|Keyword Selection Methodology|
Step by Step instructions
| 5:21 pm on Sep 23, 2005 (gmt 0)|
You have a web site idea.
You need to find the right keywords.
What is the process you go through to locate the "best" key words?
| 10:21 pm on Sep 27, 2005 (gmt 0)|
Look at what industry you are in.
Look at the keywords your competitors are going after
Brainstorm for new keywords and/or phrases
And you're set to go :)
| 1:35 pm on Sep 28, 2005 (gmt 0)|
|Look at what industry you are in. |
|Look at the keywords your competitors are going after |
In essence, steal it?
|Brainstorm for new keywords and/or phrases |
What is your process of "brainstorming"?
| 1:42 pm on Sep 28, 2005 (gmt 0)|
"Steal it" would be putting it a little harshly unless you are actually cutting and pasting their META tags and content. I'm sure no one would recommend this.
My guess is that the intention was for you to check what keywords your competitors are targeting on page and on PPC and determine why they are targeting those particular words. The keywords they are targeting may or may not be 100% applicable to your site.
| 7:54 pm on Oct 3, 2005 (gmt 0)|
IMO, researching the keywords of my competitors by looking at their meta tags is not stealing. Copying their meta tags and inserting them in my page is stealing. Reviewing what the competition uses is research.
Overture has a free keyword suggestion tool which returns the results of actual searchers in a recent month for the term you type in along with extensive variations.
So, using the classic example, if I type in widgets, I get:
1582 widgets yahoo
1279 window widgets
1192 konfabulator widgets
854 desktop widgets
602 dashboard widgets
333 wandering widgets
249 widgets for apple
239 mac widgets
204 free widgets
196 widgets pc
193 widgets world
and many more ...
While that's not quite brainstorming, it's good research. Also there is another tool that you can find if you do a G search for 'keyword suggestion tool' which returns Overture's results along with those of another tool, which shall remain nameless ...
Brainstorming might take the form of talking to someone else in the company or your family who knows your business ... or even a clueless friend ... and asking them what they would type in a search bar if they were looking for your product.
But checking the metatags of competitors or using the keyword suggestion tools is faster and in the case if the latter, more reliable, IMO.
BTW, what *are* widgets?
| 4:07 am on Oct 4, 2005 (gmt 0)|
yes, indeed overture give some foods for research. BUT, the numbers, most of the time, are not accurate. From my experience, words shown to be little search vloume sometimes might be jump off as a money keyword that drive you in tons of traffics; vise versa, some high search volume keyword doesnt drive as much traffic that they first look like -- even if i am ranking on the top 3 position.
thus, overture data is always just my preliminary references. anyone experience the same as well? i would be great if anyone here's willing to share how they go further with these data ...and is there other method to 'see' on the search volume.
| 1:36 pm on Oct 4, 2005 (gmt 0)|
|and is there other method to 'see' on the search volume. |
Well, yes, one sees more when one actually manages a PPC campaign, and studies log files.
I have no idea whether the #s are reliable or not with Overture; I am not sure how we would know, and would be pleased as tahiti, to know how to judge the accuracy.
But the traffic from SERPs and PPC clicks does go a long way to show me what is effective for a given site.
I run quite a few PPC campaigns on Google and Overture for various sites, and I have found my initial research to bear fruit in most, if not all, cases. I don't recall being surprised.
Overture's #s may or may not absolutely accurate, but I have found the relative strength of the #s to be reliable. That is, when Overture (or WT) says "key west fishing" is searched more often than the next most relevant phrase, the click throughs seem to follow a similar pattern. As do the log files.
Of course, other elements, such as the text for the ad would influence whether an ad receives a similar ratio of results. And the title element also influences whether a searcher chooses to visit my site or the one below mine.
Yes, a little searched phrase often outperforms the projected #s, but, again, that may have to do with the presentation of the ad, and not the pure numbers of searches.
So, if I am searching for the most relevant phrase, I have found that the one searched most often is the one I usually find most effective when optimizing the home page. In actual practice, I often choose the second or third most popular phrase for the initial optimization, and once I am finding the site in the top SERPs I further optimize for a more narrow, more frequently searched term. It's been my experience that once I refine the page optimization for the narrower phrase, the page continues to perform well for the initial keywords.
| 7:31 pm on Oct 9, 2005 (gmt 0)|
Widgets are "a non-existent item considered for the purposes of hypothetical discussion." :-)
| 7:05 pm on Oct 10, 2005 (gmt 0)|
Thank you for the information.
I understand the use of the logs, and PPC feedbacks, but my search is for a brand new site, therefore there will be no logs, or PPC feedback.
Following on Overture's suggestion tool, do you pick the top five? Top ten? Do you slide down, and target from the fourth to the fifteenth? That is, do you have a methodology to selected within the Overture sugestions?
Same with G's suggestion tool. Are the "suggestions" in a particular order, or just random from the results? If not random, what is your method of selection? If it is random, how do you follow-up to find the better ones?
| 3:11 pm on Oct 18, 2005 (gmt 0)|
|What is the process you go through to locate the "best" key words? |
Best for what? AdSense revenue? Conversions for your own product? Raw traffic?
The fact is, you won't be able to figure out what the best keywords are except by experimentation. Therefore, a more logical approach is to focus on making a process that lets you measure your particular flavor of "best" for as many search terms as possible, with as little effort invested per term as possible.
Finding the initial pool of targets:
a) pick a handful of obvious targets, based on the topic.
b) for each of these, locate the top 100 results in Google.
c) for each set of 100, run some software to produce a sorted list of  how many times every possible word appears and  how many different pages from the top 100 it appeared in. Ignoring words that don't appear on at least X different pages is a handy feature.
d) if you can build or buy slightly smarter software, have it also do the same for each 2-, 3-, 4-, and 5-word term that appears on more than one page.
This can give you some good clues about Google's keyword likes and dislikes in your particular topic area.