PDA

View Full Version : Trellian - Keyword Discovery


tyates
22-11-2005, 15:24/03:24PM
Greetings,

I was wondering if anybody has any experience with the keyworddiscovery product from Trellian? From some of the questions and answers on this product from Trellian's forum, it seems there might be some flaws that have or haven't been worked out.

In any case, the product seems to be what I'm looking for - something that can give a broad keyword use summary in categories (such as recreation, shopping, etc.) with the ability to drill down in further subcategories.

Any other alternatives or suggestions would be appreciated. Thanks

Connie
22-11-2005, 16:06/04:06PM
Hi tyates and welcome. :hi:

I don't know anything about the tool you are referring to so won't comment about it.

Blue
22-11-2005, 16:37/04:37PM
The thing to look for in any tool like that is where they are deriving their data from.

To the best of my knowledge, Wordtracker has the least skewed, and therefore most accurate results.

However, that said, manual research works best for me.

And welcome to the forums!!! :hi:

ihelpyou
22-11-2005, 18:56/06:56PM
Welcome to the forums tyates! :hi:

That's the same company that sells the bogus "submitwolf' that auto submits to a kazillion fake search engines., so I would not give this crap software any credence either.

Danny
22-11-2005, 20:26/08:26PM
I could forward my email to you and you'll receive their weekly spammy mail :D

Connie
22-11-2005, 20:40/08:40PM
:uplaugh:

projectphp
23-11-2005, 00:51/12:51AM
Whilst Submitwolf is lame, AFAIK, and I haven't used it, Keyword Discovery is on a par with Wordtracker, if not marginally better. Similar bias (i.e. the numbers are skewed) and similar volumes of DB terms.

What Keyword discovery does better is suppossedly seasonal terms.

Here is how they say it works http://www.keyworddiscovery.com/engines.html:
How we collect the data

The search engines listed below either provide their complete search logs and these are imported in full into our database, or samples are collected by scraping search statistics from ISP logs and other sources.

Data is collected on daily basis however due to the index size, the online database is updated monthly.
So, they use ISP logs, which is slightly different to the usual method (of being given the logs by SEs).

If you have any doubts, and you can take some of this with a grain of salt, http://www.keyworddiscovery.com/testimonials.html is a pretty good list of names that like the product.

Danny
23-11-2005, 04:45/04:45AM
Here is how they say it works I'm glad you noticed they said it :)
They are the very same people who say the following about submitwolf :

"It can register your web pages with over 1000 search engines and directories. With Submit Wolf, a task which would take months of your time, is reduced to only minutes. In addition, you can also submit your URL to over 1,000,000 link pages."
a pretty good list of names that like the product My suspicious mind tells me that some people say anything to make a few bucks.
If someone, eg. JW, says or writes something about a product and puts a link to it, then they should at least be open about it that it's an affiliate link.

I'm sorry PHP, but i do receive a weekly mail from some Edward Tax fellow who asks me to join their affiliate program and put a link to them on some of my pages.
How many such mails do i have to receive before it can be called "spam" ?
I definitely don't buy my cigs and my pills (if i ever need any) from someone who spams my mailbox, so why on earth would the same not apply for a keyword discovery tool ?

Quadrille
23-11-2005, 06:25/06:25AM
Originally posted by projectphp
If you have any doubts, and you can take some of this with a grain of salt, http://www.keyworddiscovery.com/testimonials.html is a pretty good list of names that like the product. You are a trusting soul; most spammers can come up with a list of fawning idiots who fell for the scam.

And those who cannot, simply make it up - anything on their site is hardly an independent assessment, is it?

Anyway, the list you found includes total strangers, spammers' friends - and one person who would endorse anything - without understanding the consequences.

That page stands as a warning, not an advertisement.

Irony
23-11-2005, 08:02/08:02AM
To those wondering WTF is going on here - there is more information in this post (http://www.ihelpyouservices.com/forums/showthread.php?s=&threadid=19788#post206308) and below.

ihelpyou
23-11-2005, 14:05/02:05PM
The projectphp post can be certainly taken with a grain of salt. Stay AWAY from Trellian stuff. Period.

I wish those trellian folks would make a post in here. I'd love to blast them a new one.

ihelpyou
23-11-2005, 14:58/02:58PM
Hey Mike; Is that list of names of testimonials suppose to be impressive to us?

FYI; It isn't impressive to me in the least. Don't know about anyone else, and frankly, don't care.

