Welcome to WebmasterWorld Guest from 188.8.131.52
Forum Moderators: open
I compared the output from wordtracker and overture for the following terms for a clergy related clothing website I am working on.
Note that the count for overture is a lot higher than the wordtracker count. They are roughly proportionate.
I thought wordtracker pulled data in from a bigger universe - am I mistaken?
clerical shirts - 40
clergy shirts - 63
clergy vestments - 74
clergy - 160
clergy clothing - 54
clerical shirts - 197
clergy shirts - 416
clergy vestments - 231
clergy - 1279
clergy clothing - 218
When comparing the count with the number of web pages for each keyphrase (i.e. to identify how competitve the phrase is) what is the most accurate search engine to compare with wordtracker (I used google) and with overture?
And while I'm here - is there a more accurate way of finding the best keyphrase than to divide the count for a phrase by the number of pages returned in a search?
the first is the difference in the data that is collected. WT maintains a database of approx. 350 million searches collected from multiple sources over an 8 week period. That means it's not collecting every single search, but rather a sample of all the searches.
The Overture database collects all the searches conducted at Overture and its partners sites for the previous month. I've never seen any info published by Overture that tells the actual size of the database, so there isn't really away to compare them.
The other thing that will cause the Overture numbers to look quite a bit different is the fact that they combine plurals and mispellings into the count, whereas WT doesn't do this.
WT will list the different variations separately. WT also doesnot contain any search data generated by automated search tools or position checkers. In many cases, that type of software generated searches can dramtically scew the numbers.
When comparing the count with the number of web pages for each keyphrase (i.e. to identify how competitve the phrase is) what is the most accurate search engine to compare with wordtracker?
One isn't more accurate than another because the level of competition is different on each one. If you were only going to check one engine, then it would make sense to collect your competitive data from Google simply because it has the biggest market share, but those numbers won't give you any indication on which term is the least competitive on Alta Vista.
is there a more accurate way of finding the best keyphrase than to divide the count for a phrase by the number of pages returned in a search?
The other key factors you want to take into consideration are the overall range of competing pages, and the overall market share of the particuar engine in relation to the other engines you are checking.
With competing pages, you need to remember the fact that off-page factors will have a much greater impact on rankings for terms with high page counts.
The way I approach this is to assign a numerical value to specific ranges of pages that works as a competitive adjsutment.
0-10,000 = 1
10,001-20,000 = .9
20,001-30,000 = .8
30,001-40,000 = .7
40,001-50,000 = .6
50,001-60,000 = .5
50,001-70,000 = .4
70,001-80,000 = .3
80,001-90,000 = .2
90,001-100,000 = .1
100,000 + = 0
Using those numbers you can simply multiply the count (the popularity of the term) by the corresponding competitive adjustment score.
If you are going to compare terms across multiple engines, then you would need to apply a similar adjustment to each engine based on its market share. Going that will make the final score a better estimation of the terms total traffic potential.
Mike Mindel from Wordtracker here. Just to clarify this last point.
Our database consists of all the keywords from the major metacrawlers Dogpile and Metacrawler for last 60 days. This comes to about 350 million searches, but it is ALL searches on these Metacrawlers. Which is a sample of the overall web searches.
However, we also allow users to use the Overture keyword tool through our system if they want. I.e. combined with the power of the lateral tool and the baskets, and competition analysis etc. Users can plug it into the system instead of our own database if they feel like it.
Hope that clears things up.
It's interesting to see how people combine search phrases using wordtracker and the PPC's. The phrases being bought on PPC are not the same as the niche phrases on Wordtracker, it seems that you get the early adopters in there and people just jump on the bandwagon, when there are some excellent phrases going cheaply that will deliver farmore traffic than some of the high exposure phrases.
I researched a term I thought might be profitable.
On Overture is said 43,000 searches last month.
On Wordtracker is said 83 searches last month.
Some difference hey?
Well, I built and optimised a page for Inktomi. Submitted and got to #1 at MSN.
