The Science of Retail Site Selection

By Tom Ryan

A new crop of modeling tools and data analysis applications are making retail site selection much more of a science for retailers.

Wegmans is supplementing its traditional gravity modeling approach with deeper data such as income level and ethnic makeup. Kroger is staying on top of population trends by using a demographic package that’s updated quarterly. Lifestyle segmentation and psychographics are also playing an increasing role in site selection.

More chains are also relying on in-house data from loyalty card programs to estimate the impact of building new stores in markets where they already operate, thereby helping avoid cannibalization. There are even tools measuring store loyalty to gauge how a new store will steal sales from an existing one.

Like other retailers, grocers are also beginning to realize that real estate decisions are “joined at the hip” with merchandising and marketing, notes Matt Germain, Acxiom’s client representative for the retail industry.

“Companies like Whole Foods and Trader Joe’s understand,” he noted. “The closer a retailer gets to a neighborhood concept, the more they need distinct household-level data.”

Site selection has become much more complicated for grocers in recent years, due to factors such as encroachment by alternative retailers, fiercer competition in general, and the drastically longer times it takes from initial site selection to store opening.

“It’s challenging. The easy sites are gone,” Paul Gilbert, director of real estate at Wegmans, told Progressive Grocer. “And if you’re looking at a high-growth area, you can almost guarantee that your competitors are going to build there, too.”

Yet with all of the promising new tools, retailers must also walk the market.

“The amount of data available from different vendors doesn’t substitute [for] the amount of time an analyst or consultant must spend in the field reviewing competition, understanding strengths and weaknesses, knowing where the growth areas are, and knowing what the pricing in a certain market is like,” Bob Gorland, VP of Matthew P. Casey and Associates, told Progressive Grocer.

This involves getting to know the intricacies of a market from local builders and developers, websites that offer information about building and construction permits, and real estate specialists.

“As sites get more scarce and building costs and break-evens go up, it becomes even more important that you’re getting good, competitive data, and good data from planning agencies,” Mr. Gorland said. “The last thing you want to do is look at a site and not dig hard enough to see that a Wal-Mart Supercenter, Costco, Trader Joe’s, or an independently owned ethnic grocer is looking, too.”

Discussion Question: What do you think of the promise of some of the new tools being used for retail site selection? What are some tried-and-true factors in site selection that may be getting overlooked?

BrainTrust

Discussion Questions

Poll

9 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Leon Nicholas
Leon Nicholas
16 years ago

Site selection tools, in my view, are only as good as the data that they utilize and the way that data is modeled. One can incorporate irrelevant variables into a fancy software suite and produce colorful bubble charts that point one in the wrong direction. Similarly, if the model employed in the software isn’t reflective of the drivers of store demand dynamics, then even good data will be mis-applied. The other element I would point to is the use of historical, as opposed to forecast, data. With rapid demographic changes occurring across the landscape, believing that the past will be an indicator of the future for site areas is likely to generate mis-fires on new sites.

Site selection is not a video game, in other words, where one is enamored by visual effects made possible by the underlying software. An interface with sex appeal should not be masking poor data quality and unsophisticated modeling–the real problems when it comes to site selection.

Mark Lilien
Mark Lilien
16 years ago

Regardless of the tools used, the #1 reason bad sites are selected: a goal of too many new locations. There are 2 major errors: growing too fast and growing too slowly. Error #1 is by far the most dangerous. In contrast, retailers who build very few new locations each year have the luxury of choosing the best possible projected return on investment, and they don’t strain their existing stores, either. IKEA builds very few new stores. They don’t seem worried about being in a footrace for real estate. When Wal-Mart and Barnes & Noble announced their domestic new location slowdowns, investors praised them, knowing that only the most profitable alternatives would be selected. Is it better to have a chain of 24 stores, 17 of which have mediocre performance, or a chain of 7 stores, with only 1 mediocre performer?

Kunal Puri
Kunal Puri
16 years ago

Interesting topic–reminds of me a famous story/rumor in India.

What is today the largest Indian Private bank wanted to install a large network of ATM in the early ’90s when they just went private…they spent about a year on site selection and analysis–then the CEO had a brainstorm–simply identify the best available location near every competitor ATM. Worked brilliantly for them….

But yes, it’s not always repeatable and hence the need for the site selection software products–their output should always be distilled with commonsense!

Jon Seltzer
Jon Seltzer
16 years ago

Given the investment of time and money a new store requires the additional data cost for quarterly data, consumer segment data and other relevant information is money well spent. Have you ever heard someone say, “Gee, our store is performing well. I guess we spent more money on site analysis than we needed to”? However, how many times has someone said, “I wish we had looked at more data before we built the store”? Data is as reliable as its source, and the more you work with a data-set the better you understand how it can be best used.

