Transparency and the Location Mystique

Transparency In Retail Site Location Models
Transparent: Free from pretense or deceit, frank.
Obfuscate: Mask, conceal, disguise as part of a hidden agenda.
Location Decisions: Science or Intuition?

When it comes to selecting retail locations, knowledge is power. Experience plus intuition are the key drivers. The large majority of location decisions continue to be made on the basis of intuition and experience, not science. Why? Well, a large part of the reason is that this common sense decision process works pretty well in many situations. Good locations are good locations after all. Mom, my neighbor George, or even Aunt Sally can all tell the difference between good and bad locations — to a point. The next step, however, that goes beyond good and bad to actually estimate a store’s performance in that location, does require special intuition and experience. This also works quite well — to a point.

The best decisions come when real estate expertise is complemented by objective, scientific knowledge that can come from demographics or other data — and from predictive models. Because so many companies use models, or other kinds of intelligence, today to evaluate sites, it’s important to take our understanding one step further. Both intuitive predictions and model predictions tend to be used in ways that are not transparent. And, if a decision process is not transparent, it’s also not generally open to shared feedback and improvement. If you are on the receiving end of model predictions delivered as either mystery statistics or the oracle, as real estate or development personnel often are; then a lack of knowledge about how the model works may be beneficial. Usually, the normal “good” locations are not the ones people remember; it’s the dog sites.

In the long run, however, everyone (especially your company) benefits when the basis for a decision is objective, understood and shared. This enables the criteria for current good locations to be extended into future “good locations”, and the problems with bad locations to be understood and avoided. Companies that have this attitude usually benefit enormously from predictive modeling because the model results are simply one piece of the decision pie, and they don’t step on anyone’s toes. If everything is on the table and above board — transparent — then new information is always welcome.
Unfortunately, the more common problem is that a history of relying on the company’s version of a “location mystique” (which just means your company’s shared, intuitive understanding of what makes a good or bad location), is combined with a general lack of transparency and understanding around the factors that drive store sales. This results in a decision process that attempts to be good, open, and feedback-oriented, but ends up driven by tradition and opinion.

Why Location Decisions Often Lack Transparency

Real estate information and decision-making, certainly on the commercial side, lacks transparency because the key players benefit from controlling access to information at different points in the process. We see, for example, that commercial databases listing available sites are very incomplete because the local brokers have a wealth of inside knowledge. They know which property will be available soon, or could be available at the right price — and their competitive advantage comes from controlling this knowledge.

Sales Prediction Models Typically Do Not Please Everyone

It won’t surprise anyone reading this to know that most real estate decisions are driven by experience and intuition, not models or objective criteria; or to learn that the decision-making process for site selection often loses transparency because of internal politics, communication barriers that prevent knowledge sharing, hierarchical administrative structures, and IT or technological issues.

The truth is that, much of the time, brokers and real estate VPs don’t want the interference of a model — unless it supports their decision. Why? Because models generally get in the way of doing deals. In cases where the model predictions and expert opinions converge, great! The model is not a problem, per se, but actually it’s agreement with the expert is setting up a big problem because now, anyone examining the results is likely to conclude that the model has some validity. Right? Well, look at what happens when a site comes up that the VP likes — but the model is not enthusiastic about. Ouch. The VP wouldn’t have proposed the site if he/she didn’t think it would work. Now the model’s second opinion really hurts because the last site got high ratings and everyone was happy. Who is right in this case: the model or the VP?

Transparent Site Features Help All Parties Communicate Objectively

The problem here is transparency. Store sales are obviously based on a combination of factors. Some are location-related (like the site attributes or demographics), and others are not, like operations, marketing or even brand strength. If the VP and the model were talking the same language, they could agree or disagree on the strength of each of these factors, and eventually reach a consensus opinion on the site. Lacking this transparency in the decision process, both are left only with their differing opinions, and supporting justification. This matters little, since they are not speaking the same language.

