I found two examples on the internet:
Campaign 1: Region and City Campaign 2: National
Location Targeting: Columbia, MD Location Targeting: United States
Keywords:
- bank - columbia md bank
- banks - columbia maryland bank
- banking - columbia md banks
- savings account - columbia maryland banks
- checking account - howard county bank
- free checking - howard county banks
In the example above (campaign2) someone made a National campaign, isn't that double? because Campaign1 should be displayed anytime when someone's search phrase contains columbia.
Why couldn't I just make Campaign1 which is regional
then WHEN someone e.g. in New York searches for columbia md bank it would display my ad in New York because the newyorker has columbia in the search phrase.
I don't understand why campaign2 is needed.
Is the IP address of the searcher known?
If this is a yes, serve the ad.
If no, then did the searcher user a geographic keyword mapped to that location?
If yes, serve the ad.
if no, don't serve.
That's very basic, but should be a high level example.
So, then the question becomes 'is the geographic keyword mapped to that location?'.
We'll use your Columbia example.
Off the top of my head (so I'm sure I'm missing some), I can think of:
columbia, md
columbia, mi
columbia, sc
To confuse matters, what about:
columbia records
columbia house
So, in this instance 'columbia' can be used to describe both a geo area and brand names.
Columbia could be used to describe multiple geo areas.
Therefore, I don't think Google should serve a geo ad for columbia searches as there isn't a high confidence factor that the proper geo is known or that it's actually a geo search.
In instances like this, it's up to the advertiser to make sure they have a national campaign with geo-keywords so that there ads are shown when it's appropriate.
