so Google might consider it doubled.
Any thoughts?
How I can do to recover my visits?

hipotecamix - I think you've mostly answered your own questions by identifying the problems. Either you add unique and useful content, or you suffer the consequences.

I mention your questions right here because you also describe your site as being primarily just images. Several of the pages that JohnMu cited as being of low quality were simply pages of images with no relevant text... essentially just the page template. He also cited text pages with not much useful content on them. Conceivably, this kind of thing may run through a lot of the sites that were affected by this drop.

If you want to rank, every page really has to be good. To quote an important point that JohnMu made...

...make sure that all of your indexed pages are of the highest quality possible and that they are fantastic representatives of your website, in a way that all new users will be drawn in and want to recommend your website to their friends.

Google is going to continue to raise the bar, and either you pay attention to that, or you don't.

[hitchhiker]

@morpheus83 - thanks, where specifically? Most of the urls self 'push' back to their canonical. They self 301 if the usr gets the case wrong, or the url changes.

Cheers!

[hitchhiker]

Those pages might be 'low quality' for search landing. They should be ignored. They are not 'spam' in any way, they are people's profiles, anonymous users frequently use them to get to results they want.

@Robert Charlton - could you read what I wrote after Mu's reply, and tell me if you think I'm 'somewhat' justified or not. We've both been doing this a long time; am I wrong in my 'observations'?

[nrep]

Thanks for posting that link hitchhiker. Just to clarify, were you hit on 16/17th Nov specifically - it may be worth asking John Mu what happened around that date, as it seems like some sort of step change for UGC sites. He's given good general advice, but if he doesn't know that it happened at that specific time then it may not be the reason.

FWIW, we've not got any user profile pages indexed, no printable-thread versions, no archive pages etc... we've been very selective about just getting beefy content indexed, and we've still been hit with that Nov update.

[hitchhiker]

@nrep - Yep, specifically 17th if i remember rightly. And we were great throughout the Panda updates, only Up and UP. I'll try and bring it up again later, please lend your support to that thread.

[1script]

@hitchhiker: kudos for getting a response from JohnMu. I'm sorry to be a party pooper here but I've done the exact thing he suggested - no-indexing of all pages with less than stellar content quality parameters (amount of responses and other visitor engagement the UGC has generated and the overall length of content being the primary characteristics in my case). I've done it almost immediately after the Nov 17th drop (around the beginning of December) on multiple sites and NONE had exhibited any signs of recovery yet. There has been a slight uptick around Penguin 2.0 time but I have to assume that was because some of the competitors have been negatively affected. So, it's not my recovery, it's just the competitors' troubles. Perhaps temporary, we'll see.

So, yes, what JohnMu had said did make sense but the page quality issues have been identified in threads in this very forum within two weeks after Nov 17th and yet even if fixed, there has been no recovery 6 months since.

Additionally, I don't get why this Nov 17th update has not been numbered / cataloged as a Panda run. It definitely bears all signs of Panda (pardon the pun). It's not Panda #21 (too late) and not #22 (too early) and yet it affected quite a few sites, all of them pretty large if I understand the situation correctly. Shall we just call it Panda #21.5 for clarity?

[hitchhiker]

Panda 21.5 sounds good.

Well, I'm making the changes today - I'll write back here in a few months, I too assume nothing much will come of it. This does simply appear to be an algo change that harms the bigger UGC/Forums in general.

My guess: The advice we're getting is meant for 'article' sites / blogs - that's why it's just not 'adding up'.

Please contribute to that thread, it will help get this issue some light. My bet is Google slipped up a little with panda 21.5.

[1script]

My guess: The advice we're getting is meant for 'article' sites / blogs - that's why it's just not 'adding up'.

Why do you say that? It looks like JohnMu had taken some time to browse your site before posting, so you would think he'd recognize the forum format. At least that was my impression. Articles and blogs (except for comments section) are the antithesis of UGC, so it would be hard to think of one thing and give you an advice on something opposite. In any case, forums and forum-like sites are particularly susceptible to poor content quality, so whatever he was talking about (and the entire conversation around Panda issues) is very much applicable to forums.

