STI: PopViews (Spring 2007)

New in STI: PopStats January 2007 Release - Puerto Rico!

STI: PopStats now includes population estimates for Puerto Rico! Several PopStats users have requested estimates for this part of the country and now they are ready - all 2,477 block groups. The basic population estimates were created using the same methodology PopStats uses in other U.S. markets. However, a unique methodology was required to calculate income and home values, since Puerto Rico does not have IRS or OFHEO data. Also, several fields of data are not available for this area, including Workplace, Landscape, expenditures, transience, ancestries, and most of the 1990 Census. Please let us know if you need access to these fields of data and we will try to fill your needs.

Understanding PopStats' Projection Boundaries

Unique Methodology Employs Monte Carlo Simulation

PopStats users occasionally ask us about our radically different approach to creating population projections, which, unlike other data providers, includes not just the projection, also includes a high and low estimate. As many users know if they have attended a PopStats User Conference, we use a simulation process called Monte Carlo to create our unique set of projections.

In fact, the strength of our process is derived from it its ability to create and upper and lower projection boundaries for each trade area based on existing growth trends. In other words, we let historical trends "bind" the simulation and influence the final outcome. In mathematical terms, this is called a "triangular distribution" technique.

Calculating the projections for each trade area begins with determining the current "shape" of its growth curve from the four classic shapes:

• Linear
• A fast-rising arch with a quick leveling off
• A slow-growing arch with a rapid upswing
• A classic "S" pattern, which is a theoretical curve called a Gompertz curve

We examine these curves from four timeframes:

• Short-term based on the recent eight quarters
• Mid-term based on population patterns since 2000
• Long-term, if growth has occurred since 1990
• The theoretical projection based on a Gompertz curve

With this information in hand, we run the Monte Carlo simulation for each block group. The result is the primary growth projection and both high and low projections. The upper boundary is a figure that basically identifies the trajectory of the "highest" curve in the simulation. The lower boundary identifies the lowest curve in the simulation.

Essentially, if the values of the high and low boundaries are close together, the model is indicating a high degree of confidence in the projection. This is typically of a block group that is experience slow growth rates. If the high and low boundaries are far apart, the model is less confident in its population growth projections. This is common in high-growth markets.

Understanding how projections are created helps users determine how they can most effectively be used in market research. One of the most valuable ways to apply projections is to use them as a "confidence" factor in determining trade area stability. So when there is high confidence in the projections, a company may need to conduct less additional research. But when the confidence is low, additional research may be required to understand the market's future potential.

The Problem with Business Listing Services

We have been asked: Why not just use a business listing service to calculate the number of businesses in a trade area for STI: WorkPlace? The reason is accuracy. And we have the facts to prove how skewed the average business listing service is. One of the most popular services reports that there are 12.5 million businesses in the U.S. However, three government agencies cite a very different number:

• The U.S. Bureau of Labor Statistics - 7.5 million businesses
• The U.S. Census Survey - 7.5 million businesses
• The USPS - 7.5 million businesses

As always, PopStats seeks out the most accurate - not necessarily the most accessible - information for all of our products.

PopStats Offers Five Years of History

PopStats users know that every release of the product includes eight quarters of population history. However, everyone doesn't realize that the product also offers five years of history - starting from April 2001. Several users had asked for this data and we are proud to report that it is now available in the product.

Announcements