This map flags zip codes
areas which contribute customers to "M" restaurant,
but doesn't show the number of customers. The dot-density map
offers this later information (Fig. 2).Olga Medvedkov
This project is an example of cooperation between two academic departments, Geography and Management, at Wittenberg University. Other parties in the cooperation are business communities of Urbana and Springfield, and the State of Ohio. In this project students from Business Geographics and Marketing classes were brought together to solve real-life marketing problems. They helped a local restaurant in targeting more clients. With the results obtained in the project the restaurant's owners were able to achieve much better response from adverticements mailed to potential customers.
An old family restaurant "M" moved to a new location, and its owners were trying to increase their cash flow to cover overhead expenses. Urbana is a relatively small city (population 11,000) and a move over 2 miles (from downtown to periphery) does make a difference in maintaining the old customers and attracting the new ones.
We had several tasks to accomplish:
The data came from different sources. Our client has provided a data base which required some improvement. In addition our students conducted license plate surveys. During several weekends they canvassed parking lots of about 20 different local restaurants-competitors and wrote down license plate numbers of all cars entering parking lots. The same procedure was repeated for our client's restaurant. The tables with the license plate numbers were submitted to the State of Ohio Department of Motor Vehicles for obtaining car owner's addresses. We expanded the initial data base for the customers of each restaurant. In order to define different customers "profile", we used Census data, 1990 and postal data. A combination of client data, Census data, postal, and field surveys data is necessary in this kind of projects. Usually this stage of the project creates one of the stumbling blocks and requires cooperation between different forces.
All our data is highly geographical in its content. We have the client's restaurant at a particular location as well as whereabouts of its competitors. The data base for the customers is space referenced by street addresses and zip code areas. The social, demographic and economic Census data is also organized in accordance to geography for different territorial units, starting from the national level, state level and all way down to urban blocks level. Postal data is aggregated for yet another geographical units, zip codes areas. In other words, geography is the major organizing aspect in our information. The GIS analysis with its spatial data manipulation is clearly the most promising tool for our project.
As the first step, we mapped our client's customers by geocoding them on the zip codes level (Fig.1).
This map flags zip codes
areas which contribute customers to "M" restaurant,
but doesn't show the number of customers. The dot-density map
offers this later information (Fig. 2).
The map shows where the majority
of customers are coming from, and how they are distributed across
different places. It is obvious that the majority of customers
live in Urbana itself and in Springfield (the bigger city of 70,000
population, 15 miles away from Urbana, with 11,000 population).
Since the zip code area where Urbana is located is bigger than the city,
we had to conduct our project on more detailed scale, on a street level.
Our second step was to geocode customers by their street addresses in Urbana and Springfield (Fig. 3)
. The maps revealed that our
restaurant's primary catchment areas make two belts: one belt
stretches North-East from downtown Urbana, the other covers the North-West
sector of suburban Springfield.
Our third step dealt with the "M" competitors located in nearby area. We selected the restaurants which are similar to our client "American cafeteria of 60's" atmosphere, by the type of cousin, services, etc. We surveyed several fast-food chain restaurants which created competition by intercepting our client's patronizes. Since our client moved to the periphery, east of downtown, a potential customer can see quite a number of intervening opportunities on his/her way from downtown to cafeteria.
Our fourth step was to geocode customers for competing businesses,
and to overlay these maps with the previous once (Fig. 4-5).
The fifth step was to define our customers "profile". We created several choropleth maps reflecting the gradient in median income, age, education, family size, ethnicity, etc. These choropleth maps are detailed by zip codes areas and by block groups in the case of Urbana and Springfield (Fig. 6-8)
. The choropleth maps reveal that
the most common customers patronizing the "M" restaurant are middle
class people with median income ranging between $30,000 and $45,000.
They are, typically, of middle age, predominantly white, often empty nesters.
Using buffering technique we could show the distribution of customers
within five miles from our client's restaurant (Fig. 9).
On the base of our maps we advised our client to do direct targeting of their current customers, and to expand their catchment area in Springfield where there are niches for potential new customers similar to current customers.
Our client has conducted direct mailing and got a response which exceeded initial expectations. Now, the competing restaurants are looking forward to get our maps.
Obviously we have satisfied our client, but of no less importance is that our students were exposed to real-life problems, and they contributed to finding a solutions. The students could learn new concepts and new techniques which are applied immediately in a real life situation and prove to be successful. It was a great hands-on experience for the class.
Acknowledgments: I would like to thank Lisa Murphy, VP for academic affairs at Business Geographics American Association, who encouraged me to teach Business Geographics, Pam Schindler, Professor from Management department at Wittenberg University, who agreed to combine her Marketing class with my Business Geographics to conduct joint project, and our students without whom this project would not be possible.
GIS for Business and Service Planning (1995), edited by Paul Longley & Graham Clarke, published by Bell and Bain, Glasgow, UK
David Martin, Geographic Information Systems: socioeconomic applications (1996), Routledge, UK
Mark Monmonier, How to Lie with Maps (1996), The University of Chicago Press, Chicago and London
Lisa Murphy, Why aren't Business Schools Teaching Business Geographics? Business Geographics Journal, February 1997, Volume 5, number 2.
Profiting from a Geographic Information System (1993), edited by Gilbert H. Castle, III, Published by GIS World. Inc. CO, USA