Panda Hangover - Why Things Are Going Well This End!

[celente]

Hi Warriors,

I see alot of people writing and panicing about panda slap, and how they are not getting as much traffic and how google is the biggest A**hole on the face of the internet.

I want to you to think for a second. Since the emergence of bum marketing, many article writers went to town on it, found out more ways to exploit this until the internet was over run with quantity, and not quality. something google hates.

I have seen some articles out there that I read the first sentence, and start laughing so loud I nearly wet my pants. WHY? its like someone learning english for the first time wrote it. Words muddled, makes no sense and just really a waist of internet real estate if you ask me. SO I do not blame google for doing what they did.

If you are going to do article marketing, this is changing, I would do the following :-

1) Don't spin, that is the clear message that is being sent to you. You are not really going to get good results. I have one word for you JIBBERISH, ok you might have the kickass best spinner out there, but who really cares. Its not original content the second you hit the "spin it" button is it.

2) Write longer high quality articles. My articles that are 800 words long that are high quality, pull in more sales and more subscribers than 30 odd 400 words articles that are just written quickly and not much info at all.

3) Think the word FRESH. Update your blogs / websites / web 2.0 more often. Write quality, write to help people. Focus more on this.

4) Go to the sources and websites that already have high quality traffic flowing and ask yourself how you can get more of that to your site. i.e. comment on high traffic blogs, Guess post on the the top blogs in your niche, see if you can get some media buys or banners on large networks or websites that rank PR 1 - 3 for the buyer keywords.

5) Post less "high quality" articles if you use this technique. Quantity will no longer help you win the race. Remember Google is after the highest quality... I now have the mentality that LESS is actually MORE if you know what I mean. Post more fresh, original and high quality info or outsource this to people that can do this for you. And you will be rewarded.

6) Get more high quality backlinks. The sites we have good backlinks for are actually ranking higher and getting more traffic after the panda slap. Do not just go to fiverr and pay $5 for someone to spam the crap out of a forum, again 20 high quality backlinks from high PR sources is MUCH BETTER than spamming a forum with 6000 profiles every day. So keep that in mind. Google does not like this.

The internet is changing, it always does. We just need to adapt around these changes. Once someone finds the new BUM MARKETING gold mine, (and who knows what that is) people will grab hold of it and go to town and abuse that too, so be ready when the changes come for that too.

Oh and to all you people that write to me, and explain how spinning articles is still working, you need to grow up and investigate why all the changes are happening and read GOOGLE rule books 101.

They are always after content, High quality content, not jibberish that you keep putting out there, even after the panda update. So no need to start debating in here. Just adhere to the rules. Go out and actually write stuff that is going to get you subscribers and also help you get more paying clients. Heck you might even make a sale if you start writing quality. LOL (winks)

Take a stab at what the next bum marketing traffic technique will be? but whatever, don't just rely on one traffic source to bring visitors to your site. That way you wont see huge drops in traffic when another update comes (because they will come)

Good luck.

[dburk]

Hi celente,

1. Trust me, you don't need a spinning tool to create "JIBBERISH". It has nothing to do with whether you spin, or not. That's is just a myth. You cannot tell a well crafted spun article from a manually written article. It's not the tool, it's how you use it. Spin responsibly and you get better results, not worse.

2. I haven't seen any indication that longer content has fared better under Panda. There isn't even a correlation that I have seen.

3. My websites that I haven't updated in years have risen to top spots after the Panda Update. Again I see no correlation in your assertion.

4. sound advice.

5. This advice seems to contradict your point number 2. I agree quality is important, but I see no reason you should limit the quantity of your high quality. I would recommend that you not choose one over the other, choose both as the best option.

6. I strongly agree with this advice, don't spam, it never lasts anyway.

Yes, the Internet is always changing, but the core principles of SEO has remained unchanged. Useful relevant content is always the key ingredient.

As they say, content is king!

[cagliostro]

I don't really know ... I think i don't care about what Google Panda (or whatever) wants me to do, i will do what people need. And that is not up to Google to decide.

