Dramatic Differences in Keyword Search Numbers in Different Tools

Discussion in 'Keywords' started by Thomas Holmes, Feb 6, 2006.

  1. #1
    I built a site based on statistics from wordtracker that said that there were over 400 searches a day on a particular keyword. Wordtracker gave this keyword combination a KEI of 68 Google, 190 Yahoo and 735 MSN.

    Now i've built the site and first of all it ranked number one in MSN for that keyword combination and number two for a second keyword combination (also said to receive around 400 searches per day on wordtracker). So i noticed there wasn't really any traffic coming from it so i thought ok well maybe they've just over-estimated the percentage of traffic that MSN gets for this keyword.

    But today i just noticed my site is now listed six in google under this keyword and I have received no traffic from it at all. So i'm a little bit more suspicious now.

    Looking through a few of the keyword tools:

    (1) Wordtracker is still very insistent - in fact it now says there are 550 searches per day on this.

    (2) Overture is reporting 48 per day but this is all variations on the order of this three word phrase.

    (3) Keyword Discovery is reporting 12 (!) per day.

    Has anyone else had any experiences like this - such a discrepancy between the different tools? or have any insight into this?

    Its really not looking good. Lucky it was just a site to test some of my seo learnings... :D
     
    Thomas Holmes, Feb 6, 2006 IP
  2. Trellian

    Trellian Well-Known Member

    Messages:
    35
    Likes Received:
    0
    Best Answers:
    0
    Trophy Points:
    121
    #2
    Hi Thomas,

    Unfortunately some terms can be skewed, but as a general guide WT and KD figures tend to be similar or at least close enough.

    Now 550 per day in comparison to 18 per day indicates that the WT data would be skewed, especially if you have top rankings for that term and do not receive anywhere near the 550 clicks.

    In relation to KD data, also a 18 daily searches number is a small sample which should not be used as an exact figure either.

    Cheers
    David
     
    Trellian, Feb 6, 2006 IP
  3. Thomas Holmes

    Thomas Holmes Peon

    Messages:
    54
    Likes Received:
    2
    Best Answers:
    0
    Trophy Points:
    0
    #3
    Hi David,

    thanks for that - yeah, it does really look very skewed. At least now I know in future to research on more than just wordtracker....

    But at the same time i'm rated number nine in dogpile and number seventeen in metacrawler. My understanding is that wordtracker figures are based on these two meta-engines. So with those ratings I should at least be getting some click-thrus - i'm on the first page for both!

    Alternatively the data isn't just skewed - its wrong.

    Yours confused...
     
    Thomas Holmes, Feb 6, 2006 IP
  4. Trellian

    Trellian Well-Known Member

    Messages:
    35
    Likes Received:
    0
    Best Answers:
    0
    Trophy Points:
    121
    #4
    yes WT get their data from dogpile and metacrawler, as for clicks yes if that term is actually searched for.....

    Cheers
    David
     
    Trellian, Feb 7, 2006 IP
  5. Thomas Holmes

    Thomas Holmes Peon

    Messages:
    54
    Likes Received:
    2
    Best Answers:
    0
    Trophy Points:
    0
    #5
    Ah, was wondering if you were related to the Trellian of Keyword Discovery. I guess you are.

    Its a nice looking tool. I've only used the Trial Version - as an independent person working in this area, the subscription is a little steep.

    I think I might have sent in a question but never got a response. I could however just have thought I sent it in - all a bit fuzzy.

    Anyway I wanted to know is it possible to seperate out data from the regional google databases so you can get data for certain search terms on google.cl for example?

    Its a nice tool though, Keyword Discovery with a lot of data. Its clear from my website statistics that traffic is not only coming from the american or .com sites but also the regional ones. This is something that WordTracker probably underestimates given its based on what I -guess- are predominantly US based search results.
     
    Thomas Holmes, Feb 14, 2006 IP
  6. FireStorM

    FireStorM Well-Known Member

    Messages:
    2,579
    Likes Received:
    88
    Best Answers:
    0
    Trophy Points:
    175
    #6
    try buying adwordskeywords software
     
    FireStorM, Feb 14, 2006 IP
  7. Trellian

    Trellian Well-Known Member

    Messages:
    35
    Likes Received:
    0
    Best Answers:
    0
    Trophy Points:
    121
    #7
    Unfortunately not at this stage.

    The next phase is to create a few regional versions, starting with Dutch and French versions, that only include keywords from those specific regional search engines, but again this will be pooled from a number of engines, and not a engine by engine case.

    Dutch is due out in 3 - 4 weeks.

    Cheers
    David
     
    Trellian, Feb 14, 2006 IP
  8. Thomas Holmes

    Thomas Holmes Peon

    Messages:
    54
    Likes Received:
    2
    Best Answers:
    0
    Trophy Points:
    0
    #8
    Great! Am in France at the moment - definitely an expanding market in terms of internet.
     
    Thomas Holmes, Feb 15, 2006 IP
  9. Brad Callen

    Brad Callen Peon

    Messages:
    854
    Likes Received:
    50
    Best Answers:
    0
    Trophy Points:
    0
    #9
    Yeah I have found that sometimes wordtracker is not entirely correct and overture seems to be a pretty accurate tool.

    Do checks with multiple different tools to determine where you think you should go.

    If more then one show similar stats, then you may have a better chance at making a good keyword selection.

    Brad
     
    Brad Callen, Feb 17, 2006 IP
  10. Trellian

    Trellian Well-Known Member

    Messages:
    35
    Likes Received:
    0
    Best Answers:
    0
    Trophy Points:
    121
    #10
    Did you not mean instead:
    "overture seems to be a pretty inaccurate tool"

    Cheers
    David
     
    Trellian, Feb 17, 2006 IP
  11. Thomas Holmes

    Thomas Holmes Peon

    Messages:
    54
    Likes Received:
    2
    Best Answers:
    0
    Trophy Points:
    0
    #11
    hey brad, yeah that seems to be the learning from that experience. back up your research with a few sources. looking forward to hearing more about keyword elite!

    David - in my case, overture seems to have been giving more reliable data than wordtracker.

    thomas
     
    Thomas Holmes, Feb 17, 2006 IP
  12. Trellian

    Trellian Well-Known Member

    Messages:
    35
    Likes Received:
    0
    Best Answers:
    0
    Trophy Points:
    121
    #12
    I use all the main 4 tools and find this not to be true...

    Wow I'm defending the competition. But I can not imagine how Overture can provide any reliable data on terms when they do some "conversions" to the most popular terms that they display. Such as, misspellings, hyphenated terms, punctuation variations, singular and plural combinations are "merged", "combined" and "mapped" to the most common form of that term.

    Some examples:
    - merge all plurals as one: "fly" is the same as "flies", or "hotel" is the same as "hotels"
    - merge all same words together regardless of order:
    "sydney hotels"
    is the same as
    "hotels sydney"
    and the same as
    "hotel sydney"

    As a result the Overture data is over inflated and inaccurate as you do not know which phrase and spelling variation is the more popular search phrase.

    Why does Yahoo/Overture do this? All of these features are great for Yahoo Search Marketing PPC advertisers, in short Overture simplified their keyword tool for use by their PPC advertisers, in order for them to see the potential search volumes they can expect. Yahoo make more $ by delivering PPC listings for keywords that an advertiser might not have bid on.

    Cheers
    David
     
    Trellian, Feb 18, 2006 IP