Dan Raftery
Dan Raftery
16 years ago

This is not a new skill being honed by supermarkets as the lead-in suggests. Successful retailers have traditionally been able to harness growth by knowing where to build stores. All that’s changing is reference sources available to expanding retailers. They still need to do their homework to find optimal sites, just like they’ve done since the first store owner decided to open a second location on the other side of town where the developers were planning new housing tracts.

The new databased resources are great, but surely can’t be as predictive as digging into the local planning commission’s minutes and talking to a few village trustees about their take on the future. New sources are great, but their availability doesn’t automatically make all of the old sources obsolete.

David Livingston
David Livingston
16 years ago

One of the best tried and true methods that is often overlooked is simply looking for over-performing stores operated by an under-performing retailer. For example, if there is a successful Winn-Dixie, Ahold, or A&P store, these are usually extremely vulnerable to a new competitor and easy to take volume away from.

Getting population updated quarterly is probably overkill. Only a few isolated areas are growing fast enough to be concerned with. Second, the source of the data is often questionable and unreliable.

Other demographic information such as household size, race, and income can be formulated into the per capita expenditure levels which are then part of the gravity modeling process. Wegmans has already been doing this for many years.

I can tell you that Bob Gorland and Paul Gilbert are probably two of the best analysts in the country. Retailers would be wise to listen to them.

Elliott Olson
Elliott Olson
16 years ago

With the rapid increase that we have seen in technology, communications and the internet, it is now possible to gather more precise information than one ever thought possible. One needs this information in an age where the lifecycle of retail concepts is becoming shorter and niche marketing, specialty stores and conflicts with various classes of trades, are becoming more frequent.

While we do not dispute the need to evaluate and quantify the physical characteristics of a site and trade area, we find that the market is being separated into two camps. The first group is based on the old gravity modelers who still enjoy getting into the old biplane and flying by dead reckoning and the savvy analysts who realize that they can’t run the modern models with out technical support, just as the modern pilot can’t fly the f-15 and drop smart bombs without GPS and computer controls.

The comment that quarterly population figures are questionable and unreliable is without foundation and contrary to the scrutiny that the data has received. To say that quarterly data is overkill is to grossly underestimate the impact that location can play in retail success. Another benefit of the quarterly data is that growth areas can be spotted immediately and a chain can become proactive in their site searches. To say that the new data methodology is only necessary in fast growing areas obviates the ability of quarterly data to show population loss from abandonment or condemnation, over building, ethnic changes and the quality of growth that can only be obtained by trending quarterly figures.

To benchmark a supermarket site against a successful A&P totally misses the mark on what will succeed in the 21st century. The consumer is rapidly changing and the real winners will create their own unique space and concepts in the retail environment. The analysis of those opportunities takes more than an ability to role play and case stores.

Kevin Anderson
Kevin Anderson
16 years ago

I think Mr. Livingston and Olsen are both right and wrong at the same time. To simply look at an high performing, perceived weaker competitor, as a primary criteria for a successful site is way oversimplifying the process, as is over stressing the need for “up to the minute” population projections as necessary for determining a good location and site. Most companies I know use multi-tiered approach to the site selection process and use some pretty sophisticated tools to get there.

We are limited by time itself on what we can do that is practical and actionable in the site selection process. Most potential programs I know of work on a per capita weekly volume estimate that is NOT seasonalized; as are most sales forecasting programs that use gravity or regression as a basis for the estimates. And more to the point, what we do as sales forecasters as to format is really driven by the type of financial analysis undertaken by the company for validation of capital expenditures. How many proformas do you see on a project that used quarterly sales to calculate returns on invested capital? Odds are not many.

Remember also, most stores are either leased or built by the supermarket company. Quarterly population reports will give you history and some forecast out to the future, but will still be based what has happened in the past. After you sign the lease and start building the store, what are you going to do if you find out from your population estimates start showing a trend downward? Odds are you are contractually obligated to finish the store and open it. So you may have the luxury of knowing that the potential is heading south, but like the Titanic, you are still going to hit the iceberg.

Are quarterly population projections useful? Yes, but more to the evaluation of existing stores and helping make the decision to close the store or to add some sort of capital to the facility if it is successful.

Don’t even get me started on the weaknesses or psycographics and what they do to good companies (hint: they take the merchandisers from doing field work and make them stay in the office – the start of the downward spiral).

Remember the more you know the stores and their characteristics, the better the sales forecast. And like Bob Gorland said, you can’t do that sitting behind a desk.

Dave Wendland
Dave Wendland
16 years ago

I agree that site selection is both an art and a science. I hope that Wegmans does not venture too far down the science route and give up the art.

I remember the early days of USA Today (although this may be an urban legend), when they chose to place their all new paper boxes near existing local newspapers. One cunning newspaper in Iowa chose to place one of their newspaper boxes in the middle of a corn field. Sure enough, USA Today placed one there also.

So…the moral of the story is not unlike “Field of Dreams,” you have to build it and hope they come.