As Google's expanded broad match broadens its horizons, advertisers continue to suffer. The rationale behind the change is that Google is doing the legwork for you, so you can get more clicks and impressions off a small selection of keywords. The example provided by the Google Adwords Team was "Honolulu hotels," showing how, when broad-matched, you could reap the benefits of traffic from "Hawaii Hotels" and "Honolulu inns," etc. What they failed to mention is that you may very well show up on "Austin Hotels" or even "NYC Hotels." Unbelievable? Well it should be, but tell that to www.nyc.hotel, whose paid search placement was appearing on "Honolulu Hotel."
tremendous issues have popped up with paid search campaigns in the automotive industry. In this example, the advertiser is regional, so there are two Adwords campaigns. First, a
geo-targeted campaign includes general words such as "Toyota" and "Corolla." A secondary nationally targeted campaign is leveraged to gain volume potentially lost due to anonymous IPs or to IPs that
are not able to be targeted from certain Internet Service Providers. In order to limit the risk of unqualified clicks in this nationally targeted campaign, only very specific keywords with
geographic modifiers ("Cheap John's Toyota Farm Boston," "Massachusetts Toyota Dealer," etc.) were placed in this second campaign targeted to the United States.
Frantic reports began coming in from regional sectors outside of the ads' targeted regions, which were showing on keywords such as "Toyota dealership Baltimore" and "Pittsburgh Toyota" and, even more alarming, "Toyota Solara." Since these regionalized sectors share a domain name, these discoveries caused controversy among the regional dealers whose ads were being blocked within their approved geographic areas.
The tech team at Google looked into the issue, and the results came back simply as an "expanded broad match issue." Apparently, "Cheap John's Toyota Farm Boston" is the same as "Toyota Solara" and, logically thinking, Pittsburgh=Boston. Therefore, searchers across the nation were seeing a targeted localized ad countrywide.
There were several options presented as potential solutions:
1- Build a robust negative keyword list. In order to avoid ad serving in areas outside of the target, each locale not targeted must be added as a negative word. The U.S. Census in the year 2000 counted 25,375 places in America. A campaign can have up to 10,000 negative keywords -- so there goes that option.
2- Phrase or exact match the entire campaign. This will involve massive keyword build-out, or you risk losing out on relevant volume.
3- Run Search Query Reports to isolate unwanted traffic. Have you run one of these? Granted, this report is still in its infancy, but at this point it is not very helpful. 18% of my clicks came from "1457 other unique queries."
As expanded broad match matures into a bigger problem for advertisers, one question continues to come up: When is Google going to do something about it? Unfortunately, we are all bombarded with Google's reports of how beneficial it is for most advertisers. Since articles like this one now appear in various industry publications on a weekly basis, it is hard to believe that "most" advertisers are reaping benefits from this feature. Whatever happened to providing the best possible user experience?