We addressed this issue in our modeling systems by listing specific site features, or rating the quality of certain demographic measures. This does help some. What stops it from being very beneficial, is that we didn’t take the issue of transparency far enough. The problem was confusion about the role of the location in store sales. This varies from one extreme, “If it’s a dog site then the location is at fault — bring me the head of the real estate manager who picked this site:”, to the other, “The location may be part of the problem but lets take a good look at marketing, operations and competitive positioning first.” In general, the way this confusion appears, is that because the influence of all key factors on store performance is not measured quantitatively, or well understood, the role of the location is typically perceived to be much too strong.

In my experiences over twenty-five years, with several hundred retail companies, I would have to admit that the number of companies willing to make the location decision process transparent could be counted on just a few fingers. Dick Riveria, the former president of TGI Friday’s did a rebuild of their sales prediction model — just to open up the process and the information to a new team of real estate managers and executives. Ron Stegall, the founder of BizMart, insisted that everyone, both core staff and brokers, be familiar with their site model’s components — so they could validate them. Gary Wyatt at Lowe’s introduced modeling to bring transparency to a process shared by both marketing and real estate. Jim Kirkpatrick at Applebee’s, and Jim Torcivia at Cracker Barrel, always tackle problem stores from a “bring me better measures so I can understand what is happening” perspective, rather than simply as location problems.

As a forecaster, my biggest challenges in making store performance transparent were situations where the concept was part of the problem. In these situations, the concept name, the differences between store prototypes across markets, the competitive positioning of the concept, and so forth, — often impacted sales. Yet, raising the concept flag was not something the owners were willing to handle. I won’t share any bad examples here, but Ruby’s Diners, Staples, and Red Robin are good examples because their hard work in the other direction — to fine-tune the concept component of performance in objective, shared ways that impact the bottom line.

One of the major ironies of location modeling is that it is common for sales prediction models to be the ultimate in non-transparency. They are often statistically complex (which is intimidating by itself for many people), difficult and time consuming to learn or use, access is often controlled and limited to a special group of analysts, — and who gets to see what reports when may be determined by political agenda rather than “need to know.” You would be shocked at the major corporations for whom I’ve conducted workshops that have never had all the key people involved in location decisions — real estate, finance, marketing, operations and executives — in the same room together to discuss what is needed when from whom.

Transparency Helps to Create Actionable Results

Another problem with most models is that they are either largely based on statistical “mumbo jumbo” that could not be made transparent if you wanted (some of the early neural network models fit this description well) — or are so simplistic (like common regression or gravity-based models) that they lack credibility because even a statistically naïve person can understand that they cannot explain the complexity of many stores in many retail situations in many markets. Even worse, what good is transparency if you cannot do anything with it? We’ve built models with fairly transparent reports for some time — but so what! The user can disagree with what they see, but the battle is already lost because the results don’t change! If I disagree with the answer, and I can see why the answer is wrong — because the system is transparent — you had better let me be able to act on it.

Now that you know my biases, you know what’s coming next. There is absolutely no good reason for Site Selection and Sales Prediction to be a covert, mystical process with hidden rules, hidden agendas, and controlled access to information. One of the biggest misses I’ve seen has to do with franchisees. I would guess that at least 25,000 franchisee locations have been selected using our modeling systems and reports. Most of these franchisees saw a Site Quality Rating, the factors that contributed to the rating, and a sales prediction; yet, not a single one has (to my knowledge) ever been able to sue the franchiser successfully over a poor performing store based on information in our report, largely because it is transparent. The courts have made it clear for many years: not sharing critical information is generally risky. Sharing (being transparent) with an appropriate disclaimer offers more benefits to all parties, and more protection from liability.

