Is there a pattern to Panda demotion?

[dibbern2]

I can't see any regularity to the depth of Panda's reranking of my pages. Some went from top 5-10 to somewhere in the twenties, another group slid into the fourties, and a third group got smashed, they are in the 100-300 spots. (There is also a small group with NO demotion, or perhaps one or two places.)

Is this typical of what others are seeing?

[indyank]

maxmoritz, I know that because none of the theories have helped anyone so far and I too don't have anything left now.

Since I don't see a pattern in this like you, I am suspecting google's manual tools in play.

But yes, every point of yours makes sense as you seem to have tried everything and from how you explained it, I could easily see an expert in you.

Suggy definitely has a good theory but we are still scratching the surface.

[edited by: indyank at 5:26 pm (utc) on Jul 5, 2011]

[maxmoritz]

I do think there is something to Google wanting to generally show certain types of pages for certain types of queries. Obviously transactional queries deserve top results that are mostly transactional. I just don't think that was a primary Panda factor. And, I think that can be trumped by site authority/trust. For example, I've seen numerous Home Depot pages with "product guides" ranking for transactional searches. Obviously Home Depot has realized that they can leverage their site authority and they're creating these pages specifically to rank them. So that's an example of information pages ranking for transactional queries due to site authority...trumping any page/query matching factor.

Again, I do think it's something that's important to consider for both Google AND users, but I don't see it as a primary Panda factor.

[mrmobility]

What if panda resorted the SERPs, but user behaviour is the route to recovery?

I was a panda winner, going from 9 to 4 on our main keyword and getting a bump on most of the long tail. Since that initial bump we have been alternating 4 and 3, then 3 and 2, now 2 and 1. It looks like Google is trying us out with different rankings to see how the users react (we have done no SEO for six months so it isn't improvement from traditional methods). If that IS the case they could do similar sorts of things further down the SERPs. This would mean you would be less likely to make a huge recovery, instead having to beat the sites above you one by one.

[maxmoritz]

Since I don't see a pattern in this like you do, I am suspecting google's manual tools in play.

I'm not sure what you mean by "manual tools", but consider that there are a near infinite number of queries and millions if not billions of sites. Whatever Google is doing, it's got to be "algorithmic". Of course they could have white/black lists, but otherwise this is an algorithmic issue...with Panda being something in addition to the regular algorithm, obviously, and as Google has confirmed. Anyway, it's time for me to get back to working on my sites and tests as this speculation isn't really going to go anywhere without data.

[SevenCubed]

The simplest way to try to figure out what happened is to set up a spreadsheet. For your targeted key phrase gather a half dozen sites that currently rank for that phrase. Pick them apart according to whatever criteria you have determined is meaningful and compare it to your site. But that in itself is not enough. You then need to have something to compare it to so now you also go to the page your own site is on and grab a half dozen sites from that page and build them into your spreadsheet. Most likely you will then get a better picture of the underlying factor of difference. It's not good enough to just pick apart the tops sites.

[maxmoritz]

I was a panda winner, going from 9 to 4 on our main keyword and getting a bump on most of the long tail. Since that initial bump we have been alternating 4 and 3, then 3 and 2, now 2 and 1.

That kind of thing happens all the time. It could have something to do with usage data, but more likely has to do with time and or the reenforcing affect of a quality site ranking (gaining natural links/mentions/etc.).

As I mentioned previously, both of my sites that were hit were objectively "better" than many of the sites above and below them that were not hit. AND, the changed I made to my e-commerce site decreased the bounce rate by 50% and page views per visitor by 40%. There is no doubt in my mind that my usage data is FAR, FAR better than most of the sites above mine. Yet, no soup for me. I'm 95% sure this has nothing to do with usage data.

[indyank]

sevencubed, I have tried that too.

But then there will be a few pages which would remain on top, despite most factors prove them to be of low quality.I thought that they would disappear as soon as panda catches up, but to this day they remain there and it is 4 months.

[maxmoritz]

The simplest way to try to figure out what happened is to set up a spreadsheet...

Myself, other people I know, and many, many others have done that and more. Trust me, there is no simple way to figure out what happened. There is no "one thing" that happened anyway. The SERPs are most definitely based on a complex combination of factors and we may never be able to point to exact causeS of Panda devaluations.

[indyank]

maxmoritz, what is the avg.time on site like?

[maxmoritz]

maxmoritz, what is the avg.time on site like?

4 minutes and 11 seconds, the past 30 days. But I think numbers like that are largely irrelevant as they'll be different for different types of searches.

[indyank]

After about four months my search engine traffic percentage had come back to 80+ (same as pre-panda) but no signs of improvement in the overall traffic yet. Post panda, the SE traffic percentage was around 70 (varying between 69.5 and 70.5).

[edited by: indyank at 5:44 pm (utc) on Jul 5, 2011]

[indyank]

But i think if metrics is in play, we will have to see them as a whole and not in isolation as bounce rate or pageviews.

[SevenCubed]

In my example above your comparison criteria is what matters most. Google claims to have 200 such, Bing says they have almost 1000 factors. For those saying they tried it, how hard did you try? Break it down too, onsite factors vs offsite factors.

But more than anything I have to say what I keep saying. It's too soon to put in all that effort because this thing isn't settled yet. There are still too many results in tops positions that do not belong there according to Google's "what's best for users" mantra.

