GCensus: Let the Real Local Targeting Begin

Spotted via John Battelle: GCensus is a Google Earth mashup of census data.

There are lots of already folks layering census and demographic data onto dynamic mapping interfaces. I discovered this at last year’s “Location Intelligence” conference (a small show) in SF. Most of it is highly specialized or very B2B and thus not visible to general audiences.

Beyond the interesting consumer-facing things that can be done, what Census data permits on the “back end” is lots more targeting using location as a proxy for other things. Here’s my prior write up of AdCenter’s demographic targeting from the launch last May:

Let’s assume the user is online in his/her home (putting aside that affluent users might not be adopting free Wi-Fi). Once you know where the user lives, you can layer Census data on top of location. Suddenly you know things like:

  • Household incomes
  • Home prices
  • Number of children in the home under 18
  • Numbers of home offices
  • What percentage of residents take public transportation

This is only the beginning of the data that are available at the neighborhood level. You can start to see that geotargeting takes on a whole new character with this kind of data. But all of that depends on locating the user with precision, which is in part what Google’s Wi-Fi initiative is about.

One might expect that advertisers would pay a premium to target only those neighborhoods where high-probability customers were. If I’m a financial planner, for example, I might like to target prospects within 15 miles of my office who have incomes of more than $100,000. I could use this hypothetical enhanced local-targeting system (with information on income leves) to only advertise in those neighborhoods where my potential customers were (e.g., “Pacific Heights”) — as opposed to a broader geographic area (e.g., SF Bay Area). The benefits are obvious for local realtors soliciting potential sellers to list with them. And if I’m a regional car dealer I could similarly use this enhanced local targeting to reach only those potential prospects who fit a would-be buyer profile for VW Jettas or Toyota hybrids.

Better local ad targeting, as these examples hopefully indicate, is ultimately about demographic targeting. AdCenter’s debut will put some pressure on Yahoo! and Google to offer more and better targeting options to advertisers, which they’re now working on and rolling out to varying degrees: Yahoo! with behavioral targeting and Google with demographic targeting by vertical.

When you combine local with Census data you get all sorts of other potential targeting capabilities that are tied to local. That’s really what’s at stake and what’s significant about the marriage of maps and Census information.

One Response to “GCensus: Let the Real Local Targeting Begin”

  1. Aidan Says:

    Good analysis – I’ve worked on customer loyalty programs where this sort of data is critical to good targeting – in fact, the perception that targeted needs to mean one to one is mis-guided. Because of my interest, I’ve started a project similar to GCensus, in which I’m working to generate Google Earth/Map KMLs of Census data for the entire United States. You can track my progress at http://censuskml.blogspot.com/.

Comments are closed.


%d bloggers like this: