Let's Post Our Panda Solutions - Things That Have Worked

[flanok]

I, like probably many of you have looked through this forum for answers of what we can do to recover from this Panda 2.2 update.

The truth is I am seeing an awful lot of “we have lost this and lost that” and very little of “we did this and got better” results.

I thought it may be an idea to restrict one thread to “Things That Have Worked” whilst we all experiment with content, links and everything else.

Let’s leave all the other chat to other posts, and if you have found something that gave you some sort of return form Panda 2.2 then lets post it here.

I will start will some small gains.

History
My losses consisted of many pages losing 5 to 10 places for key terms. I.e. first to second page rankings, across the board leading to a 30% traffic reduction. But not huge rankings loses that have been reported from other members.

I noticed I had a big issue with existing “SUPPORTING” pages no longer being cached within Google’s index (or appeared not to be) around a third of my site. I got this message when clicking on the cached link per each page. (My key pages that had most rankings still were cached.)

Your search - cache: Mysitepage: Did not match any documents.

At first I thought Google had an issue with my SUPPORTING content, so I moved this section (around 2000 pages to a subdomain, all handwritten over a long period, but in honesty probably lacks real data).

These SUPPORTING pages also did not cache after 3 weeks (only a very small quality were cached).

What I Did
3 days ago, I went into webmaster tools and increased the crawl rate for these supporting pages and in 3 days have seen a dramatic increase in how many pages now show as cached.

As all these pages had important internal links throughout the site to my KEY pages, I believe I am now regaining these internal links to my KEY pages .(or as now on a subdomain these may now be classed as external links to my KEY pages ).

Sure enough this morning I saw not a full return, but saw several KEY pages return back to first page status. There are still around 50% of these supporting pages to be cached, so I keep finders crossed for further gains.

I had also added links from my home page deeper into key pages that had been linked from internal pages, but does not account for all the improvements just some.

This is not a full return but a big enough indicator, for me to understand maybe if it was the quality of content, moving it to another subdomain has helped.

But also there is still a lot of uncached information out there and I do not think we will see the full status until all pages within google are rechached under Panda.

Of course this is all just my opinion only, even if you disagree with my comments and have your own solutions please post them here.

[canuckseo]

@Shatner

on sites where descriptions were changed we have returned to pre-panda levels. Higher in some cases.

on sites with no changes - there's been no recovery

[Swanson]

Shatner, re your comments about the problems discussing this Panda recovery. There is so much "noise" about what people are doing or not doing that it is quite impossible to find a set of things that work or don't work.

In fact I am going to put another theory out there.

Almost everything everyone has been doing has not affected their "Panda scores" - the reason people have seen part recoveries and then fallbacks are because of Panda itself being updated and refined NOT because of things they are doing.

I say this because what I hear from threads like this mirror my own experience. I have sites that got hit - on some i did work on, in fact lots of different things on each - on some i did nothing. ALL and I mean ALL recover on the build up to the next Panda update and then fall back to where they were when it has settled down.

I believe there is an update going on now and that is why the OP is associating work done on the site with a recovery yesterday.

I think that Panda has another element that everyone is missing and that is why all the changes being made do not seem to result in a sustained recovery once Panda updates settle.

Basically I think all you are seeing with each Panda update is dials being turned to refine the algo as well as crunch new data.

I need to test this long term but I think that if you don't fundamentally change the user experience and hold on to your visitors, improve bounce rate (substantially) and return rates then it's just not going to happen.

You remember that list of things that Matt Cutts said they looked at as regards to what makes a quality site and then Panda did his stuff.

What if what he worked on was nothing to do with those actual items - what if they looked at the user engagement data that goes with sites that fit that profile. What if they looked at other data they could measure that correlated with sites that had those characteristics - not those characteristics themselves primarily.

What about that as a theory?

[Swanson]

And to me that would make more sense and require more machine learning than simply looking at pages on websites.

That is why you can't recover because you are not actually making the site any better to Panda you are simply adding or deleting content or pages or a combination and not thinking about the user - you aren't pushing your site into the correct "metrics".

