Overture vs. Wordtracker

Huge differences in what they say

[Small Website Guy]

Overture reports that searches for a certain TV show happened 14697 times during the month of July.

Wordtracker reports that this TV show was searched for 1155 times during the last 60 days and predicts that there will be 1111 searches for that term in the next 24 hours.

I happen to have a website about that TV show which received a little more than 400 visits last month, that's 14 a day. (Although my rankings suck even though I thought it's SEOed well and it has a PageRank of 5 and it's listed in DMOZ. Number 32 on Yahoo and 100 something on Google. No wonder why no one finds it.)

Anyway, I looked at the search results for a more commercial two keywords (let's call them "keyword1 kekyword2") that I recently created a new site for. According to Overture, there were 57246 searches for keyword1 keyword2 during the last month. Impressive.

But then I looked at Wordtracker which reports a dismal 164 searches during the last 60 days and predicts 158 total searches during the next 24 hours.

Now this is a pretty huge difference between those two examples. For the TV show, Overture's number was 12.7 times greater than the Wordtracker number (reflecting Overture's greater search traffic?).

But for the more commerical keyword pair, Overture's number is 349 times greater than the Wordtracker number.

What's going on here?

I tried another TV show. Overture's number is 21 times bigger than Wordtracker's number.

I tried another very popular commerical search term, Overture's number is 432 times bigger.

Wow, is Overture VASTLY overinflating the number of searches for commercial words (by a factor of TWENTY?), or do people use the search engines that Wordtracker uses for its results to only search for old TV shows?

[penfold25]

You are not using the free version of **** are you? The standard wordtracker trial version only produces results from altavista

[Small Website Guy]

Yes, I was using the free version of Wordtracker.

How would the non-free version's results be different?

Maybe they have a really poorly designed website, because I didn't read an explanation of how I'd be getting anything different from a paid subscription.

[bufferzone]

I don’t know if this holds true still but in the old days when overture was called GoTo it was commonly known that you needed to take the suggestions and search estimates with a grain of salt. GoTo was a pay per click SE and the searches got inflated by webmasters checking out their bids for keywords.

This, or something along the same lines, might hold true today. What you need to do is to correlate your own actual traffic with what the two tools estimates. Over time you will discover what the true ratio between the estimates and the actual traffic is in your line of business.

[bigjohnt]

I've found that regardless of the source and scope of data on the web, there is an enormous amount of "fudge" or variance. All accumulated data is skewed based on source, duration, and depth.

I recommend using any of those numbers as relative numbers- not absolute numbers.
The skews for each are variable based on innumerable factors, including webmasters checking positions, fraud clicks, single users searching more than one site on the same term... on and on.

Make your own rule of thumb, and test how it works for you. Its anybody's guess, and it will change month to month, or even day to day. I have yet to see any quoted number match source to source.

Example: Even checking popularity of search engines, using media metrix, nielsen, Alexa, or any other. You will never get the same percentage. You MAY however get the same relative positioning, or the same top 4 - 5.

[cabowabo]

Overture counts queries from bots the same as they do from actual browsers, which is why their numbers are inflated. I have done testing on a keyword that Overture said was getting 1200 queries per month. I bidded on the keyword and had a #1 ranking. Zero traffic. WordTracker reported 0 queries per day on average. I would suggest leaning toward WordTracker to the point of falling over.

Cheers,

CaboWabo

[ogletree]

These are just tools for kw research. I have terms that don't even show up on WT but I get 40 visitors a day from. All you can do from these is compare the terms you know with the numbers and give terms a high med or low rating as far as traffic. OV includes all kinds of things in there number. I have not used OV in some time. I use WT (paid) every day.

[wattsnew]

ogletree,

<terms that don't even show up on WT>

Yes, all my best generic positions and clickthough rates are for keywords that WT seems to vapourize as insignificant. My best phrase (30% of traffic) did not turn up in OV, GG or WT as suggestions.

But the traffic for them was there on Google and Yahoo.

Makes you wonder if there's a (gasp) conspiracy of non disclosure going on somewhere... :-)

W

[Illah]

WordTracker reports DAILY results. It's not 'XXX' searches in the past 60 days and then 'YYY' predicted for the next 24 hours. Why would it say there were 100 searches in 60 days and then say 100 in the next 24 hours? Doesn't seem very logical.

What the numbers mean are 'XXX' APPEARANCES in thier database over the past 60 days, not total searches online. The second is the predicted daily searches.

Now Overture reports on monthly results.

Sooooo, multiply the predict number on WT by 30 to get monthly and compare those. Now they should be much more similar (but there are always freak occurances).

--Illah

[HyperInterActive]

Regarding keyword analyzer software...has anyone ever tried webceo.com's software?

BTW: was up greg!

[Vork]

oooofff
finally found a decent thread on what has been bothering me for months...
thank you guys!
What I was trying to do all this time was determining the so called weighted average of keyword popularity index based on the WT & OV values.
I hear somewhere at this forum that OV tool displays not overture queries only but Yahoo, AV & alltheweb as well - areas were Overture preferred listings are also displayed (is this a correct assumption BTW?)
If so, knowing the SE traffic coverage of the Overture suggestion tool (acc. to nielsen say currently - 40% of all SE traffic) I multiply overture data by 0.4 and given that WT free daily stats for a given keyword represent AV only (nilsen stats again) multiply it by 0.02 (appr. 2%). Then I sum up the values and divide them by 0.42 (40% & 2% respectively) to obtain weighted average.

Does anyone here do anything like that?
Does anyone try to calculate effectivess index based on WT and OV stats?
Anyone willing to share their know-how formuli? :)