Benefits of a Scientific Approach to Site Selection

There actually is a science of site selection with a good deal of research on what factors matter where. My book, The Site Book, is a synthesis of much of this knowledge as we apply it in our modeling programs. Companies that have made the culture shift to use and share these objective rules for location decisions — not just the financial estimates — have seen tremendous benefit on many dimensions, not the least of which is their bottom line. One of the simplest statistics that illustrates the value of this scientific approach is sales volume. When we looked back over the last 7,000 locations for 13 different concepts that were evaluated using the logic in The Site Book, we found that locations that had a Site Quality rating of 65 or higher (the average rating is 50 on a 1-100 scale) also had 17% higher sales than the average store for that concept! Yes, there is a strong relationship between an objective measure of site quality and sales.

PAI’S Approach to Modeling

The rules used to make decisions in models need to match the rules of the retail world, not just fit a set of mathematical criteria. This statement, which is really talking about transparency, simply means the models need to be thinking about the world in ways that match the logic that you or I might use as real estate professionals. To some degree we met this goal in the past by creating reports that listed the criteria used to estimate site quality and sales — but I don’t think we went far enough. Why? Because the best, transparent models are also:

Logical: The rules in the model match what happens in reality
Open: You can see the rules and logic operate
Robust: The model can predict reliably despite the “noise” or error associated with most sales modeling
Adaptive: There is a way to learn from the errors in prediction to improve the model

If you had asked me five years ago, when I wrote The Site Book, I would have touted our logic for location analysis as pretty complete. And, I might have said the same about sales forecasting. Yet, how can this be when locations are responsible for only approximately 50% of store sales! What about the missing 50%? For many years in our presentations to clients, we represented this 50% with the following pie chart:

This is transparent to a degree if you know the Location Factors and can specify their influence. Yet, you cannot have a truly transparent sales prediction without all of the components. It’s impossible to know if a store’s weak performance (when you think it should be doing well) is due to a problem with your location models, or store operations, marketing, etc. This shift in thinking led to a series of analyses to explain the other factors: the market, competitive positioning in the market, operations, the concept, the brand and many others. Could we explain the sales contribution of these other components? Typically the answer was “yes” because we had so much detailed customer data that let us get at problems with operations or marketing — and in doing theses analyses, we also learned that once the non-location components could be predicted, it was also possible to predict the contributions of many small location factors in our original model. How much is a new sign worth? What about being on the corner versus down the street two blocks? What is the contribution of the tourist population to sales? What’s the contribution of Market A versus Market B to sales?

Answering these questions led to a very detailed Transparent Report in which all of the factors underlying store sales could be viewed and analyzed. The high-level version of the report is illustrated in the table that follows:

Making Changes to Model Predictions… Adding Validity or Fudging

Logical, open, robust, adaptive… that’s a great list of attributes to define transparency; but they don’t mean much without the ability to make changes in the model’s predictions to reflect what you’ve learned. Model predictions, even from the best systems like PAI’s, the National Weather Service, or election polls, are not perfect. You should expect that they will change when knowledge improves. Remember that transparency in almost any context — interpersonal, financial, business management, politics, and so forth — usually goes hand in hand with getting good feedback and being flexible enough to act on it. Part of the value in watching the individual gears turning in the “big machine” is to be able to notice the wobble and make a correction when needed. This couldn’t be truer in many domains than it is for location analysis — where the change in a single factor (a new manager, construction on the road in front of the store, or a competitor opening down the street), can dramatically impact performance in one week!

FOUL you say. A fudge is still a fudge. We wanted a model to provide an objective, unbiased prediction, not one that could be fudged to fit someone’s biases.

Let me see if I understand what you are saying. When you are ill, would you rather rely solely on the results of blood tests and temperature readings than to add the interpretation that is part of the physician’s expert opinion? Or, if surgery is called for, to would you forego the expert opinions of several physicians? In the case of location analysis there is a lot of knowledge about a site, the market, competition, and experiences in similar markets that may not be in your modeling program. Besides, transparent is as transparent does. In other words, because it’s transparent, everyone can see the gears turning. There can be no mathematical cover-up, as you might experience with a single prediction driven by complex statistics. Transparency means you are looking at the facts, at least as the model sees the world. There can be no hiding because everyone else sees and shares the same view. If there is disagreement about a “fact” such as the quality of the market, name recognition, the manager’s performance, visibility of the store, or any other parts of the transparent report — that’s actually good! You are working at a concrete level where a consensus can probably be reached. If the consensus opinion is that the model is wrong about one of these components, you want to change it and determine how this change will impact the predictions.

