Does this sound silly?  That’s because it is!  However, there are some folks who claim to be able to find these lost customers, for a fee, of course.  How do they do it? 

First, like any good consultant, they ask you to provide them with a list of all your customers.  Without this list they can’t help you find them.  But wait, if you have the list, are they really lost? 

The truth now comes out.  This is not about finding your lost customers.  This is about developing a PROFILE of your customers so that you can find more customers like them when you are opening new stores or selecting lists for marketing campaigns. 

A PROFILE is not the same as a list of your actual customers.  It is a classification scheme that is based on certain attributes of your customers such as age, income, race, media preferences, and shopping behavior.  Your customer list is a sample of the universe of people who have visited your stores or might in the future.

Some companies actually know who all of their customers are.  Bentley sells about 10,000 cars per year.  They know their customers.  McDonald’s sells about 4.2 million burgers per day.  They don’t know who all of their customers are, so it’s important that they develop a profile based on a sample of their customers.

 There are two primary sources of information about the universe of people who might be your customers:  census-based data and consumer files maintained by credit companies such as Experian and Equifax. 

 The census-based data starts with government information and enhances it with other sources to provide annual updates of neighborhood profiles for the US.  There are hundreds of variables including breakdowns by age, income, race, household type, occupations, and much more.  The population counts tie out to county level estimates that are frequently updated for 100% of the population.

 The consumer files are based on data captured from credit related transactions such as mortgages, car loans, and credit cards.  Although the vast majority of the population borrows for something, some people pay cash and don’t make it into the files.  Even if they are listed, the information about them may be missing or out of date, such as income (a few people have lost jobs lately), presence of children, and marital status.  However, to the extent it is correct, it’s a much better way to learn the profile of your customers than assuming that a customer is an average of their census-based neighborhood profile.

Regardless of the source of customer profile data, you need to know more than their demographics to understand if they will shop at your store.  How far are they willing to travel? Are there competitors nearby? Is the nearest store visible from the street? 

 Some people in your customer list live very close to the store, and if you assume that their demographic profile is the reason they visit you so often, you might be wrong.  They might be coming because you are so convenient!  Conversely, people with other demographic profiles in your sample might be further from your stores or closer to one of your direct competitors.  They would be happy to shop at your store if you were more convenient! 

Now your customers really are lost!  They are hard to find in the noise level created by the problem of convenience and competition, not to mention the quality of the real estate where the store is located.  How can we reduce the noise level so that we can hear the voice of the customer clearly?

The answer is that we must group them into categories of convenience AND demographics simultaneously to get a true picture.  To ignore convenience and competition is to risk a completely inaccurate customer profile, and that could lead to a store sales forecast that is also wrong.

The next time you see a map with color-coded dots representing households or “heat maps” that shows the location of your customers, be sure to ask how it was made!  Was it just based on demographics or lifestyle segmentation, or did it consider the set of choices they had when they shopped?

Let’s take a closer look at the “science” part of the equation.  The science consists of information systems that help decision-makers visualize important data about markets and sites as well as predictive models that guide the development of successful store networks. 

Although it’s great to philosophize about the blend of art and science, at the end of the day there is a cost-benefit analysis that must guide any decision to invest in software, data, and modeling services. 

There are four key factors that must be considered:

1)  The marginal impact of better science on the financial performance of stores.

For most companies, the opportunity to maximize profitability by putting the right number of stores in the right places is a compelling case for investing in more science. 

If decision support tools reduce the number of underperforming stores by a small amount, the return can be very large.  However, there are some cases where this may be less true:  sales forecasting for stores with very low capital outlay (kiosks), market planning for stores that require highly specialized real estate (zoning, parcel size), and market planning for stores that are “opportunistic” and sign short leases (e.g. Christmas decorations).  

2) Ability to define and measure critical success factors. 

Predictive models must be able to quantify the key drivers of the business.  

If success is predicated on the presence of a trade area population with certain demographic characteristics, daytime population, and competitors that can be identified and rated, a model can reliably estimate sales.  However, stores that depend on specialized venues such as tourist attractions, sporting events, or interstate highway traffic will rely more heavily on “art,” and mathematical models may not be worth the investment.

3) Consistency in historical data. 

Science is a data hungry beast. 

The better the data, the better the tools, whether they are based on visualization or mathematical models.  Most companies have pretty good data these days with the exception of customer transaction data, which can be acquired through surveys and other methods.  However, if there are few stores, or a totally new concept is being launched, the best approach might be to defer the investment in science until a critical mass of data can be developed.
  

4) Culture of senior management.

There is a VERY broad range of management cultures among chain facility operators when it comes to analytics and the perceived value of science.

 Although the trend in the past 10 years has been significantly in the direction of greater emphasis on science, it only takes a CEO who operates very fast by the “seat of the pants” to squash any attempt to capture the value of investment in visualization tools and analytics.  CFOs are usually the most supportive because of their quantitative view of the world, and if you are a visionary who sees this opportunity first in your company, you might want to take him or her out to lunch to begin building support.

Each company can be profiled on the basis of these four key factors in order to generate a “high level” ROI assessment.  If the preliminary indicators are favorable, the financial returns can be estimated more precisely and compared to the expected cost of implementing a system to capture the value.  Companies like geoVue and others can assist in this process which includes identifying such line items as staffing, software, data, and professional services.

However, the growing use of “science” in decision-making has led many to ponder the role of “art” in real estate planning and site selection.  Let’s consider the ways that chain operators apply art and science in their real estate decisions.

Science is objective and well-defined by its very nature.  Whether it’s a map of the market showing key information such as store locations and competitors or a mathematical model that predicts sales, it’s still data and equations that can be exposed at the finest level of detail for all to observe.

Art is much harder to define, but for real estate analytics we can focus on two types using the analogies of pottery and paint-by-numbers. 

Pottery art is practiced by experienced real estate field people who have learned to identify the key factors in successful sites from years of driving around retail markets, sitting in real estate committee meetings, and seeing first-hand which sites worked or failed.  It is “Pottery” art because they apply their intuitive knowledge of the  “dirt” (clay) to the observed facts (photos, site plans, and their new favorites Google Earth and Bing Maps). Most of the time they do a good job picking winners and losers. 

Analysts use Paint-by-Numbers art to estimate the potential sales of a site.  After years of generating trade area data using rings, drive times, or hand drawn boundaries, they can “feel” the patterns in the numbers and set thresholds for acceptability that are similar to keeping the paint inside the numbered areas of a coloring book picture.  The good ones can be very reliable, but they often disagree with the Pottery Artists because their approaches are so different.

The most successful chain operators combine these different types of art in their decision processes on a foundation of fact-based predictive and visualization tools to maximize their profitability.  In the coming months we will be exploring this topic in more detail to gain practical insights into the question of blending art and science in real estate planning and site selection.

We welcome your thoughts and comments on this journey!