The Different Keyword Search Results Between Overture and
Wordtracker
Keyword Selection is fundamental to Search Engine marketing. Get
the wrong keywords, your online business is doomed to fail. Find
the right keywords, and you will drive masses of targeted
traffic to your website via the Search Engines.
There have been discussions in webmasters' forums why are the
search query results of Overture so much different than those of
Wordtracker? For example at the time of this writing a search
query for the key phrase 'internet marketing' yielded 342, 848
searches for the last 60 days by Overture, and 2, 356 by
Wordtracker. Now, which is more accurate?
Overture
It is a Pay Per Click Search Engine. According to Overture, its
statistics of searches in previous months are compiled from its
partners, which include AltaVista, Yahoo, MSN Search, HotBot and
All the Web. So the statistics from Overture are broad based
because it has a larger network.
Its data, though, has some drawbacks.
1. There's no distinction between...
a. Singular and plural terms.
You have to figure if surfers are searching for the singular or
plural form of the keyword.
b. Upper and lower case.
c. Human queries and automated queries.
Queries by automated bid optimizers, position and ranking
monitors, link popularity analyzers are registered as hits.
2. Duplicate Searches
For example a person doing a search for a particular key phrase
in Yahoo, and then in MSN would be registered as 2 hits.
Wordtracker
It is a keyword generator and analyzer, and it does not have
direct access to the major Search Engines' databases.
Wordtracker gets a lot of its analyzer data from Meta-crawler
and Dogpile, which are meta search engines. Metacrawler and
Dogpile search the major Search Engines like Google, Yahoo, MSN
Search and Ask Jeeves, and retrieve the best results.
The data of Wordtracker, mostly gathered from Metacrawler and
Dogpile, represents only a small percentage of total searches on
the Internet.
Wordtracker has about 350 million searches over an 8-week
rolling cycle (Source: Search Engine Workshops Weblog, June 30,
2005). Now for some Maths. 350 million searches over 56 days
would give an average of 6.25 million searches per day. Google
with a share of 36 per cent of Internet traffic (Source:
comScore qSearch, July 2005), registers about 112 million of
searches per day (Source: Top Ten list, Wordtracker). So in
comparison to total searches on the Internet Wordtracker
represents only 2 per cent.
When Wordtracker returns zero query for a particular search
phrase, it does not mean that no one searches for it on the
Internet.
However, with Wordtracker, automated queries are not added to
searches, and duplicate searches are eliminated. Singular and
plural, upper case and lower case search terms are distinguished
except for keywords where singular, plural, lower or upper case,
are similiar in meanings eg.'keyword' and 'keywords'.
Overture or Wordtracker?
For Overture the figures are inflated, whilst for Wordtracker
they are under reported. Nonetheless, these are helpful tools
for keyword research and selection. Use these figures as guides,
and not absolutes to make comparisons in your choice of
keywords.
For example if key phrase 1 has 10,000 searches in Overture, and
key phrase 2 has 2000 searches, then it is obvious that key
phrase 1 would bring in more traffic if your web page is
properly optimized. But don't expect 10,000 visitors with key
phrase 1.
I have read on Webmasters' forums of people who had selected
keywords based on the promising number of searches on Overture,
optimized their web page, and achieved within the top 10
positions of major Search Engines, yet saw very little traffic.
In my opinion, if you are comparing popularity between keywords
or key phrases, Overture will do the job. If you are selecting
keywords or key phrases to start an online business, or
advertising on Pay Per Click engines, conventional wisdom would
tell you Wordtracker is a better bet. It is better to set lower
expectations based on lower figures, and be pleasantly surprised
when things turned out otherwise.