Anyone who would allow their name associated with such a product is a total idiot in my book.

tyates
23-11-2005, 17:02/05:02PM
Thanks for the reply. It sounds that there are good/bad opinions on Trellion in general. I'm somewhat on the fence between going with the keyword reports put out by word tracker and the word discovery. I'll have to think this one over.

Quadrille
23-11-2005, 17:33/05:33PM
Originally posted by tyates
It sounds that there are good/bad opinions on Trellion in general. ... But not many good ones in this forum. In fact, the only good one is from a visitor from another forum, here to defend his pal promoting the stuff!

To be fair, there are forums where this sort of stuff gets a good press - but I hope you'll always consider the nature of the comments, not just count the votes!

ihelpyou
23-11-2005, 19:30/07:30PM
hmm, let's see; tough decision;

I could keep on using www.wordtracker.com which has been doing ONLY this software forever and a day now.

OR: I could change over to a company who is known for sending out "BULK" spam emails, and for selling a useless piece of AUTO submission software to newbie webmasters for years and years, and that which does a site zero good, and goes against ALL legitimate search engine TOS's, and can actually hurt a site in the long run.

Oh, gee wiz; A tough decision huh?

Not for anyone who cares about their websites.

projectphp
24-11-2005, 00:08/12:08AM
But not many good ones in this forum. In fact, the only good one is from a visitor from another forum, here to defend his pal promoting the stuff!
Sigh. Sometimes, I really don't know why I bother.

Quadrille (or is it Andrew?), I am coming to Europe next year, so why don't we have a beer sometime and let all the hate go? I am sure you will find I am, like shylock, a real human with real feelings, don't have two heads or whatever it is that Trolls have, and I am even good company, on most occassions. I'll even get the first shout, and I promise the pub will have white curtains :cheers:

In any case:
<settingrecordstraight>
Trellian is located in Melbourne, Australia. I, too, am located in Australia. Australia is a big country. Melbourne is 1,000 KM from where I live in Australia, near Bondi Beach in Sydney. This is similar to the distance from Chicago to New York, or London to Milan.

I did, it must be told, meet a Trellian dude at a booth once circa 2002. I don't remember his name, what he looked like, or anything 'cepting I gots me a free disk called, from memory, "SEOTools". which, yes, had a copy of the lame SubmitWolf.

Lastly, whilst Trellian is not my enemy, calling us "pals" is a bit of a stretch, as they haven't even yet bought me a beer,a nd we certainly don't lean on each other when we are down :)

I do, however, have an interest in improving what I do. If Keyword Discovery is better than Wordtracker, I am all ears (and eyes I guess).
</settingrecordstraight>

Hey Mike; Is that list of names of testimonials suppose to be impressive to us?
No Doug, it is meant as a point of reference. That is what testimonials are: a good guide to help you decide for yourself.

I personally highly recommend testimonials to clients, and that is usually not in a field with the added benefit that readers will actually know the names (even if here certain people are so scary they only get to be initials).

So, if a testimonial from Sue in Kansas is good, one from {initials only person whose name IHY dare not speak} is surely infinitely better.

In any case, apart from {initials only person whose name IHY dare not speak}, who else don't you all like on the list? Christine Churchill? Rarely have I heard a bad word said about Ms Churchill. What about Andy Beal? He seems nice enough. And lastly Beth Abernathy, a member here and SEMPO scooper? Is she no good?

Personally, I think the list is impressive, if for nothing other than Beth Abernathy's comment:
"I have been using Keyword Discovery now for about 2 months. I previously used a combination of the following tools: Google Adword Suggestions, Overture Keyword Selector Tool and Wordtracker...I have found that the features offered are better than using the 3 combined tools and saves time! My clients are very happy with the reports I have been able to provide them."
To me, that is a fantastic endorsement, and one that made me think "Wow. Saves time, is better, makes clients happy. Works for me." That comment, more than any other, made me take notice.

If I was Trellian, I would be proud of each and every one of the comments there and, yes, those testimonials are impressive.

That all out of the way, what I really find odd is the judging here. We have lots of people that have never tried the product (including myself) discussing it.

So, has anyone actually used Keyword Discovery, and what did they think of it before I give it a go ;)

Irony
24-11-2005, 00:49/12:49AM
Hey Mike, I see you basically ignored Danny's comments. I think his points are valid.