Hey presto the ultimate algo - SEOs mess about for years trying to think of technical fixes rather than thinking about making a great site that people love.

[snickles121]

Well here is a article I found today on using subdomains and Panda.

[blogs.wsj.com...]

I so far have not had any changes good or bad since Panda and the loss of traffic. I gave up last month on trying to fix it. So im just back to writing article to help get traffic back

[Ummon]

Ok I'll Bite

1. When were you first Pandalized?
- With the introduction of Panda. Day 1

2. How heavy was your Pandalization? What percentage of Google search traffic did you lose?
- First Panda Reiteration 2000 visits per day to 1000, second 1000 to 300
3. How big is your site?
- Over 1000 Pages. Blog, Written By me. Pure adult affiliate. No ads on site (only affiliate links) no adsense
4. How long have you been recovered so far?
- This is day 2
5. What percentage of your Google search traffic have you recovered?
- Looks like I'll bust 1000 visits so 50%
6. What kinds of changes did you make (if any)
- 1) Cut back posting from 6 times per week to 3
2) Took a month worth of panda data and searched for any page that did not get traffic from google.
3) No Followed/No indexed all of those pages (Roughly 300). The ones that I spot checked were admittedly rather thin. I did this roughly 2 weeks ago right after panda 2.2

Now I know two days is not a "recovery" I was actually going to wait to mention it until I saw this thread. But I am hopeful since it doesn't look like the pattern I have seen on other false positives. I have another site nailed in the scraper update in January that every now and again spikes and ranks for everything under the sun. Lasts 3 days and gets sent back to where it started. Why It was penalized I have no idea but the spike usually corresponds a few days before a algo change

[walkman]

"You remember that list of things that Matt Cutts said they looked at as regards to what makes a quality site and then Panda did his stuff. "

Did he mention bounce rate? Everyone seems to think bounce rates are it.

Hey presto the ultimate algo - SEOs mess about for years trying to think of technical fixes rather than thinking about making a great site that people love.

Bounce rates are certainly not a way to measure the greatness of a site though.

Thanks snickles121
time to keep an eye on them

[Shatner]

@Ummon Good to be hopeful! Thanks for giving coherent, detailed, great info.

Not to rain on your parade, but just want to share my experience so it may help you.

So basically you're saying you saw a recovery to pre-Panda 2.0 levels, but not to pre-Panda levels.

I was hit in Panda 1.0 and I have had recoveries like that a couple of times now, where I'll get back to pre-Panda 2.0 levels. In every case, after a few days they turned around and eventually things bottomed back out again at my loest post-Panda 2.0 levels and stayed there.

To me it seems like most of the time, the people who recover and stay recovered, get ALL of their traffic back right to Panda 1.0 levels. Not just part.

So it'll be interesting to see if you keep any of that partial recovery as the week wears on. Let us know!

Do you know if the recovery is on older pages, like any specific keywords you track you noticed, or just something overall... or maybe even confined to new pages?

[Swanson]

walkman, you missed my point and I wasn't specifically saying bounce rate.

However, Bounce rate between sites in the site industry and keyword landscape is a valid measurement (not bounce rates from visiting your site from Google search, but bounce rates using session data - I am sure Google collate data from Analytics, Adwords, Adsense, Display, Toolbar - hell, I bet they even buy ISP data like hitwise and Alexa. What about the data they get and use for Interest Based Ads/Remarketing and logged in to Google Accounts?).

Even so bounce rate is 1 of 100+ variables you would use to measure user behaviour. What about comparing visitor return rates during a 30 day period of similar sites? Again, just thats just one thing - hundreds potentially more they could use in addition.

The point I am making is that maybe there are measures of how people engage with sites that Panda is actually intended to work on.

NOT the things that Matt Cutts mentioned - they are the physical and visual attributes that a good site would have. However that results in certain user behaviour and engagement data that might be the actual Panda algorithm.

So what am saying is that everyone is trying to get out of Panda by making changes from that Matt Cutts and Amit Singhal Q & A.

