As you can imagine, the list of widgets with each of the target cities, states I have make my keyword list in the thousands.
Using dynamic keyword insertion, this is really easy, but from what I'm seeing, putting tons of keywords into a single adgroup is a bad idea.
So is it better to create a single adgroup with dynamic insertion or creating hundreds of adgroups with just 2-3 keyword permutations in each one? Is there a difference performance or cost wise to doing either?
Thanks everyone!
Using dynamic keyword insertion, this is really easy, but from what I'm seeing, putting tons of keywords into a single adgroup is a bad idea.
It can cause problems from a performance standpoint - where lots of keywords can create problems downloading or uploading changes - or from a quality score standpoint - where lots of keywords can group poorer performing keywords with better ones. As far as performance problems go, I've seen the negative impact of having over 4k keywords per ad group, but I've never seen a downside to as many as 1-2k.
So is it better to create a single adgroup with dynamic insertion or creating hundreds of adgroups with just 2-3 keyword permutations in each one?
No, I wouldn't go that far in either direction. I'd structure your campaigns and ad groups in a way that's manageable, scalable and takes performance and quality score in mind. Sounds easier than it is, I know, but there's a happy medium that addresses all three concerns but doesn't cause unnecessary inconvenience in there somewhere. I'd just sketch out the different constellations... themes... rough quantities... ad text needs, etc. and determine the best way to slice the groups at the onset.
You've done upwards of 1-2k keywords per adgroup without notice adverse effects? I ask because i've seen recommendations of 20 keywords per adgroup and that seems way too low when I have thousands of keywords like this.
Hi peermedia - It may depend on your limits per account, but for my particular settings (1MM keyword cap per account) - I've never noticed any ill effects from 1k to 2k keywords per ad group. My cap is 5k keywords per ad group, so I'm well under the limits even at these levels. While limiting ad groups to 20 keywords would be very targeted - depending on your number of keywords, it's not scalable.
poster_boy, from a quality score standpoint what is the downside of grouping poorer performing keywords with better performing keywords as I always thought quality score was on a individual keyword level basis. Is there also a quality score assigned to the ad group, campaign or account that takes into account groups of keywords?
Hi kobayashi - My understanding is that Quality Score does reside at various levels - keyword, ad group, ad text, campaign, account, etc. While keywords have unique Quality Scores - I do strive to bundle similar types of keywords per grouping in order to minimize the variance to a degree.
If you do, then you will probably find that very carefully groomed keyword lists, in which all keywords in an ad group are about a common theme (which also happens to be the common theme of the sorts of pages on which you'd like your ads from that ad group to appear) will probably perform best for you.
If the content network is especially important to you, then thousands of keywords will probably not be work to your advantage. It'd be better to have 20 laser-focused keyords than 2,000 unfocused keywords, IMO.
AWA
I have somewhere around 15000 keywords total, between myself and my clients, and even though it's more work to initially set up, I only have a handful of ad groups with more than a couple hundred keywords in them, and most have far fewer. As of this morning, I have 17 "poor" quality scores across the whole shmear. So it's more work in the beginning, to be sure, but a lot less work chasing inactive keywords and poor quality scores. YMMV.
