Panda-loved website: What does it look like?

[EmptyRoom]

Simple question: How does a website loved by Panda look like? On-page/Off-page characteristics?

[tedster]

Simple question with a hidden assumption - that we can spot one type of profile that applies to all websites that Panda likes. That's been the "holy grail" of Google SEO for the past eleven months, and no one has come up with an answer. In fact, it would be the same as reverse-engineering Panda.

When the first Panda iteration was set loose last February, there were several lists released by different companies of websites that gained traffic, the "winners". Lots of people tried to figure it out, but as far as I know, nothing definitive has come of it.

That's mostly because we don't even know what specific metrics to looks at - and possibly because there is more than one profile that Panda likes. If Panda is so complex that Google needed to run it once a month instead of integrating it dynamically, then it is VERY complex, because Google has a massive amount of computing power.

[EmptyRoom]

tedster: Thanks for your answer.

I am sure that many of us have websites that were HIT by Panda, but also have websites that were LOVED by Panda. I'd like to know how does that Panda-loved website looks like compared to the ones that were hit.

[hispdcha]

If it's a Google owned property, it's good to go. If you're a brand that would make Google look stupid for not showing it in the serps, it's good to go. Other than that? No rhyme or reason from what I can tell.

[Reno]

and possibly because there is more than one profile that Panda likes.

Every time someone writes about Panda liking or disliking something, the words need to be understood in the context of: "In my experience, Panda...", because there is no universal statement that stands up under scrutiny ~ Panda is inscrutable by design, which is why Google is so pleased. What works for you will not work for me; my failure could be your success; my success could be your failure. The variables are too complex & unpredictable ~ there is no map or keys to the Kingdom, because if there was, it would take Google no time at all to spin the dials and make that "solution" obsolete. To paraphrase Bob Dylan, "you'll find out / when you reach the top / you're on the bottom".

..............................

[Content_ed]

Simple question: How does a website loved by Panda look like? On-page/Off-page characteristics?

My Panda loved site looks exactly like my Panda hated sites. Same HTML editor (from 1995), same layout, same contact page, same Google Analytics, and as I discussed in the ongoing instant Panda recovery thread, the content is somewhat related to the group of pages from my Panda hated site which I moved there. It had fewer pages than either of penalized sites, but not by a huge margin and I've never heard of that being an issue.

In terms of offsite characteristics, my Panda loved site has far fewer incoming links and they are lower quality links (in the PageRank sense) than my two Panda penalized sites. It's newer, rougly 3 years old vs 11 and 16, and it's only been featured in a major media outlet once, versus multiple times for the other sites.

The only obvious issue I ever turned up is that the Panda loved site had almost no offsite duplicate content, and what was there, I easily cleaned up with DMCA filings last summer after it was hit in Panda 1.1 (the other two sites were hit in both 1.0 and 1.1). The other offsite action I took for the soon-to-be beloved site before it recovered was to eliminate most of the links to it from the most Panda hated site. There were probably a dozen links, left over from a few years ago when I was just launching it and trying to give it some status.

After the beloved site recovered from Panda 1.1, the only action I can remember taking before the love turned into an absolute crush and Google more than tripled its traffic over pre-Panda levels was also related to duplicate content. A company had asked me last year to use some of my content for marketing their product, and I gave them permission providing it never appeared online. Their webmaster forgot to no index a directory that included three of my pages reworked in XML form to demonstrate their product, and of course, Google picked them up. So I got them to take that down, and a little while later, Google got really romantic.

Wait, I think I dropped the links from my other penalized site to the beloved site in the same time frame, all of them, rather than most of them as in the first case. I know I wrote this up at the time on this forum, no idea what the thread name was.

So I don't know if the recovery and puppy love was due to getting rid of two minor duplicate content problems, whether it was due to unlinking from my two penalized sites, a combination of the two, or none of the above:-)

[EmptyRoom]

I wasn't aware that the situation is THAT doomed.

No predictability means very high investment risk.

[Reno]

No predictability means very high investment risk.

That will be true if a person expects to have Google be the primary source of traffic. That's the point of so many threads these days ~ don't expect anything from Google. Diversify, diversify, diversify, and the risk will not be quite as great, though with that said, without Google it is certainly more of a challenge.

......................

