Content network ad groups

[SEMblahblah]

I know that the ideal strategy would be to have tightly focused ad groups with with keywords that form ONE particular theme. So, for 'reviews', 'downloads', 'articles', etc etc is it better to have a unique ad group for each of these themes?
Whats the advantage of doing this? Is it just that I can see where my ads are performing and where they aren't and make bid changes accordingly?

What if I club them together? The ads would still show on all such sites..right? Atleast, this is what I've observed for some of the campaigns.
Ofcourse, I won't be able to manage bids for each theme..but for some reason, I've notice that this sort of an ad group receives far more impressions.

Maybe I'm missing out on something..because having unique ad groups for such themes sounds far more logical. So let me rephrase the question - how would you structure an ad group for themes like reviews, articles, downloads, etc (what type of keywords, how many keywords, variations, etc etc)?

[SEMblahblah]

Also, would you recommend using the keyword tool for creating content ad groups? Does search volume matter AT ALL?
Can I come up with keywords (whichever fit the theme) irrespective of the search volume?
How do you guys generally go about it?

[pavlovapete]

Hi SEMblahblah,

my .02c

Whats the advantage of doing this?

- think about a portfolio of shares in different companies. If one company is tanking you decide to dump it and that action alone increases the overall return of your portfolio. Same with keywords in an adgroup. You need to be able to track the performance of the keyword.

What if I club them together? ... this sort of an ad group receives far more impressions

- google likes tight themes - then they can ascertain if your ad and landing page/ offer are relevant to the user's query. If you have broad matched a ton of keywords in one adgroup then yes, you'll probably get a load of impressions. Depending on your objectives a large number of impressions may be a good thing.

I'm gradually coming over to the idea of "suck it and see" - i.e. throw some money down and use it as a learning experience and the basis of further testing.

HTH

Cheers

[LucidSW]

Pavlov's portfolio analogy is a good one. Bottom line is, as he says, better to track performance for each theme.

[briggidere]

I have gone down the AdWords recommended route a few times with good success.

Theme them but have 3 AdGroups for each one. The specific target, a complementary target & the audience.

Don't stop there though. Once you have some data to go on, get into your placement targeting for each one as you can see which ones work best and target them specifically.

A targeted placement campaign can work wonders.