I am saying that could be completely missing the point. Could it be the user behaviour with sites with those characteristics in industry verticals and common site types that Panda is really measuring.

That is why making changes to content, ads etc. does nothing - because you haven't changed user engagement on your site significantly.

Again, it's a theory - but one I am going with, and I won't mention it again in case I find it is correct as I don't reckon anyone believes it!

[walkman]

Thanks Swanson,
it's one of the things we've pondered ages ago as we lay in Panda land. I have something coming up but it makes it a bit harder for the users to be honest.

[Ummon]

@Shatner

After things settled down I was still on page one for my main keyword Just not #1, probably around #8. But that was just for the US. Other area's I was no where to be found.

It seems with this reiteration whether its what I did or what google is doing I have recovered in the UK, Canada, and Australia. Also I am seeing considerably more longtail keywords.

I am not saying you are wrong. Time will tell but as I said it doesn't fit the pattern I have been seeing for false positives but then again Panda is a whole other game

[tedster]

Whether Panda measures user engagement directly or indirectly, improving user engagement is an important part of running any business - especially an online business. Putting efforts here is quite likely to improve your business, period.

The kind of problem Google wanted to resolve with Panda was the type of site that barely thought about their user at all. They were search engine plays, working their code and content with little concern for much else.

Assuming that the HubPages report is accurate, then putting good content in a walled garden "uncontaminated" by junk-food content seems to be a wise step.

everyone is trying to get out of Panda by making changes from that Matt Cutts and Amit Singhal Q & A.

Yes - and jumping to wild conclusions from little or no discriminative reading of that interview.

[walkman]

NOT the things that Matt Cutts mentioned - they are the physical and visual attributes that a good site would have. However that results in certain user behaviour and engagement data that might be the actual Panda algorithm.

Swanson,
Aristotole mentioned something like this a while back [webmasterworld.com...]


Why do you think sites keep going down when nothing's changed, at least not for the worst?

How would you explain the rank of certain pages? Better link juice, content? I doubt Google has enough data to measure "social engagement" for each page. In my case, I rank for the same pages more or less and they have nothing special in 'social engagement.' Not all pages rank all the time but I see the same group even when I get a boost.

I am just trying to see if we find a common theme, not trying to be a prick, since none of us knows, google isn't really talking (not on a language we understand at least) and we're on month five.

[indyank]

Matt cutts did not mention about bounce rates but another Googler did. i have a strong feeling that you people are talking about bounce rates measured from actual visits. But what google might be doing here is deriving the bounce rate and other metrics from the content, for some sites.There is a research paper that suggest that they can do it well. This derived bounce rate takes into account the content quality.

The features are extracted from the content and the model is trained to predict bounce rate from them.

Bounce rate might not be the only metrics that they are using but they may use the predicted bounce rate as one of the elements to arrive at the quality score.

It is the content that contributes to all these predicted metrics and they use some techniques to extract parsed terms and score them.

Source - [bayardo.org...] (4.3 and 4.4)

[MrSavage]

I've really tried to keep my commentary under control the past while. I'm trying my best but it's hard sometimes to stay silent. I'm trying to keep my posting elsewhere, but since we're talking solutions, I need to say something.

Netmeg linked a great article.

Quote #1:

Vanessa Fox: I talk to many people who have sites that have been hit and I certainly sympathize with their plight. However, there is no quick fix in these cases.

Quote #2:

Vanessa Fox: Yes. It’s hard to hear because this is affecting people’s businesses. I think it is going to be a lot of work to figure out who your audience is, what they are they looking for, are you engaging them well, and are you providing value beyond all the stuff that we talked about. It is a process.

==============

I'm just going to speak to Panda solutions. I would strongly suggest that my opinion is no more valid than yours and yours no valid than mine. Unless you work for Google you won't have any clue what is going on. Everything Panda solution is based on speculation with all due respect. I think it's foolish to take any apparent solution and base your strategy on it. That's just my opinion of course. I'm not trying to sound negative here.

