keyword choices ... which would you choose?

[Crazy_Fool]

i've got to rebuild a site soon, and there are many keyword combinations i could use in the titles of the pages. this site is not likely to rank highly for any of the keywords as they are all very common. even the combinations are very common.

as an example (and this is not for the site i'm doing), imagine a site that offers computer training in office software. any keyword pairs such as "microsoft word" are going to be very common. google shows 2 million or more results for the common terms. bigger combinations such as "microsoft word computer training" are going to be less common (between 500,000 and 2 million results in google), but still far too common for the site to stand any chance of top rankings.

other combinations are slightly less specific, such as "microsoft word 2000 training", so there should be a greater chance of appearing towards the top. these combinations tend to show between 200,000 results and 1 million results in google.

however, a look at overture tells me that two word pairs such as "microsoft word" are searched for 20 or 30 times more than the less common pairs / combinations.

should i use the less common combinations in the hope of appearing somewhere near the top? or maybe i should use the common combinations to be in more searches, but ranked lower?

whatever happens, this site isn't vital and isn't going to rank highly for any search terms, so these changes are more for experment and practise than anything. any comments are welcome.

(BTW - i've already got several sites ranked highly in google for their various search terms, so i have a good idea what i'm doing, just for this site, there are too many keywords to choose from!)

[fathom]

go to google adword
make a ad (its not real anyway)
select a market (region and language)
type in your current ideas for keywords, (uing commons between them)

google will then allow you to look vertically and horizonally for different keyword combos and niche market (that you may not have thought of)

once finished making a list - go next and google will price what each keyword is worth. The higher the price the more competition and the harder it will be to "freely get to No#1. For google - target words under 5 clicks per day that meet and match your target market with products or services.

Downside - this is Google only and based on daily click rates not long trends (what's good today maybe nothing tomorrow when your site is ready)

[startup]

I like fathom's approach to selecting keywords.
Part of your keyword research should include how many sites are targetting any KD or phrase you are planning on using. Do a search and take note of how many site index pages come up for the search. Some searchs will produce over 100 sites before you will see any subpages in the serps. If you see subpages in the top twenty even though the return is 2+ million, that KW is very do-able.

[Crazy_Fool]

i normally do this for other sites, and the relevant terms are easy to spot and its easy to get a top 10 place. with this particular site, there are many very very common keyword combinations with large numbers of results. the less common combinations still have fairly large numbers of searches. in google, even rare keyword combinations have PR7 and 8 sites in 100th place - there is little chance of a small PR4 site beating any of them.

i guess what i'm asking is whether i would be wasting my time using rarely used keyword combinations in the titles etc, considering that the site still probably won't get listed in the top 100 for most of them.

[JamesR]

do searches for the more common ones on Google and see how well they are optimized. I bet you could beat most of the competition out there if you apply the techniques learned here. I have two sites that are in some of the most common phrases in the universe and rank pretty well. The key is how competitive the phrases are. I measure competition by how many other SEOs are sweating it out day by day to rank better than you. If you read here and apply the good stuff, you are way ahead of the ball game.

[WebGuerrilla]

Less common, higher positions wins every time. While it is quite possible that a common term ranked at #47 could draw more search activity than a less common phrase ranked at #9, the #9 phrase will still perform better because the conversion rate drops significantly as you get deeper in the SERPS.

Bottom Feeders Don't Buy:)