alphasource - 7:32 am on Apr 13, 2007 (gmt 0)

buying long tail *is* effective in several cases:

1. to add to arieng's point, google has been known to change its matching algorithm on occasion. i remember two distinct cases last year (april and november, but i could be off a little). when a broad matched phrase is suddenly getting mapped to different queries, the whole dynamics of the game changes. if you're buying "widgets" broad, and really the query driving most of the volume to that keyword is "cheap widgets", and suddenly "cheap widgets" is not mapped to "widgets", but instead "buy widgets" is, well... you get the point. so it always pays to be thorough and cover your bases.

2. i agree 100% that new advertisers pay through their nose to get distribution on google (and msn and now yahoo). the beauty of long tail terms is that if you get enough good CTRs, this enables you to buy head terms slightly more cheaply. i have seen the CPCs on some head terms go from $18 to $14 for the same position while i do nothing but tinker with ad copy and optimize for high CTR on tail terms. optimizing tail terms CTR is cheaper than head terms, and i play that game even if the conversion on tail terms is sucky because i get discounts on the head terms.

finally, i totally agree with OP that at the margin these initiatives are useful. issue is: the game is so competitive and when you get to the majors, the efficiencies are gained at the margin..