[Johnsonton]

Those are really great tips. Using articles always helpful in terms of SEO for sure as far as most of the search engines prefer contextual link building.

[Grindstone]

Originally Posted by dburk

Hi celente,

1. Trust me, you don't need a spinning tool to create "JIBBERISH". It has nothing to do with whether you spin, or not. That's is just a myth. You cannot tell a well crafted spun article from a manually written article. It's not the tool, it's how you use it. Spin responsibly and you get better results, not worse.

Agreed, spin it right and it's 100% unique and 100% human readable. Takes time tho, I personally like to just write unique content these days.

2. I haven't seen any indication that longer content has fared better under Panda. There isn't even a correlation that I have seen.

Data derived from close to 20k page one results being analyzed (via our tool in my sig) disagrees with you on this one, however:

[dburk]

Originally Posted by Grindstone

Data derived from close to 20k page one results being analyzed (via our tool in my sig) disagrees with you on this one, however:

That is interesting, but it is an average, where are medians? How were those 20k keywords selected? Are they consistent when you use a subset or repeat with a different set? By averaging the word count you are building in a bias towards longer content, a handful of pages with very long content will move the average up significantly while the shorter content may fare far better.

I can tell by looking at your chart that your data isn't made up of highly competitive keywords of commercial value, as you typically see word counts that high mainly on obscure documents that almost never rank highly for competitive keywords with commercial value. Perhaps you data includes mainly blogs which tend to include the last 10 posts on the homepage which spikes the word count.

More importantly, did you do a Pre-Panda update vs Post Panda Update comparison?

[bobrittany]

Thank you for the good advice. I completely agree with you. In fact, I am actually quite happy with the Google update and my sites have seen huge increases in traffic. I was getting really annoyed when every time I googled something in my personal time the first 5 google results were so obviously internet marketers....its not cool. Provide real information for people, and they'll buy. Thats what its about.

[wfstin]

Originally Posted by dburk

That is interesting, but it is an average, where are medians? How were those 20k keywords selected? Are they consistent when you use a subset or repeat with a different set? By averaging the word count you are building in a bias towards longer content, a handful of pages with very long content will move the average up significantly while the shorter content may fare far better.

I can tell by looking at your chart that your data isn't made up of highly competitive keywords of commercial value, as you typically see word counts that high mainly on obscure documents that almost never rank highly for competitive keywords with commercial value. Perhaps you data includes mainly blogs which tend to include the last 10 posts on the homepage which spikes the word count.

More importantly, did you do a Pre-Panda update vs Post Panda Update comparison?

I'm 99% sure that chart is from serpiq.com internal database of user submitted keywords, which is a tool used almost exclusively by internet marketers so I think it's safe to say that nearly all of those 20k keywords have a commercial intent.

[RevSEO]

This post has some very sound advice across the board.

Posts like this really make me happy as you are getting WF members on the right track.

I completely agree with what you are saying and thanks again!

[Grindstone]

Originally Posted by dburk

That is interesting, but it is an average, where are medians?

It's in the neighborhood of 1500.

How were those 20k keywords selected?

Input by (mostly) paying clients. We have a free trial that allows 5 keywords per user, so there's a small % that might have been throw away queries.

Are they consistent when you use a subset or repeat with a different set? By averaging the word count you are building in a bias towards longer content, a handful of pages with very long content will move the average up significantly while the shorter content may fare far better.

Law of large numbers - Wikipedia, the free encyclopedia

I can tell by looking at your chart that your data isn't made up of highly competitive keywords of commercial value, as you typically see word counts that high mainly on obscure documents that almost never rank highly for competitive keywords with commercial value. Perhaps you data includes mainly blogs which tend to include the last 10 posts on the homepage which spikes the word count.

Yeah, no. It's averages from the FRONT PAGE of more than 20,000 google queries that our paying customers felt competitive enough spend their hard earned cash to investigate. I can't disclose any client info but all the terms I pump into it daily are very competitive. I only play in the big boy niches: pharma, diet, insurance ecig, reputation management, forex, stocks, make money, etc.

More importantly, did you do a Pre-Panda update vs Post Panda Update comparison?

