homepage Welcome to WebmasterWorld Guest from 54.166.110.222
register, free tools, login, search, pro membership, help, library, announcements, recent posts, open posts,
Become a Pro Member

Home / Forums Index / Marketing and Biz Dev / Keyword Discussion
Forum Library, Charter, Moderator: open

Keyword Discussion Forum

    
About most useful of useless tools
Keyword suggestion tools don't work the way they should. Period.
Julia




msg:273074
 10:27 pm on May 19, 2003 (gmt 0)

Hi,

I have a sad story to tell you about my experience with such tools. I’ve read a lot of discussions on the subject and as I could understand just a few people are satisfied with the quality of their suggestions. To me it looks as though these tools were almost useless. Here is why:

Once upon a time I decided to optimise the home page of my main site for a tough two-word key phrase that was moderately general. I used WordTracker and WebCEO to see whether it was popular or not and make sure the results were valid. Both tools suggested it was a perfect keyword requested 6,000-7,000 times…daily!

I won’t bore you by telling how long it took me to tweak the page and set up high-quality text links. So, the results were devastating — I expected at least 150-200 visitors on a bad day because I was #5 in Google/Yahoo and #2 in Altavista. I got that very amount divided by 100. All my efforts brought me (take a deep breath) 2-5 visitors daily. Silly me! I was too naïve to bgeleive they couldn’t lie so blatantly.

I dismissed WordTracker and kept WebCEO for advisory opinion. I wonder what would happen if I used Overture — they counted plurals as singular keywords (keyword=keywords), which distorted facts even more. The volume of searches was, if I’m not mistaken, 12-14,000. It's all a lot of hooey. What do you think?

[edited by: WebGuerrilla at 9:43 pm (utc) on May 20, 2003]
[edit reason] delinked [/edit]

 

lorax




msg:273075
 4:42 pm on May 20, 2003 (gmt 0)

Hello Julia,
Welcome to WebmasterWorld!

>> It's all hooey.

Well...I think everything Internet/WWW related is a bunch of hooey - or at least hocus-pocus. ;)

Seriously though, gauging traffic for a particular keyword or keyword phrase based soley on the info provided by any tool seems awfully risky to me. I think tools make up some part of the equation but it's a smallish part. I'm more inclined to go with my instinct and hard research data from actual users within the target audience I hope to attract. There simply isn't a tool on this planet (this includes Google) that I trust 100%.

I think Wordtracker and WebCEO have thier place - and it's more a matter of which one (or both) suits your needs - but I wouldn't redesign a web site or a web page based soley on thier data.

Julia




msg:273076
 8:39 am on May 21, 2003 (gmt 0)

Lorax, thanks for your reply.

Being extremely challenging the Path of Research is not that good. All I managed to find out were a few keywords they would supposedly use to find my site. Unfortunately those suggestions had nothing to do with the expected volume of searches/visits. What good will it do if a Bobbie tells he’d use a certain keyword to find you? What we need to know is the approximate number of other Bobbies searching for this term daily/weekly.

I agree that both tools have their palce. Both Worttracker and Web CEO were good in discovering what people search for. When I enter my base keyword, Web CEO returns 1,000 suggestions of 1,137 found. The problem is I can’t figure out which of them work and which don’t. Numbers lead me a dance :(.

So to speak, I have to optimise a page for a possibly good phrase and see how many people click on my listing to understand whether it works or not.

As a side note: I noticed the results of Web CEO are more credible to some extent. Its suggestions look as though they were not completely filtered out. They include some weird combinations whicj I wouldn’t believe people search for but which still get me a little traffic according to my stats. Wordtracker’s results are more appealing but they don’t include these pesky keywords. I guess they have some kind of filters to get rid of obscure keywords so that they don't clutter results.

ncsuk




msg:273077
 8:44 am on May 21, 2003 (gmt 0)

You should also think in your head about what you think people will search for. Dont try to think of specific words or phrases either. If you have optimized a site for a certain set of keywords but they are not delivering then using a tool like webceo or wordtracker is useless.