In our work with clients, disagreement with the model results, and the process to understand these disagreements (that we call “field validation”), is encouraged. We see the whole proposition as 50/50. Fifty percent belongs to the objective model results; fifty percent belongs to the intuition and experience of an expert in the field actually evaluating the site, or as a part of a real estate committee in a Board room.

Historically, adjusting the models predictions meant adjusting or tinkering with the input parameters until you got the predictions you wanted. Today, what we attempting to do with our transparent approach to modeling, is encourage shared understanding and communication about all of the components of sales, especially the large, non-location components such as operations or marketing, that are not well understood in many companies. How much do operations or marketing contribute to store sales? Quite a lot — and when you can measure this contribution, guess what happens to your location model. It gets a lot better because you are not trying to predict the performance of operations with a demographics report. You are predicting only that part of sales explained by demographics with demographic predictors, site sales with site features, market sales with market features, and so forth.
We’ve witnessed that transparent models completely change the nature of the conversation when there is a disagreement between a model’s prediction and actual sales. Instead of a conversation that begins, “What’s wrong with the model. It’s under-predicting performance for this store by 25%;” It goes, “Here’s a large discrepancy. Let’s look at the transparent model report and see if we can understand where it’s coming from!” What you want from a model is “it’s” version of the truth, not an oracle. The key to understanding and utilizing this truth is transparency. Transparency transforms what traditionally was a black box process into an intelligent dashboard, with gauges that explain what is happening in each of your stores, and dials to make the needed adjustments. Now you have a wise partner in your modeling system, not an idiot savant!

By Dr. Richard Fenker, PhD
For the last twenty years I (and more recently with two colleagues, Lynn Cherry and Selby Evans) have been tracking an elusive phenomenon, which at times is as plain as the nose on your face, and at others as elusive as a wisp of clouds on a clear day. It’s a force in the world of retail that is as obvious and strong as the force of gravity when you toss a ball in the air. Yet, it is also as mysterious as the physical principles that underlie gravity and it’s relationship to other forces – which are still not well understood by physicists. The phenomenon is retail synergy. As obvious as the value of a collection of synergistic retailers sharing a common shopping center may seem, especially their influence on the “gravity” of the center or its ability to attract customers, relatively little beyond common sense is known about this practical yet ephemeral concept.
In our everyday experience, we “live” the concept of synergy as we are attracted to clusters of retailers that work well together. In fact, we even name these clusters with familiar labels such as “shopping centers, malls, strip centers, power centers, lifestyle centers” and the like. There is no mystery here, other than a wait for the next clever name to come from a developer in California, Minnesota, or New York. There is also no mystery about the fact that some of these centers “work” (meaning they are more effective draws for retail traffic) much better than others, or that certain types of people are attracted to certain types of centers. There is also an obvious parallel to atomic physics. We study the behavior of atoms and molecules, and the attractions between heavier particles such as a proton and the lighter particles such as electrons, or the special universe of quantum entities such as charm, flavor, spin and quarks — more on this in a moment.

The mystery for me started in the early 1980’s. I was working with Norm Brinker and Ron McDougall on a customer research project for Chili’s. In developing the research instrument, we asked the question: “What other retailers are linked to a visit to Chili’s — either before or after?” Our intention in asking this question is obvious. If there are supporting “linkages”, or other retailers likely to be visited in connection with my restaurant visit, why not locate a Chili’s near centers containing these retailers, thus increasing the probability of a visit to Chili’s? The first answers that emerged from our research made sense. While we couldn’t identify specific retailers that mattered, we could say that being near “upscale shopping” or “entertainment activity” mattered a great deal for some locations. So, at the most general level, certain classes of retail activity were clearly synergistic with casual theme dining. I can sense that yawn beginning… hold on for a moment.