Considering there have been at least 5 releases of Panda, this isn't the right question to ask. Do the averages change through each version would be a better question. What about the medians? Yes, but not significantly.

[bigcat1967]

My articles that are 800 words long that are high quality, pull in more sales and more subscribers than 30 odd 400 words articles that are just written quickly and not much info at all.

I have a writer and I tell her no less than 1,000 words per page. Panda scared me to much...

[dburk]

Originally Posted by Grindstone

It's in the neighborhood of 1500.

Input by (mostly) paying clients. We have a free trial that allows 5 keywords per user, so there's a small % that might have been throw away queries.

Law of large numbers - Wikipedia, the free encyclopedia

Yeah, no. It's averages from the FRONT PAGE of more than 20,000 google queries that our paying customers felt competitive enough spend their hard earned cash to investigate. I can't disclose any client info but all the terms I pump into it daily are very competitive. I only play in the big boy niches: pharma, diet, insurance ecig, reputation management, forex, stocks, make money, etc.

Considering there have been at least 5 releases of Panda, this isn't the right question to ask. Do the averages change through each version would be a better question. What about the medians? Yes, but not significantly.

Hi Grindstone,

So are you saying that the median for Position #1 is 1500 words while the average is above 2500?

Do you have any idea what percentage of those sites are blogs? The reason I ask is because the median size for blog posts are around 250 words and and the typical blog is configured to show 10 posts on the home page. That makes your stats about average for a database of blogs.

I have my own database of 6146 #1 ranked websites and the average word count is 718 words and the median is 649 words. I'm including all text, including navigation, sidebars, etc. I would really like to know how your database is so different? I'm guessing that your dataset includes a high percentage of blogs, article directories or something of that sort. While I have a very low percentage of blogs, it's mostly brick & mortar retail, e-commerce, manufacturing, large chains, service providers and such, no article directories or white papers.

[paulgl]

It's funny how the more things "change," the more they stay the
same. Good SEO has not changed. What has changed, is google
getting serious(?).

I find it quite odd that people are still trying to figure "panda" out.

There was nothing to figure out, actually.

Get a grip people. A kung fu panda grip...

Paul

[Sabrina178]

the rule of SEO will always be Content is King, Technique is Queen

[Grindstone]

Originally Posted by dburk

Hi Grindstone,

So are you saying that the median for Position #1 is 1500 words while the average is above 2500?

I'm not. If you reference the graph I posted, it shows positions 1-10. The median for those is ~1500 and the highest, as clearly shown on the graph, is still sub 2500. I say approximate with the median because I haven't ran that query against the DB in the last week.

Do you have any idea what percentage of those sites are blogs? The reason I ask is because the median size for blog posts are around 250 words and and the typical blog is configured to show 10 posts on the home page. That makes your stats about average for a database of blogs.

No idea what percentage are blogs but I know very few blogs rank for the terms I push through the system. I think wikipedia skews the numbers upwards, as they are a common site encountered on page 1 SERPs.

I have my own database of 6146 #1 ranked websites and the average word count is 718 words and the median is 649 words. I'm including all text, including navigation, sidebars, etc. I would really like to know how your database is so different? I'm guessing that your dataset includes a high percentage of blogs, article directories or something of that sort. While I have a very low percentage of blogs, it's mostly brick & mortar retail, e-commerce, manufacturing, large chains, service providers and such, no article directories or white papers.

Again, I don't log into anybody else's account but mine so I don't see specifics, only the system wide numbers returned by whatever query I input.

I can say I did some more specific querying last night against my keywords (tough niches) and throwing out the largest and smallest content volume on the front page, from across 100+ keywords, the average of the other 8 terms was 1800+. This would lead me to believe that more content is being rewarded over lesser content, however, I have a lot of research to do before I could assert that quantitatively.

For example, "weight loss". Fairly competitive term there.

The average Content Length of the front page results is 2,237 words. The #9 result has the lowest word count at 65 words, while the #6 result has the highest word count at 6220 words.

Throw out the outliers and you get 2,010 average words per the other 8 positions. Additionally, if you look at the sites with low content length, they are all either authority sites or subdomains on authority sites.