You should try and build on what you have and use other related words that people may find your site by.

I get found by all sorts of things. I do SEO for example so you would think people would search for SEO or search engine optimization etc etc but I have had 2 clients come to me because they typed in "affiliate submission for dmoz" which I can say appears absolutly nowhere on my site.

You have to think about the stuff around what you are targeting because 75% (my opinion) of the time people are not going to come in on what you want them too. Its just too hard to judge. You have to look at the bigger picture and work around on that.

Julia




msg:273078
 9:46 am on May 21, 2003 (gmt 0)

ncsuk, I completely agree with you. A lot of customers convert after finding my site on what I consider to be irrelevant keywords. Your "affiliate submission for dmoz" s a good example of that.

lorax




msg:273079
 1:26 pm on May 21, 2003 (gmt 0)

Being extremely challenging the Path of Research is not that good. All I managed to find out were a few keywords they would supposedly use to find my site. Unfortunately those suggestions had nothing to do with the expected volume of searches/visits.

Understood. This is where knowing your audience is all important. Hard research data would be nice though it may be hard to come by for your market segment. There are things you can do to gain that knowledge. On-line surveys, newsletter surveys, cold-calls, sifting through your log files. It's very specific information so your best bet is to develop the analysis tools and compile the data yourself.

What good will it do if a Bobbie tells he’d use a certain keyword to find you? What we need to know is the approximate number of other Bobbies searching for this term daily/weekly.

Agreed. This is where WordTracker et al are supposed to play an important role. The reality is that you need to blend the information provided here with the other information you have at hand (or acquire) into a formula that works for your market. This is one of the most important obstacles that will make or break your online business. In order to succeed you sometimes have to break trail where none have tread before.

gmoney




msg:273080
 1:37 pm on May 21, 2003 (gmt 0)

Does anybody know the source of WebCEO search frequency data? I could not find any information about this on their website. A Webmaster World site search for "webceo" did not help me out either.

I find it helpful that Wordtracker mentions that they collect roughly half of their data from Dogpile and MetaCrawler and that their database contains roughly 350 million keywords collected over the past 60 days. I would think that WebCEO would mention something similar but I have not yet found anything. Any help would be appreciated.

Julia




msg:273081
 7:24 am on May 22, 2003 (gmt 0)

Knowing your audience is all important. Hard research data would be nice though it may be hard to come by for your market segment. There are things you can do to gain that knowledge. On-line surveys, newsletter surveys, cold-calls, sifting through your log files.

Everything but cold-calls! They make me feel like an idiot. As well I won't stand sifting through log files (I use live tracking).

I agree these steps are required to understand the audience but they will hardly help me correctly estimate the number of searches. Do they call it 'sampling'? If so, the sampling method won't bring definite results because my customer base is not representative.

Does anybody know the source of WebCEO search frequency data? I could not find any information about this on their website. A Webmaster World site search for "webceo" did not help me out either.

When I was considering purchasing WebCEO I asked them the same question. I was too careless to delete it so their answer was roughly like this. They admitted they were using some third-rate search engines and their database was less than Wordtracker's. The support guy mentioned an advanced math formula to make the results more precise.

In my opinion WebCEo's tool lose to Wordtracker in terms of additional features but it is pretty good for general keyword suggestion and very fast. What else? As I said it seems to be more credible because it shows some rarely searched keywords that get me a few visitors and are not displayed in Wordtracker's results.

Janet




msg:273082
 12:07 pm on May 22, 2003 (gmt 0)

One tactic I have started to use to refine and target keywords is to throw the list into a Google adwords campaign for a week with a low $/day cap on spending.

By reviewing the number of impressions over this time it gives a very good indication of the most popular search terms.

I also find it helpful to geotarget to the Australian market if that is what my clients focus is. That way we find what the aussies are searching for compared to the rest of the world.

I have found this very helpful. It has also helped to convince some reluctant clients of the value of certain search terms that they had previously discounted as unimportant.

lorax




msg:273083
 1:26 pm on May 22, 2003 (gmt 0)

>> but they will hardly help me correctly estimate the number of searches.

