Personally i stick with exact matching, as the theory of 'broad matching' could take you into an ROI territory that is scarily expensive with little return - certainly if your chasing top adword listings!
Exact - specialist niche markets
Broad - Global sort of anything markets (i.e. Ebay / Amazon, etc)
[edited by: caine at 3:17 am (utc) on Nov. 28, 2003]
I've seen broad matches like this run very high ctr rates.
There are some national examples I can think of as well. I think broadmatch (well, the older one more than the new one) can be a very useful matching tool, and don't think that automatically exact matches should be considered more relevant than broad ones - that's something ctr determines.
Take one of my terms: niche widget
The top result in the Premium slot is Walmart. I'm sure it is a Premium ad for widget, but it is so totally out of place that I laugh every time I see it. While they were the nations largest seller of widgets last year (no accounting for taste), they would never sell niche widgets (non sexual) as they are against their "company image". Remember this is the company that won't even sell Maxim because it is offensive to their fundimentalist shoppers.
That just gave me an idea. Maybe I can go to a couple of the local churches and tell them what I saw them advertising. That should get a letter writing campaign going and maybe they will install some negatives and get out of my niche. <evil grin>
Another problem with single word broard matches is that it is often hard to keep the CTR up enough to keep them from being disabled. Lot's of negatives can help. but if your going to that trouble it is often easier to just do multi-word ads.
Most likely, if you bid on 'widgets' and get a 0.5% ctr rate, and a bid price of $1.00 - you'll spend a lot more than if you have a bunch of phrase matches that have 5-15% ctr rates, even with the same bid price. And because your ctr is so low, you'll never be able to compete for top position, except when people search for just 'widgets' which isn't very targeted.
I like broadmatching 2-3 words for a lot of industries. That way I catch the searchers who don't input the exact term I'm bidding on, but do input the 2-3 words I'm bidding on. It's a rare company/product that I'd recomend bidding for a single word term on Google.
On Overture, if you broadmatch, you also have the option to have the same term phrase and exact matched. So what most people do is end up checking all 3 boxes for the terms they want to broadmatch.
If you bid on 'widgets' in OV, and there are hundreds of 'word widget' terms, odds are, you're widget ad will never show for those other terms, even if you broadmatch it.
The other thing with Overture to look at is CTR rate. I've seen many people who say, well, my term is number one, what else can I do? Optimize Overture ads, like you would with Google. Overture makes a big deal of bid price, but very little about CTR rate, and I think many people forget to optimize Overture ads correctly.
Does anyone have a negative keyword list "template" they can share, with a bunch of common words you can use as negative keywords for almost any ad that is selling something? You know, words like sex, ****, picture, photo, download, free, etc. etc.?
One great tool for figuring out appropriate negative keywords: the 'natural' SERPS.
For example, search on one of your keywords, then take a hard look at the first hundred or so search results on the left. These results are a goldmine of information about how your search term is used in contexts that have nothing to do with what you have to offer.
Whenever you spot a non-related example, you are likely to have found a good negative keyword or two.
Just my $0.02 ;)
AWA