If a company resorts to email spam to promote whatever they have to sell, it speaks volumes about their credibility. Do you ever buy from spam? I don't. And just like Danny, I can't see why a keyword research tool should be an exception. Even if it is really as brilliant as Beth Abernathy says.

projectphp
24-11-2005, 01:02/01:02AM
Irina, do I have to answer every question? I was under the impression that in a forum one adds the insight one has. I don't think it is fair to expect me to play question answerer to every concern IHY has, do you?

And besides, it isn't my product, I haven't received nor sent any bad promotional emails, spam or otherwise, and I have no idea about any historical background. So relly, how can I can comment? IMHO, that would be irresponsible of me!

I have no idea what else they sell, why or how, and quite frankly I don't care. I don't judge companies on any one criteria, as that is a bit simplistic. If Keyword Discovery works better than the alternatives, not using it because of spamming that happenned to others is not the way I make decisions.

Lets also remember, the original question was this:
I was wondering if anybody has any experience with the keyworddiscovery product from Trellian?
I can't actually comment, as I haven't used it, but I would like to know. If anyone has experience, I am all ears!

Irony
24-11-2005, 03:57/03:57AM
Originally posted by projectphp
Irina, do I have to answer every question?

No, you don't :) Thanks for answering mine though.

Danny
24-11-2005, 06:51/06:51AM
Hey Mike, I see you basically ignored Danny's comments. I think his points are valid. Irina, do I have to answer every question Well, i'm obviously not Irina but given that in a forum one adds the insight one has (if i may use your words), i find it a pitty that you ignored the essentials of my comments while at the same time you use the fact that i said "JW" instead of "Jill Whalen" in defence of those "great" testimonials.
You did not do it once or twice, but 3 times !

Anyway, i typed "JW" because i find it easier typing than "Jill Whalen" and because i rather focus on writing decent English which is hard enough for me already since it's only my 3th language.
So, if a testimonial from Sue in Kansas is good, one from {initials only person whose name IHY dare not speak} is surely infinitely better. Despite the spam i receive from Trellian, despite the crap they sold in the past, and despite the machine-generated pages they have on their keyword discovery site (which IMO could be not only be breaking Google Adsense TOS ), a lot of people started talking about this so called "new, great" tool.
So i went out trying to figure out why.
I saw that list of testimonials and noticed JW (oops, i did it again). She's the only one i know in that list and i admit she used to be (and still is for some easy trusting people) a well respected person in the SEO industry.
Exploring her site i was interested to read what she has to say about the tool. I expected to find a review but instead i was redirected to the keyword discovery site again. Then i realised i was tricked and had clicked an affiliate link. It made me pissed.
Soon i found that even in her very own posts in her forum, she used the same hidden affiliate link.

Even Michael Martinez calls this -probably without knowing that JW does the very same- "a sneaky way to hide affiliate links"
(posted at http://www.highrankings.com/forum/index.php?showtopic=16605 )

Which leaves me to be convinced even more that the entire tool is total crap
and no, i didn't try it nor am i going to try or pay a single dime for obvious reasons given in this thread.

Quadrille
24-11-2005, 06:53/06:53AM
Originally posted by projectphp
Keyword Discovery is on a par with Wordtracker, if not marginally better. You made the above statement; you were the only person to speak in favour of the product. Surely it's reasonable for people to want to know more?

Try to keep your response on topic; the personal stuff makes you look silly.

projectphp
24-11-2005, 07:28/07:28AM
Welcome to beat up the guy that's not a Mod night. I hope I enjoy my stay :)

Firstly, I would sincerely like to have that beer with you sometime Andy, and the invitation remains open indefinitely. I understand if that makes you feel uncomfortable, but I promise I won't bite. Besides, who passes up a free beer :)

I expected to find a review but instead i was redirected...Which leaves me to be convinced even more that the entire tool is total crap
So, Jill's affiliate link makes you believe the product is crap? Honestly Danny? That is a harsh criteria for choosing, and pretty easy to abuse, wouldn't you say?

I really don't get the issue with redirecting affiliate links through tracking scripts. Can someone explain that to me as a bad thing? I would do the same myself.

Exploring her site i was interested to read what she has to say about the tool. I expected to find a review
http://www.highrankings.com/issue150.htm - sometimes Google works (though to be fair, it took three searches).

I know Jill didn't write the above info, but she did accept it into her newsletter, and I assume that means she read it.

I don't know about you, but I don't write reviews of everything I like nor see. Often, I just say "read this review by {INSERT NAME HERE}." Not to much of a stretch to assume Jill did the same, and used David's words as she thought they were apt. No use re-inventing round things all the time ;)

Look, Andy is right, we are way off topic here. Keyword Discovery, good or bad and how to decide, is up to each individual, and I honestly am in the same boat as the original poster, and looking for the justification that tips me over into using it.

