1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

[TUT] Reverse Keyword Research - Find What Ranks Post Panda/Penguin

Discussion in 'Black Hat SEO' started by jkwilson78, Aug 23, 2012.

  1. jkwilson78

    jkwilson78 Regular Member Premium Member

    Joined:
    Jun 24, 2010
    Messages:
    224
    Likes Received:
    311
    Here is a method of doing keyword research that helps find new niches to enter and keywords to go after in the post Panda/Penguin world we are in now.

    A Warning: It's an involved processed so for those that see the potential and are willing to put in the work I think you'll be very satisfied. Plus, because it is not a one click fix I don't think you will have any problems with this method becoming saturated or over used.


    Ok, so this method involves finding keywords and niche ideas by exploiting the power of sites I call Rank Magnets. This is a fancy term for sites that rank for a lot of different keywords across hundreds of thousands or even millions of keywords.

    Sites like hubpages, squidoo, ezinearticles, etc.

    While they rank for tons of keywords across many many niches they aren't an authority on any one topic which makes them generally speaking easy to out rank. So if one of these sites ranks in the top 10 for a keyword that can be a sign that top rankings for that keyword are possible without tons and tons of work.


    Step 1: Start With Some Fancy Google Searches -

    To start off we want to use some advanced Google queries to find products, niches, and keywords that are in demand but not too competitive or too hard (hopefully) to rank for.

    Here are some advanced search queries you can use:




    For Squidoo.com:
    Code:
    "20..1000 Comments" "Ranked #500..1000 in Healthy Living" site:squidoo.com
    This will find all squidoo lenses that rank between 500 and 1000 in the "healthy living" category and have between 20 and 1000 comments. I like to use lower rank lenses instead of the very top ones because often the top lenses will have a lot more seo work done on them (more backlinks) so by aiming lower you can often find less seo optimized lenses.

    [​IMG]

    You can modify the query to change the rank and comment ranges and category.

    Some other things you can try is restricting results to a certain number of "Likes" or "pins"

    In the screenshot above, the result with an avatar shows "240 Like" so you could limit results by likes using this: "50..100 Like" This would limit results to those with 50 to 100 likes.

    Basically, anything that is visible on the page that is used to show the quality or popularity of the page can be used as a search filter!





    For Amazon.com:

    Code:
    "100..500 customer reviews" "Rank: #5000..10000 in Cell Phones & Accessories" "out of" site:amazon.com
    This one is a bit more complex. Here is what it does. It looks for all products with between 100-500 customer reviews, ranking between 5000 and 10000 in the "Cell Phones and Accessories" category. The "out of" part will make sure the average star rating (ie. 3.9 out of 5) is shown for each listing so you can easily see average rating of each product which will help you tell if it is a quality product people like:


    [​IMG]


    The number of reviews is a sign that the product has a lot of customer interest, the "out of" part let's us quickly see whether those reviews are good or not and the ranking let's us define how popular of a product we search for with the theory being the higher the rank then generally the more competitive the keywords are going to be to rank for.

    As with the squidoo search query, you can modify all of the values to make the ranges smaller or larger, change categories or add new parameters. One I would add is to limit results by the average review rating which can be done by adding "3.9..5 out of" This will only return products that have an average review of 3.9 stars or higher.

    The catch is that if you add this to the search query above Google will say the query is too long to process so you will need to remove something else to make the query work.





    For Hubpages.com:

    Code:
    "Sports and Recreation" "by * 300..1000 followers" "level 6..8 commenter" "50..1000 comments" site:hubpages.com
    Building on previous queries we get a little more complex with this one. There is a lot going on here.

    This will search hubpages.com for Hubs in the "Sports and Recreation" category, where the author has 300 to 1000 followers, where the number of comments is between 50 and 1000 AND there is a Level 6-8 commenter in the comments.

    [​IMG]

    The number of followers helps find Authors who have a following and which would indicate they write about quality products, services and topics of interest.

    The number of comments shows that whatever was written about people have an interest or opinion about.

    Now, I'm not naive. These metrics can and are manipulated by some so there is that to consider but it is still a decent gauge of quality.

    But since manipulation does exist, I've thrown in the search for a specific "Level" of commenter. The higher the level of commenter the more trust and influence that person has. So if they are commenting on a hub, chances are the hub is legit as a person with a high comment level is less likely to comment on pure spam hubs.

    Like the others, all of these parameters can be tweaked and changed out with others you might discover.





    For Ezinearticles.com:

    Code:
    "Internet and Businesses Online" "Submitted On * *, 2012. Viewed 100..1000 times" site:ezinearticles.com
    
    This is an old one that has been around in some for for a long time. I first saw it it (or a version of it) in a post by Jack Duncan on the WF. Which was my inspiration for finding similar kinds of queries for different sites. So props go to Jack for this.