[dickbaker]

After about four months my search engine traffic percentage had come back to 80+ (same as pre-panda) but no signs of improvement in the overall traffic yet. Post panda, the SE traffic percentage was around 70 (varying between 69.5 and 70.5).

I must not be understanding you. How could your search engine traffic percentage come back to 80+ (same as pre-Panda), but you've had no improvement in overall traffic?

[MrSavage]

I've basically said this before but if there is a pattern and somebody figures it out, the Google has another problem on their hands. All you need to look at is there official what is Panda looking for page. It's a list of website qualities. It's really cheesy, but it is what it is.

Part of me wonders if this is a ploy to clean up the internet. Perhaps it's getting too crowded for Google and they need pages removed and content removed so they can do a better job. I suggest this theory because people are doing this all over the place. I mean seriously depleting their site that was good for years, and suddenly because a POS in Google's eyes. A bit of pruning is good. If that's all it takes, then wow. Mission accomplished then.

I also have a theory that Google is wiping the slate clean. You might get some previous rankings back, but in a gradual way. All the authority sites rank above all right now in a lot of situations. Perhaps those pruning jobs we're doing will be evaluated and you will pop back into place. I would have some real reservations about dismantling a website based on a guess on what Google wants. It's a lot of blind faith and have you considered that Bing could be on par with Google and that you are going to reduce your Bing rankings in doing what you're doing.

[suggy]

On the testing front, I am starting some now.

We were in the unfortunate (come Panda) position of having legacy articles dating from before we were even a commerce site rank for big searches. It as a pain, but we managed to move60% of people on from these and into the store. Panda destroyed these rankings.

Naively, in response I bumped the word count into thousand-plus territory and put at least 3 unique images (to site and web) on each page. It was all good content, not spam. Panda demoted further in June update.

Over the weekend, after my personal revelation, I changed the CMS template on these pages to the store catalogue page template, complete with products, AJAX product winnowing, add to basket buttons, etc. Ditched the existingcontent and wrote about 150 words afresh (as per competition). Dropped all the links to other long articles SEO'd for keyphrases. Only content links kept were to blog. Added Verisign, McAfee Secure and Safebuy 'click to confirm' links.

This puts these 4 or 5 pages more inline with the one page we still have doing spectacularly well for a big search. Only difference is it doesn't have the assurance links and it has slightly longer copy (maybe 350 words). Interestingly, it NEVER linked to the article pages, unlike many of my pandalized sections.

I am wondering if one of my negative signals was too many pages about too closely related searches/ topics all interlinked. Looked too much like the infamous x number articles on nose picking (all targeting different keywords). In short, my content silo is now a spam silo, as far as Panda is concerned! In response to this I also deleted some of the more closely related pages, to leave one good one.

Anyway, I'll keep you posted as to what happens next update.

Suggy

[dibbern2]

A few panda pages in my group that lost the least (5-15 places) have jumped into first and second place today. The search term claims 6.5 million results.

I know this is not significant for just a few pages, for just one day --probably noise or perhaps short term testing. Am reporting it only to see if others notice anything odd.

[walkman]

After you do X and Y and Z, wait several months and compare it to those ranking or not pandalized it's obvious what's happening. It's easy to say that G has found a super-secret-top-notch formula...blah blah and must still something be wrong, until you see what's ranking. That explains the irrationality, remove 'bad pages' and ranking goes down. Or you add content and rank goes down. You are sending new papers to the teacher but he isn't reading them--yet, so you go nuts thinking you still got it wrong.

The rest is like 'God sent us this volcano because were sinners. Maybe if we did this or didn't do that....' Yeah OK.

[indyank]

dickbacker, This is another unknown mystery for which I am trying to find an answer this week. There seems to be a significant drop in direct traffic but the SE traffic is also down.I thought things would become normal after the weekend but Tuesday doesn't seem to be normal.

[suggy]

I noticed something new yesterday that might contribute to the understanding of Panda...

We sell this product by brand X which comes in multiple designs. The designs all have phrased names comprising English words and often a made up word.

So the typical search is: (brand X) (name phrase) (generic for item). The generic is an everyday item. When I conduct this search, we appear in the Googlebase/ merchant centre shopping results, but the product page is buried in the organic results. Google would rather show one of our category pages with that product (and many others) on it, or a mention on our affiliate blog or PR image database, than the actual product page.

Here's the interesting bit, when I remove the brand element, we shoot back up to our pre-panda positions on page 1, generally right after Amazon.

So, it appears the brand name is tipping us into the Panda abyss? That's even though the design name phrases are quite lengthy/specific and often include made up words. But, they maybe too far down the longtail for Panda to have reached yet?

This makes me wonder whether these pages are sending out the wrong signals (spam/ over optimised) for Panda and the pivot point is the brand (it's a niche brand).

The pages all have unique content and we write rather more than our competitors. They are fully optimised from an SEO point of view -- bread crumbs, urls, image urls, h1, h2, alt text, related links, related products, strong, title, etc. Maybe just too perfect?

This is happening on a website which can still rank category pages in the top 3 for some really big shopping searches. We are also a retailer not affiliate, no anti-affiliate bias likely.

When I look at the pages that are ranking for these searches, the SEO is much more piecemeal/ incomplete. Nobody else seems to be going the whole 9 yards.

Anyone notice similar or have any thoughts?