I am not sure I am getting any closer to an answer here, and I think my methodology for choosing a product like this may vary from those here, as does my happiness with wordtracker, which I personally don't find brilliant. So, if anyone does have some input on using the system, maybe Beth Abernathy will pop in, I would love to hear it.

LMAO ROFL LOL. One last thing: I just realised: http://www.wordtracker.com/testimonials.html also features some familiar names ;)

Quadrille
24-11-2005, 07:47/07:47AM
Originally posted by projectphp
Welcome to beat up the guy that's not a Mod night. I hope I enjoy my stay :)

Firstly, I would sincerely like to have that beer with you sometime Andy, and the invitation remains open indefinitely. I understand if that makes you feel uncomfortable, but I promise I won't bite. Besides, who passes up a free beer :)No-one's beating you up; seriously - no-one.

Free beer never ever makes me uncomfortable; offer noted with thanks :)

And I've nothing much to add on topic; I think people have enough to make up their own minds - as with all tools, "If in doubt, I'll leave it out"

Danny
24-11-2005, 08:55/08:55AM
I agree we are way off topic but can't resist replying to some :
So, Jill's affiliate link makes you believe the product is crap? Honestly Danny? That is a harsh criteria for choosing, and pretty easy to abuse, wouldn't you say? Pls note that i said "Which leaves me to be convinced even more".
If i look at the source of a testimonial for a product which i don't trust to start with, and i find a page that contains no further info apart from an affiliate link, then that testimonial is worthless to me. Realising that the affiliate link is hidden, is making it even worse.
But maybe that's just me.
I know Jill didn't write the above info, but she did accept it into her newsletter, and I assume that means she read it. Of course she read it... how else would she stuff her affiliate link in it ?

Actually, i also would like someone who actually uses the product to comment here. And if someone at Trellian knows why i'm receiving a weekly email , then they are free to let me know.

ihelpyou
24-11-2005, 12:59/12:59PM
If i look at the source of a testimonial for a product which i don't trust to start with, and i find a page that contains no further info apart from an affiliate link, then that testimonial is worthless to me. Realising that the affiliate link is hidden, is making it even worse.
But maybe that's just me.
Trust me; Not just you at all. I'm "very" sure many in here agree with you and feel the exact same way. Only people who are affiliates and friends of trellian would say otherwise.

I repeat:

Anyone who puts their name beside a company who created a submission tool to sell to "unsuspecting" webmasters, AND is still screwing them to this day with it, is an idiot. The email spam is further proof of the idiocy.

Why the heck would I try this tool when wordtracker has been very good for me? If it ain't broke, don't fix it. Certainly don't fix something with something much worse.

BTW: Most people, including SEO's out there, don't know how to use wordtracker properly. I know this as I see the posts around the internet with the silly questions....... FROM SEO's.

projectphp
24-11-2005, 20:23/08:23PM
But do Trellian sell submit wolf anymore? Seems they sell the product in one big suite of http://www.trellian.com/seotoolkit/overview.htm that integrates with their online tools.

AFAIK, their rank checker seems to use the Google API, and their other tools all seem to use ISP logs.

Perhaps you need to look at Trellian again Doug, because old prejudices are often wrong prejudices.

Connie
24-11-2005, 22:40/10:40PM
Perhaps you need to look at Trellian again Doug, because old prejudices are often wrong prejudices.
Your right that old prejudices can be wrong. On the other hand there is a principle that pretty much holds true.

A leopard does not change it's spots. The Spam Danny receives, and the fact that Jill Whalen (JW) hides her affiliate link says to me Trillion is up to the same old thing.
But do Trellian sell submit wolf anymore? Seems they sell the product in one big suite of http://www.trellian.com/seotoolkit/overview.htm that integrates with their online tools.
To be honest I haven't looked. I'm not going to. Just a gut feeling. Trillion has packaged submit wolf with their new software suite.

Mike I understand why you keep posting in defense of Trillion. Jill supports Trellian . You are going to support anything Jill supports.

Now you can make your post about how stupid I am. That's what you usually do when I get involved in a discussion with you.

If you ever come to my neck of the woods I'll be glad to drink a beer with you as long as you buy. :D :cheers:

projectphp
24-11-2005, 23:43/11:43PM
Connie, can you show me where Jill hides her affiliate links, and I mean show me a specific page? Can you then also show me an example of a good affiliate link, and point me to some best practice on the issue.