Yes it's great to think that I've recovered from Panda. Indeed I had, for 5 days only. The fact is I can't take any recovery at face value right now and the reason is simple. In my eyes, Google search is a cluster F for results right now and my sites are thrown into this clothes dryer and it's spinning. The real issue is, when will the spinning stop? At that point? Then I'll dig into solutions if any exist. When everything is tossing and turning and the ground is up and down and all over it's essentially madness. Yes there are some areas seeing no effects, but who knows. They haven't been visited by Panda yet? These niches won't get targeted because they aren't deemed important enough right now is my guess. Think though, your sites that are okay right now. The wave isn't coming?

To me Panda is like the boogie man. If you consider those quotes, they sound a bit discouraging to me. It's almost that I'm reading that there is no quick fix, you had what you had, but now you need to start over with your site, brand it, built it differently. Why? Well because there is a brand new criteria for judging content called Google Panda. I want no part of it quite frankly.

There are the brightest minds regarding websites and SEO who are posting here. On the planet! The honest reality is that nobody can say what the hell is going on. Panda false positives on their implementation and false positives on the recoveries. See the madness? That's what it is quite frankly. I would suggest sticking with the official Google ideals about websites under the Panda way. It's pretty darn flower sniffing to me, but it's all about "trust". That's the key message. Get trust how?

The reality is, there are a good number of webmasters who don't even know what's up. I showed up here about a month ago just learning about Google Panda. So for the average webmaster, they may have no clue just yet.

Some people hit, others missed and the answers we won't know until the onslaught settles. If it doesn't settle? Then I need another occupation.

Panda was introduced to improve search results. We can all look and try and take our bias out and decide if the results are better or not. Ambitious plans and are the results there? In my situation the best solution is to send Google a message that I'm using Bing 50% now and it's now my homepage. My solution is telling my wife to use Bing. Need proof the Panda world isn't working? How about a drop in usage? That might be a clue.

In closing, my solution to Panda is read their 15 or 20 commandments about trust and about what Panda is valuing. Read every snippet that insiders are saying about Panda. I think figuring out Panda is like somebody saying they have figured out how as humans we got here on earth.

I'm trying to contribute and I hope this post is taken as that. The solution is to put your nose up at Panda, the implementation and the impact on yourself and your business.

Panda is essentially taking the currency which we've been using since the beginning. It's saying, that's great, you have a bunch of money there. Okay, here's the Panda buck. This is what is valuable now. Go, seek answers my friend. The ones with the Panda buck to start are the obvious choices, you know. The giants of the web. Sure it's not all about big brands but where I look that's what I see. I don't see Johns blog about blue widgets right now. So there is a new currency. That's what I think. What you had? Enjoy it. Google changed the currency and I don't think that can be denied thus far.

Not feeling effects of Panda? Enjoy your niche because that's what's left for organic traffic. Enjoy is because everyone is coming for that niche because that's what's left of the pie.

[whatson]

@ MrSavage - some people's opinions are more credible than others, some have a lot more sites and are able to determine more similarties between their Panda'd sites and non-panda'd. For example, if all my sites had hit Panda, then I would be totally lost, but some have been un-affected and I use these to determine what to do with my other sites.

We can say what is going on, as it is obvious, Google simply want to improve their search quality, and increase their stock price, which it just did - massive jump.

There are certain assumptions we can safely deduce from Panda:
It is an algo
It is sitewide
It updates about once a month
If you have duplicate content you are affected
If you have thin content and just putting up something very basic to have a page for a specific search term - you got caught
If you simply make sites to rank as your main focus and not to provide some benefit to visitors, then you are playing the wrong game
It is likely to get out of, but unknown how long the "penalty" will last.
Panda is here to stay

Oh, and unless you have your traffic higher than pre-panda times, and ALL your positions are back to where they were previously, then I am afraid that you are still in Panda jail. Although slight recoveries might demonstrate to us that there is an arbitrary scale to a Panda penalty, only PR (Panda Rank) 0 is a good thing.
If you want to read between the lines of what they were looking to do, then I suggest reading this article, [seogadget.co.uk...]