What we didn’t understand at the time were any specifics. What types of retailers were best — or did we need to get to the next level and deal with specific retailers? We could count the stores in a shopping center (although the databases at this time were much more limited than today), but we didn’t understand the relationship between the stores or how much this mattered. Finally, there is obviously some relationship, or interaction, between the people who live and work in the neighborhood, and the retailers. However, it was not obvious how this interaction contributed to the “synergy” of the area. Yes, we were scientifically curious, but were driven by a much more practical problem. Our forecasting model was telling some of these early clients (such as Steak and Ale, TGI Friday’s, or BizMart) that they would do well in certain centers because of the “synergistic” activity — and we were dead wrong. I remember driving a site in Washington DC with a retailer that had used our model, in part, to make the decision to open the store. The retail energy was awesome. With a regional mall nearby, it was also a great neighborhood – filled with our client’s customers. What was the problem?

The answer wasn’t obvious. However, as we studied the problem, one area where our thinking was muddled did become more clear. We had been mixing the idea of Retail Draw and Retail Synergy, essentially treating these as the same concept in our model. The more good retail around YOUR concept the better. Unfortunately the world doesn’t work this way. As we studied the DC mall carefully, it was obvious that our client was “near” a hundred supporting retailers in the mall or in other shopping centers near the mall. However, they were actually “adjacent to” a smaller center filled with junk retail, and badly positioned with respect to the competition. I’m sure you’ve heard this story before — “it was a great location, except for the…”

At least the concept of Retail Draw seemed clear. And based on the results of a couple of million customer surveys that we’ve administered since, and over 100,000 sites evaluated, it has remained so. Retail Draw is essentially a measure of the “pull” of a retail area based largely on the number of businesses in that area. Regional malls have huge pulling power, while local strip centers have very little, with power centers and lifestyle centers in between. The importance of “pull”, or “retail gravity” as it is called in many modeling approaches, is that the following principle seems to be consistently valid:

The larger the number of retailers in an area, the stronger the Retail Draw, and the larger the effective “trade area” for most of the retail businesses.
There is little mystery here. The city center in healthy cities is the ultimate retail area, drawing from the entire city. Regional malls can have as much or more pull. Everyone in the community, to a degree, becomes a customer of the businesses in these areas. What was still a mystery, however, was the synergy of the area, or the synergy of the businesses in the malls or shopping centers of the area. A definition of retail synergy was emerging in our thinking:
Retail Synergy describes the degree of compatibility between a collection of retailers such that customers who use one of these retailers are also likely to use other retailers in the same center.

This is not a bad definition of synergy as we would view it today, but there was still a major flaw in our thinking that took several more years to fully appreciate. Can you see what we were missing? In any case, armed with this definition of synergy, we could begin to understand that the best centers were ones that attracted distant customers, not only because of Retail Draw, but because the stores in the center had some degree of Retail Synergy. Looking back, this seems like an obvious step, and in some respects this is true. However, as our understanding evolved in real-time twenty years ago, a philosophy was also evolving, and it is one that has influenced our thinking about location analysis and sales forecasting models, from the 1980’s to the present. What was happening was that in the pursuit of answers for why the “statistics” in our models didn’t work in some cases (this is a euphemism for “bad predictions”), we were forced to look more carefully at the behavior of the customers we were trying to model. This shifted us from a “find a better methododological” approach, to one of explaining how shoppers and diners actually use the retail world.