[Marshall]

If it's a Google owned property, it's good to go. If you're a brand that would make Google look stupid for not showing it in the serps, it's good to go.

Also include sites that pay a lot in Adwords. Regardless what may be thought, there is a correlation.

Marshall

[hispdcha]

If it's a Google owned property, it's good to go. If you're a brand that would make Google look stupid for not showing it in the serps, it's good to go.


Also include sites that pay a lot in Adwords. Regardless what may be thought, there is a correlation.

Marshall

I was serious. Look at the top gainers in the SERP's and there you have it. YouTube? What does YouTube have that you don't? Traffic? Budget? UGC? Google Property?

Google Shopping results. What does it have over the other comparison sites? Unique content? No. Higher quality? Not really. Google owned? Bngo.

Amazon, Walmart, etc. Why do they win? More money? Check. More brand recognition? Check. Would Google look stupid if they didn't have them in their search results? Double check.

If you don't fall under these rules. You're in the dark and the writing is on the wall. Play fast, play hard, and it's time to get dirty if you haven't already.

[edited by: hispdcha at 6:50 pm (utc) on Jan 8, 2012]

[londrum]

to be fair to youtube, it is the leading site for that kind of thing. it was at the top before google bought it.

sites like amazon deserve to be at the top too. but its a little harder to figure out when they dont have an actual page about the query, and other sites do, and yet they still rise to the top through one of their more general pages. trusted brands obviously get a big boost in panda, even when their subject matter isn't quite spot on.

[tedster]

There is a common theme I often see in "Panda-loved" websites that is at least problematic for "Panda-demoted" websites - and that's trust/authority that is strong enough to stand up to extensive scraping and syndication.

It's not the exclusive Panda focus by far, however content duplicated on other sites is often a big factor in getting demoted. After all, the Jan 2011 "Scraper Update" was considered an important prelude to Panda. And trying to fight duplicate issues head-on can be a losing battle of whack-a-mole. I think it's essential to be proactive and directly build trust and authority signals.

Yes, execute some DMCA take-downs in egregious situations but that alone doesn't do it. Fat pings immediately on publication to Google through PuSH can help. Authorship mark-up can really help. If your site publishes the original content authors, then make some strong claims in every way you can. In the area of intellectual property, you always need to defend your own or lose it.

----

Another area that Panda-loved websites avoid is generating multiple article pages that are very close variations, usually created to target small keyword variations. For years this was a successful tactic - so much so that it became the hallmark of content farms: "How to Drink Milk", "How to Drink Chocolate Milk", "5 Tips for Drinking Milk" etc, etc.

Websites that are filled with this kind of internal near-duplication often run into big Panda trouble. One focused and well-written article is what you need. Stop trying to eat the whole cake!

[suggy]

Panda loved: small, tightly focused, clear topic, varied content, content matches user preferences (format, length, etc), relevant deep links, not over SEO'd, not massive overlap between pages, not sharing lots of same content (eg products repeated).

Panda really loved: Amazonian (!) domain authority.

Panda hated: medium-sized, no strong brand signals, plenty of written for SE content, overlapping topics between pages, wrong kind of content (eg cheap words but limited use of video or pictures), operating in generic (broad) niche, targeting generic search terms - ie opposite of Panda loved.

In other words, you're either big and authoritative or small and focused. In-between is no-mans land, where ferocious Pandas roam!

[EmptyRoom]

What is the definition of "over SEO'd" now? Where is the limit?

Is getting links as important as it was before?

[suggy]

For me, over SEO'd means too much round peg into square hole...forcing big key phrases into pages and using too many related biggies on the same page. You know the kind of thing. When we're seeing pages returned that don't even appear to have the search term in them, this stuff must stick out like a sore thumb.

ps: ditto links with 'optimised' anchor text

[AlyssaS]

Given that there was no Panda in December, I'm of the opinion that issues to do with duplication of your stuff on other external sites (or you duplicating other sites on yours) is a completely separate algo from Panda. I think there was a small change to do with this on Dec13, another on Dec29 and another on Jan5.

But that's just my take based on some observations and they don't constitute a big enough data set to say for sure. Anyone else have a different view on this/has different observations?

[aristotle]

I think that Suggy has done a good analysis, and I agree with most of it.