In 72 hours so far I have had 2 hits on this term. Now, which do you believe, Wordtracker or Overture?
On a separate issue, does anyone know a way to query the UK overture by using a URL with keyword in it? You can do this on the US database by using:
You can query the UK database by going to :
But how do you add a search term to this uRL?
that normally does it.
What I am discovering more and more is that both wordtracker and overture's results are totally wrong, either that or being number one for your keyword on a SERP doesn't mean you're going to get a the lions share of the clicks, but even then 2 visits in 72 hours for a keyword that supposedly has tens of thousands of people searching for it seems a bit rediculous.
Incidently the same search term that got 43000 searches at US Overture got Zero at UK one. Not a US based term either. However, I would prefer to get a ZERO returned than a 43000 so I can not waste so much time in the future identifying bogus niches.
When I ran an overture.com campaign a few months back I noticed that on average, despite the amount of affiliaite sites overture work with, that it only accounted for 8% of all my traffic.
If the overture search suggestion tool shows a total of 43000 searches in any particular month for a keyword, then what I would do as a general rule is times this number by 12.5 to get an overall idea of what the total number of searches made for the term accross all search engines. It seems to work for me. Again, it will vary depending on the keyword.
Note these are three 2 word phrases with mispellings of the correct word as the second word in each case.
These counts are not accurate by a long way and multiplying them by 12.5 would be rediculous in this instance.
I do not mean to offend you or anyone here, but MSN would certainly have more than 2 searches in 72 hours if these figures were right, and my page is #1 for all of them.
thanks for the link. Never heard of this one but will check it out.
"On Overture is said 43,000 searches last month.
On Wordtracker is said 83 searches last month." - perfectG
Here is a link to another related thread titled wordtracker and overture disagree [webmasterworld.com].
Anyway I have some thoughts on your example but take them with a grain of salt since I don’t know really know what keyword we are talking about. I’m guessing that the 83 number is the Wordtracker’s “count” data which is the number of searches in the last 60 days on search engines that Wordtracker tracks. Currently, “count” happens to roughly correspond to the number of daily searches on all search engines combined. Also, I am guessing that the 43,000 Overture number is a victim of Overture combining search frequency data for variations on the keyword (i.e. keyword search frequency of plurals and other variations are mapped onto one keyword and the keyword they map to isn’t necessarily the most searched variation). The Overture data may also be a victim of automated rank checking programs.
"Well, I built and optimised a page for Inktomi. Submitted and got to #1 at MSN.
In 72 hours so far I have had 2 hits on this term. Now, which do you believe, Wordtracker or Overture?" – perfectG
Having abandoned hope for the 43,000 Overture number, I will see if a “count” of 83 in Wordtracker would have helped predict the results you see. If you view 83 as roughly the number of daily searches on all search engines combined and you take Wordtracker’s estimate that MSN is gets roughly 13% of all searches then you would expect there to be roughly 32 searches (83*3*0.13=32) on MSN for your keyword during the 72 hour (3 day) period. Since you received 2 of them it appears that your #1 listing, which is below sponsored matches and directory sites and above the other web pages, received 6% of the MSN searchers. There are so many approximations going on and a 72 hour time period is so short but at least it appears that the Wordtracker data can put you in the right ballpark of what to expect.
Sorry I don't know how to reply with quotes to a previous thread.
This is incorrect. There were no sponsored pages or directory listing for these keywords. This is why I chose these keywords for my testing purposes. Sorry for not sharing the keywords but I prefer to keep my URL private. Therefore my page would have been the first result in all searches for this keyword combination.
When I have done this same "testing" with Wordtracker results (again using MSN/Inktomi), I have gotten the approximate traffic I would have expected from the figures Wordtracker predicted.
You are correct however that these keywords are victim of the pluralis & misplellings lumped together fiasco.
Have you guys compared the results with the E-Spotting tool?
Do you have any idea what to multiply the espotting results by to get a better idea of the total number of searches made for that particular term?
perfectG - is your site in Google yet and if so how many refferals do you get from there?