    This query looks for articles in the "Internet and Businesses Online" category submitted this year that has been viewed 100 to 1000 times:

    [​IMG]


    You can get a lot of modification ideas just by looking at the result snippets. For instance you could search for articles with a certain word count "Word count: 500..600" or how many articles an author has written in which more articles written might indicate someone who is successfully ranking articles "500..2000 Articles"

    Or you could look for authors that have been around a long time: "Joined December *, 2009" (This would search for authors that joined in the month of December 20009)

    Finally, you can find articles posted on a specific date: "Submitted On August 22, 2012."




    Step 2: Find Keywords They Rank For:

    This step is pretty straight forward.

    Once you perform the searches and find some specific pages you think have potential, you can take the url of the page and dump it into SEMRush.com to see what keywords if any a site ranks for.

    This can be a bit of a time consuming process as you have to check one url at a time but if you stick with it you can find some worthwhile rank targets.

    Here is an example using the first result in our Squidoo example :

    [​IMG]

    As you can see this lens ranks in the top 10 for multiple phrases with over 1000 searches a month. Another thing to point out is that SEMRush uses the LOCAL exact match number from Google's Keyword tool and not the Global numbers as this screenshot shows.

    [​IMG]


    So it is entirely possible that SemRush under reports potential traffic numbers.

    Oh, another benefit is that you don't need to be a paid member of SEMRush. They will show the top 10 keywords for free and since they are generally sorted by highest search volume first and rarely does a single page rank for more than 10 high volume keywords you are going to see the "meat" of that page's rankings for free!






    Step 3: Checkout Backlinks:

    This could actually be Step 2 as they are pretty interchangeable.

    Once you find a page that is ranking well, you want to checkout the backlinks of the page to make sure you haven't stumbled across a well optimized page.

    Looking at our example we can see that it doesn't appear to have too much going on as far as backlinks are concerned:

    [​IMG]


    A total of 47 links from 9 different domains is not too tough to crack. Also, as you can see from this screenshot the top links to this page are all from other Squidoo lenses. What you can't see is that further below once you get passed the squidoo links the remaining links are very weak and lack authority. (Hop on over to OSE to checkout for yourself to see what I mean)

    So overall, it appear this would be a page that with a litle bit of work you could outrank...perhaps with another squidoo lens.


    Further analysis of the top 10 would obviously be required but I think you get the gist of this research method by now.




    Bonus Step: Ways To Further Automate This Process:

    If you're like me, you want to take a process like this and automate it is as much as possible. So here are a few tips to bring more automation to the table. I'm not going to mock it up with screenshots as I'm about exhausted putting this together :)

    1) You could load up a few hundred or thousand of these advanced search queries and scrape the top 1000 results with scraepbox giving you a huge list of urls to research
    2) Using a free program called Netpeak Checker you can load up all the URLs you collected into scrapebox and check them all for PR, backlinks, etc to get an idea which ones are using SEO and which are not.
    3) Once you've eliminated the obviously seo'd pages you can either manually check the ones of interest in SemRush.com like above or if you are a paying member of semrush you can have a programmer put together a script that will use the SEMRush API to automatically check which keywords the urls rank for and output everything to a spreadsheet. You can easily get something like this done for less than $50...I use vworker.com


    Well, that's it! It's not the prettiest of methods but it can really help you uncover some new ideas and identify keywords that you can go after.

    What I really like about SEMRush, or any similar service is that it shows you what is ranking AFTER the Panda and Penguin updates. Sure, the results are typically a month behind but since we are getting well past the initial Panda and Penguin roll outs we're at a point that we can see what is still ranking and use this to our advantage.

    Hope this sparks some new ideas!
     
    • Thanks Thanks x 63
  2. 50cent

    50cent Regular Member

    Joined:
    Feb 21, 2008
    Messages:
    253
    Likes Received:
    57
    wow, very creative with the queries and a top post, thanks given!
     
  3. weruncompanies

    weruncompanies Power Member

    Joined:
    Feb 15, 2011
    Messages:
    605
    Likes Received:
    286
    Location:
    California, USA
    Nice!
     
  4. Ninja250

    Ninja250 Regular Member

    Joined:
    Dec 29, 2009
    Messages:
    224
    Likes Received:
    72
    I don't understand this .. Why are you doing competition analysis on pages which rank 500-1000? I understand the part about finding good keywords, but why find out how many backlinks they have? Surely you want to know the competition for the top 10, not the 500-100?
     
  5. subster

    subster Elite Member

    Joined:
    Apr 5, 2008
    Messages:
    1,864
    Likes Received:
    1,448
    Location:
    Krauthausen
    woah, flashy and very advanced method! absolutely ubermethod. guys, please see the potential before rejecting it caused by the long read. as an retailer it can give you its real value for the decision if carrying this lr that product in your shop.
     
    • Thanks Thanks x 1
  6. araqueli

    araqueli Power Member

    Joined:
    Mar 13, 2010
    Messages:
    758
    Likes Received:
    271
    really good

    thanks given
     
  7. jkwilson78

    jkwilson78 Regular Member Premium Member

    Joined:
    Jun 24, 2010
    Messages:
    224
    Likes Received:
    311
    We're looking at pages within those sites (squidoo, hubpages, amazon etc) that have rankings/metrics specific to those sites. It has nothing to with Google search rankings. The Google search queries in my post are for finding keyword and niche ideas by mining each site.