One of our first chances to test this thinking came with the modeling research for two clients, Eckerd’s and Cracker Barrel. As we designed their customer surveys, we added sections that gave us much more detail on the behavior patterns of users, and the specific combinations of adjacent retail businesses that helped or hurt sales. It may not surprise you to learn that a visit to Eckerd’s has a high probability of being linked to a visit to other kinds of retailers or institutions; or, that travelers, service stations, motels, and certain retailers all interact in certain ways to influence Cracker Barrel’s performance. At the time, it surprised and delighted us — and even though we didn’t see a large jump in the sales forecasting accuracy of our models, our “risk models” (models designed to spot potential dog locations) did get a nice bump in accuracy. We could now identify certain kinds of retail situations that just were not “right” for our clients. This synergy component has remained a part of our models ever since.

The Cracker Barrel and Eckerd’s research was moving us in the right direction, and several mall modeling projects finally got us to yet another plateau. What every developer and mall-based retailer who is reading this article understands, that we eventually figured out, was how important adjacency influences are in driving behavior. Shoppers have patterns, driven to a large degree by the occasion. On a practical Saturday, I may bounce from the grocery store, to the hardware store, to the drug store, to the post office — or whatever. On a shopping Saturday, I visit the mall, going to six stores that sell my kind of clothing, have lunch, and then look for a wedding present. Which stores I visit in any time period is obviously influenced by my needs and the time factors; but both synergistic factors (which stores are in the retail area) and adjacency factors (which stores are near each other in the centers) also help determine which specific retailers I will visit.

Retail Adjacency Effects describe the local spatial relationships between key retailers and any nearby concept. The strongest adjacency influences occur when the convenience of having certain retailers nearby increases the probability of a “linked” visit to your store.

New car dealers figured out a decade ago that the best way to bring shoppers into their showrooms was to create a cluster of dealerships, because, when people shop for new cars, they typically visit a number of dealers on a single trip. Today, unless you are a Lexus, Mercedes or other “destination” dealer, you are asking for trouble if you ignore the presence of this common behavior pattern tied to Retail Adjacency.

Now, we are getting a little warmer. As we incorporated adjacency influences into our models, predictions did improve. In fact, the combination of adjacency and synergy could make as much as a 30% influence on the bottom line for some types of clients. If you are a fast food concept, for example, how do you feel about locating in grocery-anchored shopping centers? Our research findings here might surprise you. (Send me $1,000, and a box-top from my favorite brand of cereal, and I’ll share them with you.) What we had done was to get down to some very specific behaviors of consumers that were closely related to the way the retail world was arranged.

Retail Draw, Retail Synergy, and Retail Adjacency — what is next?
The next step is in some respects a digression, since our discussion is about retail synergy; but, it happened so naturally and with little contribution on our part, other than to say, “yes, we can do it,” that it is worth mentioning. Two major big-box retailers in different industries, one grocery-related and one merchandise-related, asked essentially the same question at the same time. “Can you help us design the layout of the store to increase the potential that customers who come to purchase one product have easy access to related products they might also purchase?” The answer was, of course, “yes,” and within a couple of months, our list of synergy-related products had been expanded to include the concept of “Product Adjacency.”

Product Adjacency, or Department Adjacency, describes the layout of store departments or merchandise in order to optimize synergistic patterns of purchasing behavior.

To meet the needs of these retailers, we modified our standard customer research instrument to include in-store behavior patterns, so we could track what customers actually did while in the store – and surprised ourselves with the results. The same strong proximity-related patterns we were observing for Retail Adjacency outside the store continued in the store. A few departments (the “destination” departments), for example, attracted most of the initial visits to the store. Other departments visited were strongly influenced by their proximity to these destinations as were purchasing behaviors. In one case, a radically new store design was created by clustering secondary departments, that appealed to certain customer segments, around the destination departments that attracted those same customers. This principal, while undoubtedly not novel, proved so effective that we have adopted it as one of our classic approaches to store design.