I ask because I wonder what she is really doing wrong. Personally, I would do the same, for several reasons.

Firstly: tracking. You need to know stats to see results.

Secondly: creating affiliate shortcuts makes the easier to remember and type and saves time. I don't know about anyone else, but I have about 4 Brazillian passwords in my head, and a hard to remember affiliate link is just a waste, and surely a shortcut is the best way to run one's own site.

Eaxmple:
<a href="/{affiliatename}"> vs
<a href="http://{www.affiliatessite.com}/{page}?{affiliateid}={myid}">

One is short, easy to remember and has one variable, {affiliatename}. The other is way long, and requires you to remember four variables (all the { } bracketed things). So, whilst you may find this unnacceptable, I think it is just good sense!

Mike I understand why you keep posting in defense of Trillion. Jill supports Trellian . You are going to support anything Jill supports.
Connie, mate, is life all one big consipracy? And how do you know what motivates me? For all you know, I might be motivated by a desire to annoy you ;)

So why not ask why I am here rather than claiming to know? In any case, I'll save you the trouble. I am here to find out the truth. Jill's position is irrelevant to what I do, and I posted because there was misinformation and a lack of help for a newbie asking a good question.

Next time, rather than reading my mind long distance, just ask :)

I really don't get your post though. You seem to want to prescribe a business decision not to some sort of logical and rational criteria, but to some bizarre unravelling of a products personal matrix.

That being the case, as Jill also supports Wordtracker ( http://www.highrankings.com/forum/index.php?showtopic=18167# is a good example), why is no one here arguing against that?

What Jill has to do with any of this is beyond me. She neither wrote Keyword Discovery nor posted here. Forget Jill (trust me, you'll feel better) and focus on what about Trellian and Keyword Discovery sounds no good.

Some examples: is the methodology sound? Is the interface intuitive? Is the sample size any good? What is the data bias (as all data has bias)? What are the costs and how does this compare to the alternatives?

Those are the real issues, IMHO, not Jill's handling of affiliate tracking nor anything else.

If you ever come to my neck of the woods I'll be glad to drink a beer with you as long as you buy.
LOL. I'll buy the first one, howsabout that? (as long as it isn't American beer ;))

Irony
25-11-2005, 06:04/06:04AM
Sorry for taking the thread slightly off topic again. Could someone kindly explain me, why all the fuss about keyword research at all?

We've got Wordtracker, Overture's free tool, Google AdWords free tool, Google suggest, etc... etc... etc... We've got a lot of tools available for free. I'm sure every SEO in the world has already chosen the one that suits him/her best.

For me, it is the free tool available at Digitalpoint's and combining Wordtracker's results with Overture's. Even so, I usually look at the WordTracker column only. It gives me a lot of good key phrases, usually, much more than I can use up (given the fact that there are only two of us on the staff of our SEO department, and both Dmitry and myself have to constantly wear 6 to 7 hats in order to provide decent SEO/SEM packages to our customers). I honestly don't need more keywords. I'm well aware that WordTracker can find much more combinations for me, but I seldom need them.

For this reason, I'm not even sure I'll renew my Wordtracker subscription for the next year. I probably will. It depends. But buying another one, for the sake of some additional features I will probably never use, is out of the question. IMO, money can be better spent.

Another consideration. Wordtracker's results are very inaccurate (fact acknowledged by everyone). From what I've heard about Keyword Discovery, it is just as inaccurate... so why bother paying for yet another inaccurate tool?

Still more. Often, the clients come with their own list of desired keywords (many of them are worthless according to Wordtracker, but not worthless at all in real life). I don't mind this sort of co-operation at all, because while I know some SEO stuff, the clients often know their niche much better than I do. And because they use their logs and their AdWords and Overture stats to choose those keywords. AFAIK, of all keyword research tools available, your own AdWords stats for a year or two give the most accurate results... verified with time and practical work.

I understand that for huge sites being SEO'd by huge companies with 30, 40, or 100 people on the staff, huge lists of keywords may be useful. For huge PPC campaigns, as well. The question is: how many huge SEO companies do we know? And how many are just small shops with 1 to 5 SEO's on the staff, like mine? Do we all really need another - PAID - keyword research tool? Isn't it logical that we never had too many of them? Isn't there a good reason behind the fact that we had all been satisfied with Wordtracker and some free substitutes, until Trellian introduced Keyword Discovery?

