The report has potential, but you do have to use some caution in interpreting the Metro and City level data. If I was to believe the stats it's giving me, I would think that people in cities love to click on ads (50-100% CTR) while those in "(all other metros)" rarely click on ads (well under 1% CTR).
I'm sure that's just a data issue rather than a reflection of reality. Maybe it's just due to the fact that the report is new and they're still working on making the right data accessible in their database. But until it's resolved, the impression and CTR stats at the Metro/City level are not very useful.
This covers at least some old data. I'm not sure how far back it goes, but I just ran a report for August.
[adwords.google.com...] says "data for the Geographic Performance report is only available from May 27, 2007".
I aggregated up to the state level in Excel and I'm seeing 20% of impressions listed as "(all other regions)". This is for a US only content network campaign.
I wonder if that means Google doesn't know where those people are located or that those folks are international.
JBrown, do you think the 20% could be AOL users?
I run a report on a campaign where CTR is around 40%. Other regions is 0.84%
It would be interesting to know where those ads were showing. Or they are maybe impressions with invalid clicks filtered out.
JBrown is that respective campaign targeted to english or all languages?
It's an English language campaign.
Some other thoughts...
1. I've looked at search and content campaigns and both have significant data in the (all other regions) bucket. It's running around 8% of clicks, but a much larger share of impressions (20%-70%).
2. Is setting up a geographically targeted campaign the only way to bid differently on geographic areas? Looking at the data, it might be helpful to adjust bids higher or lower for certain states. I would like to be able to bid geographically within one campaign. If that's not available now, it would be a great new feature.
|I would like to be able to bid geographically within one campaign |
This would require geographic settings on an ad group level, something what MSN has. I'm in favor.
Another data piece I find cool to have in this report (besides impression share) is Average Position.
Since Average position is available in other reports as well as in managing interface, it would be great to have it within this report.
It seems the only 'unit of time' option for this report is 'daily'. I would much prefer a summary for a chosen period of time.
koncept, Google has some explanation in their help pages why 'daily' is the only option. I've found it fairly easy to aggregate up using a sumif function in Excel.
if your willing to do a public service
turn off search partners JBrown (for a day or 2)
some dont handle targeting properly
or at all
which skews the number
and report back
For so long we've been waiting for search partners report.
Those that are called "search partners" but do not convert not only into sales, but even into clicks, who needs them? Why bother?
actually id be interested in targeting ask.com
i actually think i may have an advertiser thats perfect for
rushmore drive (lol) the black search engine
some of the search partners arent bad but....
would you want a shake made with half icecream and half sludge
It wouldn't surprise me if some of that is search partners. There is also missing data for the content campaigns. I would like the data, but I'm not very concerned since that "(all other regions)" traffic is still converting well.
For whatever reason, it looks like Google has a hard time categorizing certain impressions and then lumps them into "(all other regions)".
Well, I ran a geographic report.
Not sure what I am supposed to do with all this information. Seems to me this is just more data that adds to the blizzard of data I already get from our websites. I have way more information than I could possibly use at this point in time.
Who knows, someday I may find a reason to use it but right now it really tells me nothing new.
I am suffering from statistics burnout.
I'll give you this; when I search at www.google.co.uk from Malaysia; I get ads targeted only to the UK. I know this because it includes my own UK-only ads.
I'm guessing that this geographical report is based upon real geolocations rather than the localised Google site chosen.
google uses the following for search localisation in no particular order
regional site ie. google.ca or google.co.uk
dns server (if you use open dns sometimes google cant pin you down)
keyword ie. new york car rental
previous search ie. (1st search buffalo NY) (2nd search sushi) these queries are random (it may not work) however i have seen reliable proof
past search history
my also have an effect
google account history
may also have an effect
|I wonder if that means Google doesn't know where those people are located or that those folks are international. |
FWIW, my ISP (not AOL) is located in a large city about 200 miles away yet I get obviously local ads for that city.
My question is how reliable is the data - does Google really have a reliable method of knowing where people are located?
Shame the report can only be broken down by day. I've just resorted 7000 lines to get a summary of CTR by region.
I'm seeing very interesting differences in CTR by country. Guam, Romania and Bosnia and Herzegovina have a CTR next to nothing.