True. You'll need to use a blend or resources and your knowledge of the subject to determine this and it is a guess.

No sampling method is 100% accurate - obviously you know this. But sampling is all we have to get a feel for the demographics of the intended audience. The only way to improve your knowledge of that audience is to conceive/find/build/buy the tools you need.

Julia




msg:273084
 8:21 am on May 23, 2003 (gmt 0)

One tactic I have started to use to refine and target keywords is to throw the list into a Google adwords campaign for a week with a low $/day cap on spending.
By reviewing the number of impressions over this time it gives a very good indication of the most popular search terms.

Very good point, Janet. It looks like a solution, I'll try it for sure.

No sampling method is 100% accurate - obviously you know this. But sampling is all we have to get a feel for the demographics of the intended audience. The only way to improve your knowledge of that audience is to conceive/find/build/buy the tools you need.

There are three kinds of lies: lies, damn lies, and statistics. I am not keen on statistics but from my bitter experience I know at least one thing for sure -- if you are not Jupiter MediaMetrix you won't get much. I agree you can get a feel of your industry but it doesn't influence the bottom line of this topic. Even if I know my visitors to the tips of my fingers, I won't know how often they type in a certain keyword and existing tools don't satisfy this need.

lorax




msg:273085
 1:32 pm on May 23, 2003 (gmt 0)

>> There are three kinds of lies: lies, damn lies, and statistics.

LOL - how true.

tolachi




msg:273086
 5:50 pm on May 23, 2003 (gmt 0)

One tactic I have started to use to refine and target keywords is to throw the list into a Google adwords campaign for a week with a low $/day cap on spending.

I did this at one point when I was deciding what keywords to optimize my homepage for. However, I think a better method is to look at clickthrough rates and impressions. I recently combined my overture and adwords data and it definitely changed my keyword makeup. Granted it can sometimes be difficult to tell whether the click through rates are comparable, but in some cases it is obvious. The most interesting thing was that I found that a few general categories I hadn't been going for, "widgets", had a higher click through rate than a seemingly specific one, "golden widgets". Of course it all makes sense in hindsight...

Julia




msg:273087
 7:48 am on May 24, 2003 (gmt 0)

However, I think a better method is to look at clickthrough rates and impressions. I recently combined my overture and adwords data and it definitely changed my keyword makeup.

Got it. Thanks

biggles




msg:273088
 11:50 am on May 30, 2003 (gmt 0)

Does anybody know the source of WebCEO search frequency data? I could not find any information about this on their website. A Webmaster World site search for "webceo" did not help me out either.

gmoney, I asked WebCEO the same question. They said they have a different approach to WordTracker and use data
accumulated over the last 6 months to calculate approximate number of daily searches for specific keywords. Search queries are sourced from:

MetaCrawler.com
Kanoodle.com
MetaEureka.com
Galaxy.com
Yandex.ru
Fireball.de
Abacho.de
Freenet.de

A broader range of sources than WordTracker, but who can say if it's anymore representative?

Apparently new words are added to their database every day. Filtering is used to prevent the database being skewed with uncommon words that have been searched on in a particular search engine several times over a short period. However if such terms are normally searched for on in different search engines, they do get included, but some sort of special algorithm is applied.

Maybe this excludes some of the oddball results you can see in WordTracker? I only briefly trialed WebCEO so can't say either way.

Apparently world daily search figures are generated by applying a changing multiplier to get to the world daily searches number. So I guess that's similar to WordTracker's approach.

A big question I have about these tools is just how representative their sample data is of most search engine users.

I suspect most users of Meta search engines are comparatively unsophisticated web users. If that's the case, then simply extrapolating this sample set to predit total daily searches across all searche engines is flawed to some extent. Also WordTracker results are very American centric due to the Meta engines used. Possibly WebCEO's wider range of Meta search engines provides a more global or at least European perspective?