Also, here is a theory. Google obviously realizes webmasters are trying every change to their site under the sun, and when someone does the correct changes, do you really suspect that Google want them to know which changes worked? i.e. they cant take them out of panda immediately, or otherwise their secret panda recipe will be revealed. It has to be kept under wraps so it cant be gamed. This is now their major IP, and essentially the only thing that separates their SERPs from Bing.

[suggy]

It is sitewide

Not sure what this means. If you're saying it universally affects all pages in your website, then my experience suggests you are wrong; whilst I lost 70% of my traffic in April Panda, I have pages that still rank #1 or #2 for very competitive terms.

[Crystal Pegasus]

Me too re site wide... one whole sub-directory of our main site has gone up in traffic numbers (not by a lot, but definitely not down) both when compared to the previous 3 months pre-Panda, and for the same time period last year, so it's not just seasonal (it's not really a seasonal topic anyway) and remained there consistently throughout the whole Pandalised period.

And in fact one page definitely still ranks first for a few search terms as it has for a very long time.

All of the rest of the site (several sub-directories) has lost about 50%-70% of traffic. It is not the same for all sub-directories, but they have all lost significantly.

The difference between the stable sub-directory and the losing ones is that it is spot on the site's specific topic, while the others are much more peripheral and some don't really fit at all well. All articles there were written solely to be useful and engaging to the user and most are pretty significantly different from much else around of a similar nature.

They were also not written with the expectation of making money from them- and we make hardly anything from them, so as far as income goes they don't really help... so far. (plans afoot, being very cautious of the fine line!)

So, that's just how it is on our particular site.

[indyank]

All articles there were written solely to be useful and engaging to the user

how do you really write it to be useful? in other words how do you think they are useful while others are not in comparison?

[edited by: indyank at 7:28 am (utc) on Jul 15, 2011]

[indyank]

suggy, as far as I see, it does affect all pages site wide and it is more keyword specific. A page continues to rank high for one keyword but not for the other one that used to rank as well.

[Crystal Pegasus]

In my view, the articles in this section were written with a true love and passion for, and knowledge and experience in, the subject, and are basically quite different from most of what I have seen out there, as they were my own individual ideas.

The other sections did not have this benefit, and were more written from conglomerated info, and therefore much less unique and had been covered in a similar way by many others.

Hence, in my opinion, less useful to users (on our site) than the articles that were much more unique, as they could have found similar articles in many places- though not identical.

I guess it's not that the actual subject matter is not useful itself, more that it's just rehashing stuff, so users had probably seen it elsewhere in their searching on the topic- kinda like "oh ok, more of this, nothing new here... moving right along" so high bounce rate, not staying on site long. While as the other section might be more like "hey this is different! wow I'm gonna try doing that! and what else do they have that might be useful too?" so they stay on pages, maybe copy or bookmark what's there, look at other related pages.

Oh... one other thing is that, probably because of the more uniqueness of these articles, they have a lot of and better links (including some .gov and .edu) than the rest of the site.

That's just my thoughts on what might be behind it anyway... I'm not very tech savvy though, so forgive me if my ramblings don't make sense :)

[nickreynolds]

Can I say I found the first few pages of this thread helpful - however accurate or inaccurate their conclusions and however long lasting their recovery there were some useful hints in there.
The thread now seems to be degenerating into (another) general Panda discussion.

[Ummon]

Well I am well into day 4 of my "recovery" my next plan of attack should this hold is collect another month of data and nofollow/noindex accordingly

[suggy]

@Ummon

I like your style/ plan, but I can't nofollow all my product pages (which seem to be buried), so I think I'll use your process of discrimination/ elmination but then try to figure out why Panda hates the pages in question!

[sanjuu]

Following on from the 'partial' recovery I've seen for the UK site I mentioned earlier, here's some information:

1. Natural traffic now at 60% of what it was pre-Panda (so 40% down). It was 80% down until a few days ago.

2. Massive movement in rankings, with literally thousands of phrases now ranking on page 1 (as they were pre-Panda), instead of page 3+ where they've been since 11 April.