But I would like to add that another possible characteristic of a Panda-loved site is the ability to make an impact and grab a visitor's attention. Google can use the Chrome browser to measure this through user behaviors such as the amount of time spent on the site, repeat visits, and bookmarking as a favorite. Conversely, sites that generally don't make an impact or grab a visitor's attention might be more likely to be affected by Panda.

[gyppo]

We do a lot of work in the coupon industry so here's my thoughts on what I've been able to see.

We can't see what elements or signals Panda seems to like or dislike:

- The top ranking site has user submitted coupons which are poorly written, not checked & many don't work at all. Their holding company is invested in by Google & it seems they are gaining more market share with every Panda iteration. Uncanny.

- Many of the sites that have a team of people checking coupons, making sure the content is high quality & of value to the user. This doesn't seem to be helping, the user generated sites are prevailing currently. Maybe more diversity in content & writing style?

- Lots of EMD cropping up for "merchant name + coupon" searches, it seems that people are capatilising on the fact that a lot of the large players have been battling Panda to no avail (when I say lots, I mean lots of EMD spam). I've tested this myself, I can get a brand new domain to rank for a competitive coupon phrase within 4 weeks - yet can I can't get my main site out of Panda with 8 months hard work. The annoying thing is that most of these sites are just 1 page of content plastered with ads, no actual working coupons. Yet we're providing this & getting hammered for it. I guess the revenue is going to Adsense that way.

- Most coupon sites have a large number of pages, this could be part of the reason Panda has had a fondness for them (boilerplate titles, snippets, little content on page). We've cut one of our sites down from 55k pages to 500 in the last month, we'll see if that helps with recovery. There are sites out there with millions of pages that follow the same formula as above, they seem untouched (so authority could play a part in how harsh Panda hits you).

I could go on forever but it feels like we're fighting an uphill battle, Google would need special conditions for our industry to get a true view on quality:

Do the coupons work?
How often are they updated? (i.e. daily/monthly)

I've done a bit of data collection across a snapshot of coupon sites to see if there's any pattern (i.e. indexed pages, title structure, duplicated templates, boilerplate text etc), but we can't spot the Panda-related signals.


Mod's note: For those not familiar with the term, EMD means exact match domain. This can be hard to find by search.

[edited by: Robert_Charlton at 6:12 am (utc) on Jan 9, 2012]

[klark0]

I'm looking at one of the largest health information sites. Almost 1 million pages in the index. Almost all the pages are cookie cutter, insert X here pages FULL of content published elsewhere. Panda loves them real good.

[Planet13]

Panda really loved: Amazonian (!) domain authority.

Yes, seeing lots of askville.amazon subdomain pages ranking well - often better than wikipedia or university professor pages where they have written books on the subject.

[Marshall]

Draw you on conclusion from this example. It is based on a real experiment.

For sake of this example, my client's company name is Brand New Widget and the domain is brandnewwidget.com and, obviously, they have the words Brand New Widget on their pages.

If I enter "brand new widget" singular into Google, my client's site is 3rd with two other domains which have the words "brand new widgets" plural on their page first.

If I do the same thing with Bing, my client is first.

So I ask, why when you enter the singular version does Google give you the plural first? And it did not auto-suggest it.

Google is doing what, IMHO, benefits Google, period!

Marshall

[netmeg]

The "Panda loved" sites that are under my purview all have strong commitments to provide the best website, the best user experience, and, in fact, the best business model for their particular niches. There are bigger companies and bigger brands competing across the board that don't do as well. And the business plans are created to be flexible, so that if we lose our unique value proposition in any area, we can find another one.

I don't pick niches that I can't excel at, and I don't take on clients who pick niches that they can't excel at.

Oh, and all my Panda loved sites link out LAMF. We're very picky about it, but if we feel it's of use to the visitors, we do not even think twice about it.

[rlange]

tedster wrote:
Another area that Panda-loved websites avoid is generating multiple article pages that are very close variations, usually created to target small keyword variations. For years this was a successful tactic - so much so that it became the hallmark of content farms: "How to Drink Milk", "How to Drink Chocolate Milk", "5 Tips for Drinking Milk" etc, etc.

Websites that are filled with this kind of internal near-duplication often run into big Panda trouble. One focused and well-written article is what you need. Stop trying to eat the whole cake!