What do others think?

gmoney




msg:273089
 1:59 pm on May 30, 2003 (gmt 0)

Thanks for the information biggles. That 6 month accumulation of keywords could come in handy at times as I often want to get keyword search frequency data for some cyclical keywords and I often miss the event with Wordtracker’s 2 month accumulation of keywords and Overture’s 1 month accumulation of keywords (i.e. Say today I decide to go after next years Academy Awards which I believe happens yearly in March).

When you talked to WebCEO did they mention roughly how many keywords were in their six month database?

I suspect most users of Meta search engines are comparatively unsophisticated web users. If that's the case, then simply extrapolating this sample set to predit total daily searches across all searche engines is flawed to some extent.

I agree, but for politically correct purposes, I tend to call an “unsophisticated web user” a slightly more flattering “less technical user”.:)

I also agree that extrapolating from Meta search engines to other search engines does have flaws. Perhaps the results for widget related keywords are skewed higher or lower. However, I tend to believe that the relative search frequency between the widget related keywords is more immune to extrapolation flaws. For example, the estimated searches for “blue widget” and “red widget” from Wordtracker may be too low in both cases but the ratio of estimated searches between “blue widget” and “red widget” may be quite accurate and may help you decide which of the two keywords offers the most promise. Therefore the flaw with extrapolating may cause you to over predict or under predict the amount of traffic you can generate from your niche but using Meta search engine statistics may still be an effective means to determine which of the keywords you should target within your niche.

Julia




msg:273090
 10:53 am on May 31, 2003 (gmt 0)

I also agree that extrapolating from Meta search engines to other search engines does have flaws. Perhaps the results for widget related keywords are skewed higher or lower. However, I tend to believe that the relative search frequency between the widget related keywords is more immune to extrapolation flaws. For example, the estimated searches for “blue widget” and “red widget” from Wordtracker may be too low in both cases but the ratio of estimated searches between “blue widget” and “red widget” may be quite accurate.

Theoretically it is true but it didn't work for me with Wordtracker. The tool reported a three-word phrase was way too popular and beat a two-word term. I took it seriously and ended being betweem #2-9 and getting a fraction of expected traffic.

To sweeten the pill I must admit when I experimented with other keywords using WebCEO the ratio was more or less trustworthy.

gmoney




msg:273091
 7:48 pm on May 31, 2003 (gmt 0)

Hi Julia,

Just to clarify my thoughts above, I believe the two main sources of errors with using Wordtracker (and WebCEO) data to estimate things on other search engines are small sample size (i.e. the database is only a small fraction of the web) and different demographics (i.e. maybe MetaCrawler searchers have different interests and search differently than Google searchers). Anyway, when I wrote my post above I was primarily thinking about errors arising from different demographics.

I agree that the thinking in my earlier post did not apply to your example because I believe the problem you mentioned in the post above has more to do with the small sample size magnifying strange behavior than it does with the potentially different demographics of Meta search engines vs. others. I believe you would see similar problems even if you used a small sample size of Google searches (i.e. say you could track 1% of all Google searches for a day) to estimate Google traffic for a month.

It would be interesting to see the details of how WebCEO tries to filter strange anomalies out.

biggles




msg:273092
 3:01 am on Jun 1, 2003 (gmt 0)

gmoney - WebCEO did not tell me how big their sample database is. Also I don't have any furtehr details about how the filtering works to keep out oddball uncommon results. Maybe you should ask Vlad Lysik (CEO of WebCEO) directly.

Julia - are you saying you have found WebCEO more accurate than WordTracker?

Global Options:
 top home search open messages active posts  
 

Home / Forums Index / Marketing and Biz Dev / Keyword Discussion
rss feed

All trademarks and copyrights held by respective owners. Member comments are owned by the poster.
Home ¦ Free Tools ¦ Terms of Service ¦ Privacy Policy ¦ Report Problem ¦ About ¦ Library ¦ Newsletter
WebmasterWorld is a Developer Shed Community owned by Jim Boykin.
© Webmaster World 1996-2014 all rights reserved