3. Still seeing the big 'brands' at the top two or three for most results, but seeing subjectively better results now across the rest of page 1 of results - fewer 'odd' results and no US or Australian sites that we had been seeing post-Panda for UK-focused searches.

4. The rankings and traffic coming in to the site is across all the different sections - so it definitely appears to be a site-wide change rather than a boost in one specific section.

5. As I mentioned earlier, dramatic changes were made across the site, but until a few days ago Google didn't seem to acknowledge any of them. The question (that can't really be answered) is whether this 'recovery' is down to the changes we made, or if Google's update to the algorithm would have seen the same 'recovery' without any changes. The timing (almost exactly 3 months to the day) could be coincidental, could just be the frequency Google's updating, or could also be some sort of 'timestamp' placed on pandalised sites and during the 3 months the site is in some way 'tethered' or held-back and only after this timestamp has expired does Google recognise any changes to the site (and the removal of anything that is a panda factor).

6. This is only Day 3 of the partial recovery, so we're not counting chickens yet. We could easily see these rankings fall again (or possibly further improve), and are very aware that changes we make could also trip the site up and send the traffic plummeting again. Until there is sustained recovery, I don't think any conclusions can be drawn as to the work we did post-Panda being successful. Even if this recovery does remain, the question of whether the site would have recovered in the same way without the changes we made would still be there.

[canuckseo]

@sanjuu I was wondering the same about my changes however I'm confident I am seeing a LARGER recovery on sites that I have made changes to. certainly some of the sites which haven't changed have also made slight recoveries however the ones which I rewrote descriptions for are showing larger changes. It's only been a month or so so I can't say for sure that they will stick but i am hopeful.

I think as everyone is aware this testing we are doing is not done in a vaccuum therefore no one can say for sure "yes this worked" or "no it didn't work" but from what I'm seeing, 4-5 weeks in, changes I've made on my sites is having more of a positive impact (IE Traffic and ranking growth) than unchanged sites which are very similar.

[suggy]

suggy, as far as I see, it does affect all pages site wide and it is more keyword specific. A page continues to rank high for one keyword but not for the other one that used to rank as well.

OK, I can buy that. I've spotted some examples for us too.

We sell a product from brand X which comes in different types and designs.

If I search (clean/ non-personalised) for Brand X Design Y Type A we are lost in outer space.

However, generally, if I type Design Y Type A we are just a bit lower than pre panda (so only showing site wide demotion effect).

So, we are pandalized for the brand.

Design Y is often a made up word/ or phrase that's pretty unique.

Now for the interesting part; I think the pandalization stems from links. We did some intensive low grade link (articles, lenses) building we did around the brand (I know but it was a minor brand and quick and easy).

[Planet13]

Now for the interesting part; I think the pandalization stems from links. We did some intensive low grade link (articles, lenses) building we did around the brand (I know but it was a minor brand and quick and easy).

If that were the case, then ALL my competitors would be Pandalized. That is all the links they have. Low grade spam.

Maybe just whatever low grade links you had were devalued, not penalized / Pandalized.

[suggy]

If that were the case, then ALL my competitors would be Pandalized. That is all the links they have. Low grade spam.

That would be acceptable then, as the norm for those searches.

If you think of Panda as a mchine-learning exercise in spot the difference (and similarities) then, if everyone's doing the low rent stuff, it doesn't matter. But, if doing the low rent stuff puts you on the other side of the line with the miscreants, then you're in trouble!

[clickinc]

Hi all ,

I am a long time lurker , been on the SEO business for about 8 years. Never had a real problem with any algo until now...

Since Panda 1 we lost 70-80% of traffic for one of our main clients , after much reading we have changed a lot of things on the website.. with a method

We are very confident on what we did , in 3-9 days we will update you (depending if we got traffic back or not) ..stay tuned.

Johan from clickinc

[dickbaker]

Clickinc, will you keep us updated regardless of whether you get traffic back or not? It would be useful to know what sort of efforts seem to have no effect just as much as what efforts do have an effect.