Despite all of this work, there was still a missing piece in our definition of synergy. We could see from our customer research studies that some busy retail areas and malls seemed to attract customers from all over the market, while others of a similar size behaved almost like local malls or shopping centers, even when many of the same brands were present in both centers. Ridgemar Mall and Hulen Mall are two of the major shopping centers in the Fort Worth area of Texas. The difference between them is profound. Despite the presence of a Neiman Marcus in Ridgemar Mall, plus essentially the same retailers as the Hulen Mall, it behaves much like a local mall, drawing largely from a five-mile area, with many of the surrounding stores and centers struggling with marginal performance. Hulen, on the other hand, is a major destination mall that draws from the entire city with its mix of mid-scale and up-scale retailers and restaurants. What’s wrong with this picture?

You already know the answer, don’t you? You can see what obvious component is missing in our understanding for synergy —- demand, destination-driven demand. Looking back, it is obvious, but it was not obvious at the time. The retail world is filled with centers of retail activity, each with some level of Retail Draw, ranging from a dozen stores to many hundreds. These centers can become strong attractors of shoppers and diners for the reasons explained above, primarily Retail Draw and Retail Synergy. However, the strength of the Retail Draw depends not just on the number of stores, and the tendency for people to link shopping visits to several different stores on a trip, but also on the “lifestyle focus” of the center. Secondarily (because customers will cross neighborhood boundaries if the draw is strong enough), it depends on the “lifestyle” fit of the center to the surrounding neighborhood.

In other words, large clusters of retailers sharing a common focus on a certain set of customer segments or lifestyle groups, have by far the strongest drawing power. Not everyone in the market will visit these centers, but the core lifestyle groups will travel quite a distance, and deal with other hardships associated with traffic or locations, because the draw is so strong. Have you ever been to the IKEA location outside of New York City? Well, you cross the Hudson River into New Jersey, drive North on the freeway until you are in the middle of the warehouse, factory, and wasteland zone. Next, you proceed East a couple of miles into no man’s land along the river bottom, and arrive at one of the most successful destination retailers in the world – capable of creating their own draw and synergy because the lifestyle pull is so strong for some segments.
Lifestyle Synergy describes the focus of the retail area, or shopping center, on a limited set of customer segments or lifestyle groups. The larger the retail mix, and the stronger the focus, the more Retail Draw for the targeted groups.
Now, we are almost finished with the components needed to build a good model of synergy. They include Retail Draw, Retail Synergy, Retail Adjacency, Lifestyle Synergy and last but not least, the “mystique” that accompanies any successful business venture. By mystique I mean that for the best of everything, the whole is always greater than the sum of the parts — wines, personalities, art, sex, and certainly the most successful centers or retail businesses. You cannot explain IKEA’s remarkable performance without mystique!

The concept of synergy speaks at the classical level to clusters of similar or related retailers. In reality, it too is a much richer concept, linked inevitably with the properties of the retail world that are one step beyond “you get what you see,” and more closely linked to an “entangled,” interdependent, universe where a healthy respect for the mystique — of a Krispy Kreme, a McDonalds, an IKEA, or a Lowe’s is appreciated – even if not completely understood!

1. Who are your customers? Knowing “Who are your customers” can guide planning, marketing, and site selection if your research asks the right questions. Bad answers here serve as little more than an afterthought in an annual report.
2. How do they use your concept? Usage patterns feed operations the information needed to improve perceived service by meeting the needs of each user type; this knowledge also helps real estate understand the site features most important to each group.
3. How far, and for how long, will they normally travel? A concept normally has three trade areas that matter, not one. Time and distance data for users coming from work, and for users coming from shopping or other retail activity, is as important as it is for people coming from home.
4. What drives their visit, and how well are you executing on these attributes? For improving operations and encouraging return visits, there is no comparison to directly matching customer expectations on key attributes with satisfaction ratings. Higher satisfaction ratings mean higher sales. If you want to know how to boost these ratings, just ask the right questions!
5. How are you positioned relative to your competitors? Knowledge of your competitive positioning goes hand in hand with location planning, marketing, and new market development. You are handicapped if you know your competitors, but don’t know which user groups they impact or why.