That said, my own forums sites haven't been affected by any of the official Panda runs. This makes me think that Google did recognize the particular challenges forums face in terms of low quality UGC and has been giving us a pass up until Panda 21.5 at which point they seem to have made a "correction" of sorts to level the field between different types of sites.

I would hypothesize that the improvements in UGC quality (more accurately improvements in the quality of pages with UGC) that you and I are making now might only make any difference at the time of the next "correction", which seems to be a very rare and unpredictable (and unannounced) event.

[hitchhiker]

The Panda 21.5 was also our first 'correction' since then we've been slowly disappearing.

Down about 80k visits a day (since P21.5), not terrible, but enough to be very unhappy about the direction.

If you read the thread over there (it's way too long now) - you'll see that I'm not arguing that some pages are 'thin' or 'low quality'. It's that we (UGC heavy sites) are not going to ever be able to adhere to most of the 'suggested' improvements.

I've just removed 'meta descriptions' and no-indexed nearly all pages except homepage, forum list 1st page, and posts themselves.

This is not something we should need to do (my explanation is within that thread)

Either way, I'm no longer a non-seo guy - i'm back into this damn game after my nice 5 year break from it. The world has one more SEO. That's not what should be happening.

[hitchhiker]

Quick summary of my changes, for reference:

I've removed meta description. Some of my descriptions made no sense (a by-product of trying to create useful summary content from UGC content). So, rather than attempt to create something I can't be proud of or confident in - I've removed them.

I've disavowed most of my referring links that had low cardinality. Ie, one site linking to my homepage a thousand times etc.

Meta de-indexed (MDEd) all tag pages

MDEd all second pages of 'threads results' (list of threads, not actual content)

MDEd all search results (i had mostly done all that before, but tidied it up)

MDEd all image pages (even though some had 100k views)

MDEd all profile/user pages.

Other minor bits and pieces.

[1script]

@hitchhiker: I would NOT remove meta descriptions but rather work on making them more enticing to click at in SERPs. It looks like your site has a rather complicated coding with quite some JS thrown in, I don't believe Google will be able to come up with cohesive or even human readable descriptions for your pages on its own. If your listings in SERPs look awful, noone clicks, Google loses interest in showing them. Traffic falls even further, vicious circle. Don't just drop descriptions, figure out the way to make reasonably well readable meta descriptions out of whatever is considered content (including UGC) on those pages.

As far as changes, I've done all the same except for one thing: deindexing image pages. I would not deindex them, especially if some of them are popular. What was the reasoning, anyway? Too little text? That should not be too big of a deal if they have "alt" tags and there's text on pages that have links to those images.

We'll see at some point in the future if the changes make any difference, or so I hope anyway.

[hitchhiker]

Thanks for your reply @1script..

I used to combine (based on text length and word usage and simple text analysis) sentences that could be considered valid, with other sentences from later down in the thread replies (if those in the initial post were too short). This worked well for many many years.

This is UGC - really important and useful stuff can be only a few sentences long, other times it can be buried in 2 paragraphs of context.

There's nothing I can realistically / confidently do beyond this.

The decision is scary, but I can't justify writing complex parsing software to construct valid description text for Google (in a non-search context) anymore.

Google have the search context at runtime, so they can isolate from the total page content the most relevant part. If they don't then I guess we die off (I don't want to die off)

Also: The pages don't use any javascript to display relevant content.

Cheers and thanks for taking the time to check out my situation.

[1script]

I used to combine (based on text length and word usage and simple text analysis) sentences that could be considered valid, with other sentences from later down in the thread replies (if those in the initial post were too short). This worked well for many many years.

I had a similar approach only I've taken the beginning sentences of the UGC snippets because there would be no good way to find what's valid or not. I have also done it for years (at least for the last 10) and I'm not going to change it because if also works well with Bing and I value Bing traffic very much. It may not be exactly the best description but at least it's not completely built by a robot either, which would have happened if you've removed the meta description and let Google do it for you.

[hitchhiker]

Lol i thought I was the only one, and also for about 10 years. I'll let you know how this goes anyway :)