How to calculate demand for retail

By one estimate, roughly 1,500 new housing units are needed to build one new block of stores. Many communities hunger these days for “Main Street retail” or “neighborhood retail.” Few, however, know how many households — and shoppers’ dollars — are needed to make a shopping area successful. Retail has posed a persistent challenge for new urbanists. Goody, Clancy Associates, a Boston architecture and urban design firm, has used its experience to devise a mathematical formula for retail and housing. “It’s important to examine these sorts of relationships, particularly when you’re trying to revitalize older neighborhoods or to create new communities,” says David Dixon, head of urban design. “This is not a science,” Dixon cautions, “but a way to determine order-of-magnitude relationships. The broad numbers I usually use are meant to provide a range and to emphasize the imprecise nature of the calculations.” The example below, from a study in eastern Cambridge, Massachusetts, lays out the basic steps Goody, Clancy uses to match retail to housing. Start with the volume of retail you’re aiming for. In Cambridge, Goody, Clancy determined that the blocks were about 300 feet long and assumed that the stores’ average leasable depth would be 60 feet. (Housing was to be built above the retail; a 60-foot depth is suitable for upper-story apartments on a double-loaded corridor). Assuming that the retail would occupy both sides of the street and would fill 80 percent of the street frontage, Goody, Clancy estimated that the block and depth dimensions would produce 30,000 square feet of retail. “For many revitalization projects, it is appropriate to look at the needs of several blocks, so 100,000 square feet might be a planning goal,” Dixon notes. “Similarly, if one wanted to create a neighborhood-scale commercial center for a new community, 100,000 square feet might be a reasonable goal. Next determine how many dollars of sales per square feet are needed to support the retail. In eastern Cambridge, the range needed was $300 to $400 of sales per square foot — partly a reflection of the area’s prevailing retail rents. Multiply the midpoint of those two figures ($350) by 30,000 square feet, and you discover the gross sales required for a block of retail. In this case, it’s $10.5 million. In many communities, construction costs and rents are lower, allowing the retail sales figure to be lower — perhaps $200 to $300 per square foot. Next determine the percentage of disposable household income spent on neighborhood retail purchases, and therefore the total household income necessary to generate sufficient gross sales. In the Boston area, Pam McKinney of Byrne, McKinney real estate consultants estimated that roughly 60 percent of household income is disposable income (i.e., income after taxes and housing costs have been deducted). Thirty-five percent of disposable income is spent on retail purchases. Fifteen percent of the retail spending consists of purchasing in neighborhood or Main Street establishments, those close to home. Therefore, divide the total sales needed ($10.5 million) by .60. Divide the resulting figure by .35. Then divide that result by .15. This produces a figure of $333.3 million — the total disposable income needed to support 30,000 square feet of retail. The percentages would not vary greatly among different markets, according to McKinney. Then determine the number of households, and therefore housing units, required to produce $333.3 million in disposable income. McKinney assumed the average household income for new dwellings in eastern Cambridge would be approximately $75,000. Dividing $333.3 million by $75,000 reveals how many households are needed: 4,444. “Very high- or low-income communities skew these numbers,” Dixon points out. “For example, a HOPE VI community that includes 50 percent public housing residents might have an average disposable household income of less than $40,000 and require roughly twice as many housing units to provide the same degree of support to retailers.” Finally, determine the percentage of the required units that must be located within walking distance (approximately 10 to 15 minutes) of the retail to provide core support. This is a judgment call based on how much of the customer support is local and how much will come from farther away — from people who see it as a destination worth a longer trip. In eastern Cambridge, McKinney projected that roughly 25 to 35 percent of the retail sales would have to be generated by new housing nearby. The rest would be from drive-by and other shoppers who would be attracted once the retail was operating. The conclusion was that eastern Cambridge would need 1,200 to 1,500 new housing units to support one block of retail. Generally, the proportion of financial support that must be generated locally ranges from 25 percent to 75 percent; for isolated retail in new communities, it may be 100 percent. How does all this shake out? “I think it fair to say that in a great many situations, a block of new retail would require approximately 1,500 units of new housing within walking distance (plus or minus as many as 500 units, depending on the factors above)” Dixon says. “The principal exceptions are low-income or isolated communities, which require a larger number of units, or higher-income communities, which require fewer. A 100,000-square-foot neighborhood center could require roughly three times as many housing units. The larger the center, the more it can become a destination in its own right and not need as many households within walking distance.”