    Once you find some topics/ideas of interest you move to steps 2 & 3 which is where you start to find out the backlink profiles and keyword rankings for those pages in the top 10 of Google. Step 1 is the research phase so to speak.

    This method becomes a lot more useful if you try to implement some of the automation methods I mention. Then you can really blaze through the data a lot faster.

    Also, if you use these queries in scrapebox be sure you have a lot of proxies and I would even set the scraping to just a single thread. Maybe it is just my proxies but Google blocks them really fast until I reduce things to a single thread after that I can scrape a lot more just not as fast as with multiple threads. They are pretty unusual queries so I'm sure Google picks up on that really fast and lays down the ban hammer.
     
  8. balsagoth007

    balsagoth007 Elite Member

    Joined:
    Oct 30, 2010
    Messages:
    2,102
    Likes Received:
    1,378
    Occupation:
    Doctor
    Great post mate. Rep added!
     
  9. shailzrocks

    shailzrocks Jr. Executive VIP Jr. VIP Premium Member

    Joined:
    Jan 1, 2012
    Messages:
    1,113
    Likes Received:
    1,766
    Occupation:
    Dolphin Seller in a Black Market
    Location:
    No Man's Land
    AWESOME Stuff!! Brother.. Liked the Search technique
     
  10. blackjack

    blackjack Regular Member

    Joined:
    Nov 23, 2007
    Messages:
    276
    Likes Received:
    51
    WOW!! Excellent Post and loved your search strings. How do you find these advance search strings?
     
  11. SmartMan

    SmartMan BANNED BANNED

    Joined:
    Jul 25, 2012
    Messages:
    673
    Likes Received:
    1,244
    Download this PDF

    Code:
    www.googleguide.com/print/adv_op_ref.pdf
    Awesome post OP! +rep
     
    • Thanks Thanks x 3
  12. General Lee

    General Lee Regular Member

    Joined:
    Jan 26, 2012
    Messages:
    230
    Likes Received:
    1,013
    Location:
    Confederate States
    Well done, I believe I'll be looking into this a little deeper.. thx
     
  13. jkwilson78

    jkwilson78 Regular Member Premium Member

    Joined:
    Jun 24, 2010
    Messages:
    224
    Likes Received:
    311
    Glad you got something from it!

    You can build these queries using the advanced search form as a starting point:
    https://www.google.com/advanced_sea...bgL&rls=org.mozilla:en-US:official&prmd=imvns

    But really, the two most helpful are the ".." which you can place between two numbers to search a range of values (it's called "numbers ranging from" on the advanced search form)

    The other is the "*" You can use that in place of words in a phrase to get a lot different options.

    For instance, if you searched:

    "red * are cool"

    If you used something like that it would return all results that contain any word between "red" and "are cool" (ie, shoes, hats, clothes, cars, etc)

    As for the specific queries for the individual sites, just look for any words or phrases that appear on every page of a site that are used to rank, categorize, or show value of that individual page. Once you start paying attention to this you can spot them on a lot of sites.
     
    • Thanks Thanks x 1
  14. kimomalcolmx

    kimomalcolmx Regular Member

    Joined:
    Apr 11, 2011
    Messages:
    271
    Likes Received:
    77
    Great share and +Rep given, will be using this for some tests
     
  15. THECEO

    THECEO BANNED BANNED

    Joined:
    Jul 22, 2008
    Messages:
    442
    Likes Received:
    212
    I would double rep that if I could. Great job.
     
  16. mikewaz

    mikewaz Newbie

    Joined:
    Apr 2, 2008
    Messages:
    31
    Likes Received:
    15
    Thanks for all the effort. Would've taken me all day to put something this detailed out. I'm gonna read this a few more times just to make sure I didn't miss anything. Subscribed and +Rep
     
  17. themidiman

    themidiman Power Member

    Joined:
    Feb 25, 2011
    Messages:
    701
    Likes Received:
    1,536
    Location:
    root@pts/0
    Love the thought process behind this. I'm a sucker for reverse engineering. Twisting this as we speak. Thanks.
     
  18. TrevorB

    TrevorB Jr. VIP Jr. VIP Premium Member

    Joined:
    Dec 21, 2011
    Messages:
    1,185
    Likes Received:
    361
    Location:
    Canada
  19. jkwilson78

    jkwilson78 Regular Member Premium Member

    Joined:
    Jun 24, 2010
    Messages:
    224
    Likes Received:
    311
    Thanks for the great feedback! Glad everyone is getting something out of it and I really like to see the comments from people putting their own twist on it.

    Post any questions you have here and I'll do my best to answer them.

    Also, if there is interest I have a process that goes along with this that would best be described as the "shotgun" approach to all this. If there is enough interest I will put that together as well.
     
    • Thanks Thanks x 1
  20. scribeforthetribe

    scribeforthetribe Registered Member

    Joined:
    Jul 6, 2012
    Messages:
    86
    Likes Received:
    31
    This is tremendous. Thanks and Rep given. I would love to see your shotgun approach to this as well, if you would share.