The scene has played itself out hundreds of times on police procedurals: a giant map of the city is tacked with red pins indicating where the banks have been robbed or the taunting notes left. The
detectives stare at the map and realize the bank robber/note leaver will likely next strike smack dab in the middle of the triangle created by the pushpins. It's geographic profiling. It works
for detectives and it can work for advertisers.
If you’re a local advertiser, mining your data for those geo-modified queries can let you capture new customers. According to a study by Oracle, "The majority of consumers ... now research at least half of all purchases online, a trend that encompasses categories from
technology to commodity products." That means that national brands with brick-and-mortar locations can dissect their online traffic, including campaign performance reports, search query reports
and analytics tracking, to learn more about the type of customers who will be coming to their stores (including virtual) in the very near future.
Local data reveals valuable demographic
information about consumers. The more precisely you can determine a customer's location, the more you can match it to existing survey and census data. Income levels, political orientations,
lifestyle trends can be vastly different a few blocks in any direction. This means your search query report doesn’t just list your trending keywords, but hints at the sort of customer with whom
your brand is resonating.
The first hurdle is disambiguation. There are over 3,000 places named either San Jose or San Antonio. There is a Greenville in most states in the
continental United States. Even the tiniest of towns has a "downtown." Highland Park, Old Town and Riverside are neighborhood names in dozens of U.S. cities and the names of dozens more cities,
themselves. Whereas syntax can assist in deducing meaning with everyday words, we have to rely on other clues to help with named entities, including relative population densities and a thorough
understanding of all of the various iterations of a given place. Action items include weighting by population, developing taxonomies for each instance of the city name, and more specifically limiting
your geographic scope to only those cities in regions where your customers live.
For example, an Italian restaurant in San Jose, Calif. would advertise under all variations of San Jose Italian
restaurants using the keywords that contain "San Jose." Campaigns settings will be for the U.S. so you won't get South American traffic – and, depending on complexity and resources, you can do
"exact" or "phrase" match for people searching on their mobile device from San Jose.
Another hurdle is effective contextual analysis. When is a searcher from eastern Ohio using alliance
as an abstract concept, e.g., "marty mcfly and doc brown alliance" and when is the searcher talking about her hometown, e.g., "find doc martens alliance"? This can be accomplished using a lengthy
negative list (be creative).
Finally, places and especially neighborhoods can be referred to by a variety of names and nicknames. It takes a great deal of research to find all of the
variations for New York City, London or Paris (the City of Light – not to be confused with the City of Lights, Aurora, Illinois).
Once parsed, advertisers can view their
data through the lens of location and overlay demographic details on their search traffic. A 60-day query report can be tinted with shades of lifestyle type, age range, household size and more. These
insights can inform logistics, marketing messages, store hours, and future targeted ad campaigns. For brands looking to move from search to traditional display, knowing the demographic of the traffic
you’re already monetizing can shorten the learning curve.
In reviewing your search query report, where IP addresses and other quantitative location data are not present, search text is
the key indicator of intent. This is the one exception where it's good to have a regional perspective, as geographic detective work will enable a clearer sketch of your next customer.