I’ve been carefully reviewing some very interesting data that has come to my attention via a presentation done for a SMX Advanced Presentation By Marta Turek. I would highly recommend that you read this very interesting article Google AdWords Geo-Targeting: Have We All Been Doing It Wrong?
Here’s what the author tested in a quick summary:
1. If you advertise at the state level, do you get better results excluding metro areas by opting-out/exclude or better results by not targeting at the state level and opting-in at the metropolitan DMA region?
2. What impacts do each strategy have on click cost and conversions?
The results of these two limited tests are absolutely shocking and unexpected. The big caveat to fully embracing these results in accounts you manage is that the testing has been done in a very small set of clients and so although may show a trend worth investigating may not be a good approach for every client.
Here are the test results as stated by the author in a quick summary:
1. When the areas in the state were targeted by opting-in, the number of impressions and clicks significantly dropped while the cost per click drastically increased.
2. When the same areas in the state were targeted using an opt-out strategy, impressions and clicks were not impacted and conversions actually improved.
My explanation for the results:
I see several specific reasons for the results. First when you target using opt-in a specific area such as a region, city or even radius, Google appears to believe that this is valuable to you and in many cases will in turn charge a higher cost per click. Remember this is a supply and demand market.
I have seen this myself when I run simultaneous programs running a state wide targeted campaign with location specific phrase match keywords compared to a radius targeted or city specific opt-in program. In the account I tested the differential in cost per click was pretty significant – $4.50 at the state level and $10.50 to $12.50 at the city/radius level. So this validates the findings that the more granular you get in targeting, the higher the click cost.
As to increase impressions, clicks and conversions, I believe that how Google serves ads and how Internet Service providers categorize locations is a key factor in the results. I know from experience that there are many cases when an Internet Service provider does not return to Google the true city location for a user.
Here’s my own case, I am located in Waldorf, Maryland. Verizon has returned my location several years ago as Reston, Virginia. That’s another state and over an hour and a half drive north and west from my true location. In the last month, I see Verizon now returning my location as Temple Hills which is about 30 minutes north of my location in an economically depressed region. By targeting a state with opt-out you open up your reach and slightly negate the impact of improperly reported locations by Internet Service providers.
It will take additional testing in your client accounts to see if these same trends ring true for your needs but a worthy time investment for some accounts that target locally and not nationally. If you are looking for an experienced AdWords account manager, I invite you to review our programs.