I'm convinced that this is the main issue with the first of my company's websites to be hit, back in February. (I suspect that this might apply to at least some of our other websites that were hit later on, but I'm less confident there.) This site covers a rather broad and popular topic which itself contains a number of subtopics which are represented on the site. One of these subtopics currently has 30+ articles that are really just variations on two themes: "What is a widget," and "Is a widget right for you?" And that's after convincing the managing editor to do some trimming.

A problem, I feel, is that we solicit individuals within the industry to write these articles. (They're not big names, as far as I'm aware, so name recognition isn't a positive that we can claim, but that's beside the point.) Trying to convince the managing editor to do more drastic trimming is extremely difficult because they feel that the "variety of opinions" represented by the numerous articles is valuable. While I don't necessarily disagree, I'm afraid that, outside of the author's expressed opinions, the articles have too much overlapping information.

I'm also struggling to come up with an experiment to prove whether or not consolidating or eliminating these potentially-redundant articles would have a positive effect.

--
Ryan

[enigma1]

Panda hated:.....overlapping topics between pages

Actually to a certain degree I was able to verify this for a couple of cases at least. The end result is bad for the original author and/or authority site.

The way I saw it implemented was by scrapping the original content of several different pages and then creating a single page with all the content in it (and it was plain copy - word by word). It brought the scrapper to the #1 position of the results for very competitive keywords.

[tedster]

I've seen the same - it can be problematic to just start eliminating and combining such content. After all, for quite a few years this kind of overlap was a successful approach for search engine traffic. And in the case of well written content rather than fluffed up pages, it can serve visitors quite well.

But if a site is already suffering, then proper study of the site's analytics is essential. If a page is performing well, then I'd be very reluctant to remove it just because it doesn't measure up to some abstract concept about "overlap". But if the site is no longer performing well as a whole (even though it once was) then surely there are some pages that are immediate candidates for this kind of surgery. Just address those - start there and gradually evolve the site into something more functional and concise.

When a site includes many overlapping areas of content, then what does that do to the navigation? What does the internal anchor text look like? Either there are nearly orphaned pages, or there is a lot of chaos and redundancy that makes things quite confused.

[Reno]

From a pure & simple academic POV, a certain amount of overlap is necessary to well written articles, because different topics have different context, but that context often has its roots in a similar place. So without the repetition of that source material, a reader coming directly into the article would miss an important part of the story.

So when we read that Panda "likes" tightly concentrated thematic content, I say that has an unfortunate side effect, which forces webmasters to stay narrowly focused on the theme and the theme only. Can you even imagine how useless history books would be if historians were held to the same standards by book publishers?

As often as not, to give readers a proper understanding of a subject, it is necessary to go "off-topic", but as the word spreads that Panda wants websites to stick to its central theme only, it's inevitable that the quality of many articles will in fact suffer ~ the exact opposite of what Google said it wanted.

BTW, whether this is literally true is not the issue ~ if webmasters think it's true, it will become reality. Just like link exchanges ~ the word spread that trading links would hurt a site (whether accurate or not), so it virtually stopped. In the coming months/years you can expect to see more & more articles only providing the narrowest context, all because siteowners want to "please" Google... yet another example of good intentions going awry.

.........................

[tedster]

Probably best not to focus overly much on this. What's better is to state it in reverse - do not create overlapping content just to cover keyword variations. Clearly content written to inform visitors is not the same as inflating the number of pages with just small variations in keywords.

[suggy]

Precisely! No more writing 5 articles / pages when 1 would do.

I would also say, don't write content unless you have something extra to add. I think google is trying to discourage the yet-another-very-similar-page on X or Y phenomenon. So, if you can't bring something extra to the content (that's already in the corpus), then you'd better bring some serious domain authority.

[enigma1]

No more writing 5 articles / pages when 1 would do

This wasn't exactly what I saw. Imagine you have a domain for 10 years and you provide services for "blue widgets", "red widgets", "green widgets"
So you have navigation pointing to 3 different pages which rank pretty good for years. Someone comes along takes the content of these 3 pages creates a single one and puts it in a new domain and outranks your pages when any of the page titles entered in the google search.

My point is, if google takes into account the original source it should be able to determine who copied content. Otherwise it doesn't matter how different the content of pages is. Somebody can still do the same. Take several pages create a long page and rank better. It doesn't make sense.