Selecting local markets more accurately with geocoding
There have been many new advances in computer technology, allowing for new methods of selecting address lists geographically. Many database marketers and list compilers, however, still restrict their geographical selections to ZIP codes, counties or city names. Often your target market may be defined by boundaries that do not fall along one of these entities. Geocoding allows you to convert postal address data into longitude/latitude coordinates, enabling a more precise selection.
There are two primary types of geocoding: forward and reverse. Forward geocoding converts a postal address into a set of longitude and latitude coordinates. This is done by comparing the postal address entered against a vector database/map of a road network. I'll be focusing on forward geocoding in this article, although reverse geocoding is useful as well.
There are several sources of national vector data for the road network in the United States. The two largest commercial sources of such data are companies Navteq and TeleAtlas, both of which make their data available for a fee. This is the same road data that is used for vehicle navigation solutions such as in-car and handheld GPS units.
There is also free vector data available from the US Census Bureau in the form of the Tiger/Line Data. A notable difference between the Tiger Line Data and the commercial data sources is that it lacks some attributes, such as one-way street designations, which are useful for vehicle routing. However, if our purpose is to locate the longitude and latitude positions of addresses on a map, it can be quite useful.
In order to make use of the road network vector data, to convert addresses into points or the reverse, you will need a piece of software called a geocoder. A geocoder works by parsing apart the postal address entered into its component parts, and then looking up those parts against the list of streets in a given area. For example, it might pull a list of all streets in the parsed ZIP code component, and then match against the street name. If a match is found to the street name, it would then look at the house number and find out how far down the street the address is and then interpolate to find the actual position.
Once you have appended the latitude/longitude coordinates onto a data file, you can then use spatial querying software to make selections from a graphical user interface. Being able to select through the process of drawing a geometric shape is generally far more precise than being limited to the traditional means of selection, which are already present on the postal data, such as ZIP code, or carrier route.
You can also combine this selection technique with demographic fields present on your file to allow for additional selections. New tools allow for all of these processes to take place within a simple user-friendly interface that allows for advanced data selection by a non-technical user.
Brian Carroll is chief technology officer of DirectMail.com. Reach him at email@example.com.