I'll link to a screenshot of it rather than embed it, as it's 3.47MB, for those who access this thread without high speed.

http://dl.dropbox.com/u/5534744/weig...screenshot.jpg

[dburk]

Originally Posted by Grindstone

I'm not. If you reference the graph I posted, it shows positions 1-10. The median for those is ~1500 and the highest, as clearly shown on the graph, is still sub 2500. I say approximate with the median because I haven't ran that query against the DB in the last week.

Hi Grindstone,

I'm sorry I wasn't more clear, by "median", I meant median word count of the data, not the median average on the chart you posted. By knowing the median word count it tells us where the middle of range is for the dataset.

Originally Posted by Grindstone

No idea what percentage are blogs but I know very few blogs rank for the terms I push through the system. I think wikipedia skews the numbers upwards, as they are a common site encountered on page 1 SERPs.



Again, I don't log into anybody else's account but mine so I don't see specifics, only the system wide numbers returned by whatever query I input.

I can say I did some more specific querying last night against my keywords (tough niches) and throwing out the largest and smallest content volume on the front page, from across 100+ keywords, the average of the other 8 terms was 1800+. This would lead me to believe that more content is being rewarded over lesser content, however, I have a lot of research to do before I could assert that quantitatively.

For example, "weight loss". Fairly competitive term there.

The average Content Length of the front page results is 2,237 words. The #9 result has the lowest word count at 65 words, while the #6 result has the highest word count at 6220 words.

Throw out the outliers and you get 2,010 average words per the other 8 positions. Additionally, if you look at the sites with low content length, they are all either authority sites or subdomains on authority sites.

I'll link to a screenshot of it rather than embed it, as it's 3.47MB, for those who access this thread without high speed.

http://dl.dropbox.com/u/5534744/weig...screenshot.jpg

Hi Grindstone,

Thanks for posting that. I am beginning to see how are datasets differ.

Let's take your example of the keyword "weight loss". That is what I refer to as a "research term" keyword. The reason I classify it as a research term is that a large portion of search queries for that particular term has low commercial intent. Yeah, I know a lot of marketers target that term, but it is a poor term for marketers since a significant portion of those searches are for information only (no commercial intent).

There is a category of terms that unsophisticated marketers seem to gravitate towards that are of low commercial value yet they target them anyway. The "weight loss" term is a classic example of one of those terms. There are many high commercial intent keywords that include the term "weight loss" as root word, but "weight loss" by itself is of low commercial value, it is research term.

"Research terms" tend to have higher word count content because it is information only the user is seeking. Research terms tend to have sites like Wikipedia in the top results because those sites are loaded down with high word count content. The low commercial intent tends to keep serious businesses from investing in ranking "research terms".

My data is biased towards keywords with high commercial intent, while yours seems to be populated with a significant number of "research terms" (low commercial intent). I think that explains why are data points are so different.

If you drill down into your data and look at keywords with high commercial intent I think you will be able confirm what I am saying. For example, take the the root term "weight loss" and add an additional word that ups the commercial intent. Look at those words and note how they tell a very different story.

"weight loss supplements" position #1 word count: 2427
"weight loss centers" position #1 word count: 203
"weight loss products" position #1 word count: 1601
"weight loss camps" position #1 word count: 355
"weight loss doctors" position #1 word count: 176
"weight loss retreats" position #1 word count: 779

These words were selected from the AdWords Keyword tool based on including the root term "weight loss", having over 10,000 monthly searches and high competition within the AdWords system (high commercial intent).

As you can see with this very small subset of data the average word count is 923.5 while the median is 355. Note that two thirds of the records are below the average and median is less than half of the average word count.

While I don't mean to imply that this small data sample is representative of your entire database, only that it is a real example of how you cannot use the average word count as a valid metric for the basis of your assertion. And that excluding keywords with low commercial intent, you have a very different picture.