Just wondering....

ihelpyou
25-11-2005, 12:03/12:03PM
projectphp; First off; http://www.keyworddiscovery.com/engines.html

That's the very same way wordtracker gets their results, but only worded differently to make it seem different. Who actually searches on some of those 170 engines they list? No one does. Second; They wrote it like Google and Yahoo give them access to their server logs. That's BS and we all know it. They don't go into very much detail at all about how they get their results unless I missed it.

I agree with Irina; Who needs another keyword tool anyway? Anyone?

Further:
http://www.trellian.com/swolf/features.htm

Read it. Nice features, Huh? 1000's of search engines, 100's of directories, AND of course, over 500,000 links pages. It also emulates all browsers so as to "TRICK" those engines into thinking it's not a robot. Oh gee, that software seems like great stuff, right? What would you say if I told you that the software could harm any website who uses it? Oh yeah.

If I recall, www.websavvy.cc had a big problem with that submission software and finally found a way to block it. Is that right Deb?

All of this defending of this company is giving me a headache. Of course, the majority of the "elite" in this industry give me a headache anyway, so this is normal stuff. Most of the "elite" back this crappy software as well.

And we wonder why this industry has a bad reputation?

Pleeeazeeeee

projectphp
25-11-2005, 19:27/07:27PM
Could someone kindly explain me, why all the fuss about keyword research at all?
I am with you Irina. Keyword research isn't the hardest thing in the world. However, if some service did it faster, I would be up for that, especailly seasonal stuff.

Doug, sorry for the headache. I like ibuprofen myself, it really works!

That said, I think you extrapolate a bit, and if you didn't, you might feel better. I didn't defended Trellian, and especially not Submitwolf which, for the 4th time, is lame.

What I am trying to understand is whether one specific product is any good. I think judging that on the company's history is a bit harsh. Otherwise, why would anyone by Windoze ;)

projectphp; First off; http://www.keyworddiscovery.com/engines.html..That's the very same way wordtracker gets their results
Actually, no, it isn't.

There is no info I can find online now, but Wordtracker used to say (http://66.102.7.104/search?q=cache:a279ht8fUhYJ:www.wordtracker.com/support/faq/index.html%3Fcategory%3D4%26questionsecondset%3DQuestions+site:www.wordtracker.com+collect&hl=en):
Where do you collect the data from?
Data is collected from several of the largest metacrawlers (these are engines that query the main search engines simultaneously) on the internet. These include all the keywords (over 175 million per month) from metacrawler.com and dogpile.com. We combine the results together, carefully making sure to remove suspect entries (mainly people who continually request the same thing over and over e.g. 'xxx company is the best').

And this is KD methodology:
The search engines listed below either provide their complete search logs and these are imported in full into our database, or samples are collected by scraping search statistics from ISP logs and other sources.
See the difference?

If not, allow me to explain. Wordtracker get the logfiles for specific search engines. That means they can only tell you the search results for two engines.

KD does this, and also grabs ISP logs to look for any search from any known SE.

Who actually searches on some of those 170 engines they list? No one does. Second; They wrote it like Google and Yahoo give them access to their server logs. That's BS and we all know it. They don't go into very much detail at all about how they get their results unless I missed it.
OK, I think you misunderstood what they do. KD, grabs the ISP logs, and crawls through them looking for SE referrals, from any of the 170 engines listed. That is good, for whilst Xuppa.com may not be a lot of searches, why would you ignore them is ISP logs?

For those that need some background on how ISP logs work, when I connect to IHY, my connection goes through many, many computers before getting to the destination computer. Any of these can log the request.

What that means in practice is that a line is put in the logfile of the proxy that says what page the request was for, where it went and where it came from. That is what most ISPs do.

Example: you search Google for [low mortgage rates], an ISP logs it, KD gets the logfuile, they find the request string ( http://www.google.com/search?q=low+mortgage+rates) pull out the keywords and put it in their DB.

As the ISPs give Trellian access to their data, similar to how Hitwise works, KD does indeed include real searches comnducted on Yahoo, MSN and Google.

So, back to comparison, Wordtracker is given (pressumably) 100% of the logs for two infrequently searched meta search engines, while KD uses the data from a subset of ISPs, and looks for searches on any known SE.

It is all complicated, and I don't know the Maths, but bias in data from one source probably won't be as accurate as bias from a small sample of all data, unless the bias in that sample is skewed, say by all the logfiles coming only from Nevada.

KD data may perhaps offer a better representative sample of real searches conducted online than data from two specific SEs, and it is certainly a distinct difference between the two services.