|Overture's Search Term Suggestion Tool|
How reliable is it?
I recently optimized a page for a certain keyword string and felt pretty stupid after I got the #1 ranking in Google and had a whopping ZERO visitors coming into my site with that search string.
I got the keyword data from Overture's tool and it ranked the string at something like 20,000+ searches in June. There was another similar keyword string that had about 300 - and this is where I get most of my visitors from for this page (sad, I know).
I may have already answered my question about the reliability of Overture's data, but the keyword with 300 searches seems somewhat accurate. So does Overture just have some data thats iffy?
The data is routinely corrupted from various sources, however, it *is* the best data of it's type around. (eg, you can't get what they give you anywhere else...)
It's probably worth remembering that Overture's data does not contain Google searches. Having said that, with 20k searches in OT, you would expect some activity in Google!
Researching in OT then optimising for Google is a pretty widely used trick, though. It can return some good results.
|I got the keyword data from Overture's tool and it ranked the string at something like 20,000+ searches in June |
A majority of those could be the results of software queries.
|the keyword with 300 searches seems somewhat accurate |
There have been lots of good discussions on the board about this. Try this search, and similar searches, on Google...
site:www.webmasterworld.com overture search term suggestion tool how reliable
I was looking at the Overture data and it seems they they take the real amount and times it. There are terms that no one whould search for (10 and 20 word searches) and list 1,000 last month.
This post may help. I got it straight from Overture in August.
Keep in mind that I got this information by phone from their tech people.
|I was looking at the Overture data and it seems they they take the real amount and times it. There are terms that no one whould search for (10 and 20 word searches) and list 1,000 last month. |
These are automated searches... Some folks get nervous and like to check their positions frequently. ;)
I'm not exactly sure of the history, but I think Overture tried to fix this (when it was GoTo... on the GoTo site itself, before it went wide) by giving its sanction only to certain bidding software that wouldn't skew the stats (though I never saw that this got fixed).
Once Overture went beyond its own engine, though, they lost more control. No one was likely to check position on Overture unless they were bidding, but on other engines, with competitive serps, this isn't true.
I assume that position checking software used on any Overture partner messes up the Overture stats. There's always that mom and pop site in Omaha that wants to check all search engines every few minutes to see how it's ranking for "mom and pops tourist widgets in south omaha."
The use of automation is obvious for the long phrases... not so obvious for the one and two word terms. It used to be, when the bid prices were posted on Overture, that you could get a sense of how close some of the bids were and how much skewing there would be. I look for patterns that suggest automation... and try to compare with Google stats when I have them.
For existng sites, a test run on AdWords, either with exact targeting, or with partial matching and careful log analysis, can give you a better sense of what ads are clicked on... not necessarily what's searched... but there's some sort of a parallel.
|...they take the real amount and times it... |
This is called "extrapolation," and it's what WordTracker does. Since it has a small database, and extrapolation can magnify errors quite a bit, I tend to distrust it. There are obvious automated searches in WordTracker data whose numbers have been multiplied.
The demographics of all the tools are different as well.
Ultimately, you need to use all the tools you can get... Overture, Google, WordTracker, and gut instinct. Once your site is up, use your logs as well.
|This is called "extrapolation," and it's what WordTracker does |
I think they just copy the Overture results. For my kw I get exactly the same results in both including weird extrapolation results.
Wordtracker=Overture as far as word count is concerned.