The CTR on the winner was 21% and the loser almost 15%, so I was a little surprised that this dropped the minimum by over 30%. So, it go me thinking. Certainly the CTR is high enough to justify a low minimum bid but there have been times when the CTR has been higher than it is now and the minimum bid did not change. So, why now?

I think I figured it out. The winning ad’s copy is a paraphrase of the landing page’s headline. The ad that was paused also reiterated copy from the landing page… but that copy is in an image and appears further down on the page. I think there is something to be learned there.

Although, I have paused the split-test, this is a powerful testimony to what can be accomplished using the technique. The winning ad was put together quickly; written only to serve as the sacrificial lamb to the ad that I had painstakingly composed… or so I thought. Because of it however, we are now enjoying a fifth of the traffic to a keyphrase that when converts can produce revenue deep into the 5 digit range… and paying nearly half of what we initially bid for a click.

Both of these suggestion helped maintain a good quality score and .04 cent minimum bids, but in limited testing I was not able to beat the CTR of the redirected URL. I will continue to experiment with them.

But if at first you don’t succeed… try, try, again. I finally found an ad that has showed some promise and early in the test is defeating the www.geocategory.com set up… and it uses the original domain, which is even better. What I did was add the geography and the telephone number to the 3rd line of the ad. I tried to make it look like an in-column yellow page ad. The 3rd line reads: Fillintheblank County… (123)456-7890. I was surprised that Google allowed telephone numbers in their ads but I now understand why they do… they are good for click-thru rates! Google will allow 3 consecutive periods but not 4… between the geography and the number.

These results would not give me the warm & fuzzies if I were running a campaign that was geo-targeting Newark.

This is important because the purpose of Geo-targeting is to put your ad in front of a person who is using a generic search query that has local intent. For Example, “Dentist” as opposed to “Dentist Newark NJ.” Studies have shown that half of local queries are entered without the geographic indicator. And Geo-targeting should give you a leg up in trying to capture these sales. The above Google Trends report argues otherwise.

If the IP addresses of searchers shopping online in Newark are in Oakland, NJ and Clifton, NJ; advertisers’ results are going to be adversely effected. Such campaigns would be plagued by low CTRs and/or wasted spend.

Take-Aways for when you’re targeting a small geographic region with adwords (City, County, a few towns etc.):

• Don’t rely on Geo-targeting. Use keywords that include the geography you’re targeting as well.
• Put Geo-targeted ads in a separate ad group, so that there lower CTRs don’t effect your higher performing ads.
• KNOW the addresses of the local Cable/DSL provider in the area where you are going to Geo-target. Many of the cities listed above in the Google Trends report are those where Cablevision is located; or in the case of Madison, NJ… where the IP originates for Verizon DSL customers. My IP is located in Madison, despite me being some 30-40 miles away from there. And Oakland is a small town about 50 miles to the north of Newark; but the IPs of many Cablevision customers originate from that town. The results of a local campaign could change dramatically by just adding or subtracting one of these important but sometimes very small towns.
• TEST!

Some Assumptions I Made Interpreting This Data:

1)People who most frequently use ‘Newark NJ’ as a search term, do so because
they live in or around Newark. 2) And that group would be representative of those who are shopping locally but use the generic search term.
Because we are studying this from the “reverse angle.” Google trends is showing us the location of the IP address for those who type Newark. Geo-targeting in Adwords would return results based on the IP address of a shopper who did not type in Newark. 3) Google uses the same IP tracking in Google Trends as it does in its Adwords system.

I believe those assumptions are true based on the results I’ve seen personally in the geo-targeted advertising that I’ve done, though admittedly a small sample. But the more relevant cities I saw in Google trends for the applicable geography, the higher the CTR of the advertising…. again very very small sample.

For those who are running “tight local campaigns” could you do me a favor? Please use Google Trends and search for a town or 2 in your area
and let us know if you find anything that would make you agree or disagree with what I’ve asserted in this article. Thanks!

Good Additional Reading on Geotargeting:

Geotargeting Adwords by The Merjis Blog

Geotargeting: Core To Local by Chris Silver Smith