Business GIS application stories

Retail GIS: The Park Slope Food Coop
	The Park Slope Food Coop is, by many measures, the strongest coop in the country. Founded in the early 1970s, the Coop has a membership of over 5,300. Members work up to 3 hours every four weeks, and enjoy tremendous savings in food prices, as well as a terrific selection of organic and healthful food and other products, and the chance to take part in a cooperative venture. In 1997, the building next door to the Coop was on put on the market, and aggressively so. The Coop saw opportunity for growth, as well as the need for a strategic plan. The leadership of the Coop realized that a firm understanding of the membership, the neighborhood, and the competitive environment was required for effective planning. DSS helped them gain and act on that understanding.
Demographics Classic thematic map of median household income by Census block group	Park Slope: an evolving, dynamic neighborhood
	While the Coop had enjoyed steady growth through the years, Park Slope itself was a changing neighborhood. Where once the neighborhood had been largely working class, gentrification throughout the 80s and 90s was nearly complete, and Park Slope had become a fashionable place to live. Large national and regional chains have taken notice over the last few years, with Barnes & Noble, Blockbuster Video, Radio Shack, Rite Aid and others making investments. As far as grocery competition was concerned, large Super Pathmarks were built at the edges of the neighborhood, and a growing number of local stores were now offering organic and healthful products -- though at considerably higher prices. Finally, a small organic food delivery firm now serves the area at still higher prices, but provding the ultimate in convenience.
	Research goals and tools
	Our research goals were as follows: Create a demographic profile of our membership Create a trade area Create a demographic profile of our trade area Create neighborhood level demographic profiles Calculate potential number of member households per neighborhood Determine competitive influence (supply alternatives available) Calculate and locate net demand for Coop membership
	We used the following GIS and GIS-related tools: ArcView GIS Business Analyst Spatial Analyst Microsoft Access QMS Geocoder GDT geographic data Member survey data Competitive location data (hand gathered)
	Methodology
Geocoding Map shows geocoded member households in Brooklyn	Step 1: Geocode the members. We imported the member database into Access and cleaned it up a bit. For the most part, the data was in good shape, and had previously been householded. We achieved a 94.3% match rate to the street or ZIP+4 level. We used the QMS Centrus product with GDT streets for the best possible results. The investment in time up front making the member data as clean as possible also paid off. Bear in mind that urban geocoding is much easier that in suburban or rural areas. Step 2: Create a trade area. We taught each block group in Brooklyn how many members were located within it. We then selected block groups adjacent to the Coop location, and continued adding adjacent block groups until a total of 60% of the membership was accounted for. The classic goal in this exercise is to achieve an 80% customer threshold, but we use density rules to crop our trade areas. Yes, there are significant members who live outside our trade area, but they are thin in relative numbers per block group as to render them trivial for this purpose. So our trade area definition translates to the place where most of our members come from, where there is reasonable density of market penetration, and where we can expect to successfully market our Coop. Step 3: Profile the members. Using ArcView GIS, we assigned block group demographics to each household record found within that block group. Age, ethnicity, education level, gender, family type, and income were all considered. Step 4: Survey the members. We selected a random sample from our member database, creating a subset of about 10%, or 530. Of these, we completed approximately 300 telephone surveys, or just over 5% of the total membership. An additional 450 written surveys were also included. Not optimal, but not a bad sample either. Members were asked specific questions about their demographics in line with what we were studying at the block group level. Additional information was also gathered relating to shopping and buying behavior, attitudes about the Coop, and transportation issues.
Trade Areas Map shows core trade area for Park Slope Food Coop

	Step 5: Reconcile the member profile with the member survey. The member profile was tweaked to incorporate what we learned from direct survey, and a model was created which fairly predicted what we already know about membership patterns at the neighborhood level. We truly felt that we knew exactly what sorts of households were likely to join the Coop if given the chance, and we could calculate totals of these households at the neighborhood, or block group, level. We had a demand metric in place.
	Step 6: Competitive analysis. We were fortunate in that we were working in a compact study area, and could use a methodology that did not have to be replicated elsewhere. This meant that we could walk the streets and avenues to locate our competition, and actually walk into the store to estimate shelf space allocated to products which would be attractive to our membership. We created a competitive (supply alternative) metric, placing the value at the competitor location.
The big picture Full blown map, showing pockets of opportunity	Step 7: The GIS thing in action. We brought the following dots onto the map of our study area, Brooklyn, NY. Coop Location Member Locations Block Group Centroids, loaded with our demand metric Competitive locations, loaded with our supply alternative metric We used Spatial Analyst to calculate density of demand, density of competition, proximity to the coop, and market penetration. The map calculator was used to produce a weighted algorithm which produced a continuous surface of opportunity. Higher pockets of opportunity are shown on our maps as dark green.
	Step 8: Buy the building. Our basemap of opportunity understanding showed us that there was plenty of room for additional market penetration in our target segments in our trade area, as well as the chance to expand our trade area. Also, our phone survey revealed the chance to gain a greater share of the shopping dollar through merchandising and service. Further, we recognized that a sizable ethnic segment (Hispanics), was represented in our trade area, prompting discussions for alternative merchandising and marketing strategies. We considered satellite locations, and identified two neighborhoods which could probably support a second location, but chose not to take that great a risk. We bought the building, doubling our retail space to 40,000, and are now in the process of renovating the space. The new Coop will be open toward the end of 1999.