The main point I am trying to make here is that the "average" is the wrong metric to look at. A closer look at the data shows that a larger number of web pages with word counts lower than the average. This seems to indicate a correlation of lower than the average word count having a better chance of ranking than above average word count. I bet you see the same thing in your own database. If you identify your median word count data point and it is significantly lower than your average word count, this correlates with a lower than average word count having higher rankings (the opposite of your original assertion).

I will close with one of the most famous and sage quotes posted on Twitter:
"Just because something is easy to measure doesn’t mean it’s important" -- Seth Godin

[Grindstone]

Originally Posted by dburk

Hi Grindstone,

I'm sorry I wasn't more clear, by "median", I meant median word count of the data, not the median average on the chart you posted. By knowing the median word count it tells us where the middle of range is for the dataset.

I was referring to the same median as you, across all data points.

I like your differentiation between commercial intent and research, give me a couple hours and I'll run those examples you posted and we'll see what sort of correlation, not causation, I'm sure Godin has something to say about that too , we can decipher. Good discussion, thanks!


Edit: Ran the first keyword real quick, was interested to see.

weight loss supplements

#1 site is a 11 year old exact match domain, obviously content length for this site is not a predominate ranking factor considering the age and EMD.

The average Content Length of the front page results is 2,591 words. The #8 result has the lowest word count at 660 words, while the #6 result has the highest word count at 6630 words.

Throw out the outliers and the avg content length on page #1 for this commercial intent term is 2327.

Of the two sites on the front page for that term with less than 1900 words (719 and 660 respectively), one has 164k links and the other has a very well developed backlink profile with lots of authority links going to deep pages.

Again, lots of content isn't an end all must be for ranking post Panda, but when you consider user engagement metrics like bounce rate and time on site, it certainly seems to help, especially for those SEOs working with less than authority sites.

[dburk]

Originally Posted by Grindstone

I was referring to the same median as you, across all data points.

Hi Grindstone,

Sorry, I may have failed again at being clear.

I meant the the median word count from the actual data, not the median from the "averages". We need to leave the "averages" out to get the true median data point.

Originally Posted by Grindstone

I like your differentiation between commercial intent and research, give me a couple hours and I'll run those examples you posted and we'll see what sort of correlation, not causation, I'm sure Godin has something to say about that too , we can decipher. Good discussion, thanks!

Great, look forward to crunching some data.

[Grindstone]

Originally Posted by dburk

Hi Grindstone,

Sorry, I may have failed again at being clear.

I meant the the median word count from the actual data, not the median from the "averages". We need to leave the "averages" out to get the true median data point.

Right, it's about 1500 words for front page listings over 20k plus queries. No average. The median of the averages would be 2275 (2223 + 2337/2). It slows everything down when I query the entire DB so I can't give you an exact # until I backup the DB and reinstall on another server so I can hammer it with queries to my hearts content. We have users all over the planet, so there really isn't a safe "off" time where I can slow it up for an extended period of time.

Also, I noticed your DB has different word counts than our scrapers, are you just pulling the body content div by chance?

[dburk]

Hi Grindstone,

I am pulling word count from rendered text, excluding form input fields, the same you would get by Selecting all, copying and pasting from a page rendered in a web browser.

[gmack]

Great thread! Thanks for the info. I think it's JUST saved me alot of wasted time and money.

As someone who has HAD some article writing and spinning software for awhile but never even got started with it (been in my 'to do' file for over a year), I'm NOW ready to give article writing and marketing a try for my own sites. I thought writing the articles myself would be the BEST way to go. Longer ones of 800 to 1000 words.

Given the newer state of article writing and google, does anyone have a BEST recommendation for how many articles to write for 1 site and best place to submit them to get results.

For example, my sig has the baseball products site I'm promoting. I thought I'd write two to four 800 to 1000 word articles and produce 2 to 4 videos. Once I have them completed, where would be the best places to submit them to get results? Manual submission or good software recommendations or people to do this? How far apart should I space out the submission process? Is 2 to 4 even enough to bother with?

Thanks for the great info. I would have started with the easy 'spinning' and would have ended up 'spinning my wheels' and probably getting banned.

Greg

[J3thro M]

Great, useful infos.... Still, spinning articles are Great.