SDSS for Location Planning, or The Seat of the Pants is Out

by Larry Daniel

Manager of GIS Development

MPSI, Inc.

(Published in GeoInfo Systems December 1992)

In the GeoInfo Systems May '92 issue, Donald Cooke described the general characteristics of a spatial decision support system (SDSS). This article reviews those concepts and extends the discussion by examining SDSS as it has historically applied to site analysis.

Reviewing SDSS

As described by Donald Cooke in the May '92 issue of GeoInfo Systems, a spatial decision support system (SDSS) is 'canned software that is intuitively obvious to use, solves problems efficiently and delivers immediate results'. SDSS is off-the-shelf software for carefully selected functions, i.e. bug-free point-and-shoot capabilities on very specific spatially-oriented needs. SDSSs don't require in-depth commands to operate, yet allow users to negotiate very sophisticated geographic analysis.

Many of the distinctions between GIS and SDSS are derived from the differences in their target audiences. While GIS sites often employ technical analysts to plow through the nitty-gritty of setting up databases and defining complex functions, SDSS sites generally have but one or two business analysts who want to make spatial decisions but don't want to become immersed in the intricacies of GIS. The SDSS users don't seek a smorgasbord of commands, they want a select handful of intuitive operations to yield quick and effective answers. They want to avoid technically complex operations and while they're impressed by GIS' potential, what they really want is three or four buttons and an answer.

The SDSS attitude is particularly prevalent among those users who consider GIS site analysis. To understand why, it's helpful to review the recent evolution of this application ...

From the Seat-of-the-Pants to Science

For many years in many circles, site analysis had been a black art. Resembling seat-of-the-pants operations more than the computer-aided science it is today, most organizations simply drove the market, aided with whatever data was available (which usually wasn't much). It wouldn't be unusual for a team of top decision- makers to hop on a bus and get a 'guided tour' through the streets that their decision would affect. During that tour, critical decisions were made: where to place new franchises, which existing sites to further invest in and which to close. It was hardly a well-honed, systematic approach to crucial decisions. Although articles had surfaced in the early '60s on how to systematize site analysis (e.g. David Huff "Probabilistic Analysis of Consumer Spatial Behavior", William Applebaum and Richard Spears "Methods for Determining Store Trade Areas, Market Penetration and Potential Sales") the cost of gathering data, programming and running meaningful analyses prohibited most organizations from changing their ways. As a result, the bus rides and seat-of-the-pants decisions stayed intact.

During that time, probably the best way to gain a competitive edge in site analysis was to contract with an organization to provide a suite of geographic software services. Vendors like MPSI, of Tulsa, offered clients the knowledge and skill to build custom databases and models, as well as the software for slick assess to the data. Though much of that type of software was not graphically based, it was still in many ways an early form of SDSS -- custom developed for retail site analysis, the systems were relatively friendly, quick and had the ability to derive application-oriented relationships between retail sites, street networks, and demographics. Unfortunately, these software products were quite labor-intensive to develop, and costs associated with constructing the database and formulate the models forced prices beyond many organizations' reach.

Over recent years, however, the costs associated with automating site analysis have tumbled and the technology has become more feasible for all organizations to consider. With the emergence of GIS and cheaper processing power, even the humbler retail chains could potentially automate site analyses decisions. With MapInfo or Atlas on a 386 and data from any number of sources, effective answers on siting new franchises or direct mail strategy could be but a few dozen well-planned keystrokes away. Browse through the May '92 issue of GIS World and it's difficult to deny -- with GIS, business functions like site analysis are getting overhauled. Decisions aided with the widest variety of spatial functionality and data are in, the seat-of-the-pants is out.

But before running out to purchase that copy of MapInfo or Atlas, organizations must come to grips that GIS is not a panacea. Although the technology has arrived, new business and implementation issues have replaced the technical issues that just cleared. How many newcomers to GIS are really prepared to tackle these questions (or the many others that haven't even been listed):

Different organizations will respond to these issues in different ways. If they're prepared to hire a consultant or a new GIS analyst, they might want to trot down the path with the commercial GISs. Others may simply pass the choices (and confusion) along to their IS division. Still others will decide to pass on GIS altogether. But those that remain will probably opt for an SDSS. For although the up front costs of an SDSS might outdistance other options, many organizations will find the SDSS route appealing. Occasionally that choice will reflect the sheer technical appeal of individual decision support systems. Most often, however, that choice will reflect a consideration of organizational issues and users will choose SDSSs because the choice enables them to focus on their applications and ignore the plethora of technical issues inherent to constructing the system.

What will distinguish a site analysis SDSS?

A site analysis SDSS should conform with just about everything that's been said in this article. In general, an SDSS will have::

In addition, the site analysis SDSS will include some application-specific capabilities. The following features are envisioned as standard in any site analysis SDSS:

Remember that the emphasis on an SDSS is not necessarily to provide the user with the flexibility to address all business issues, but the ability to provide the optimal software set for the specific application at hand. As such, another hallmark may well be the absence of functions available in related commercial products, but inconsequential for that application at hand. For example, while an option to identify the lat/long at the mouse crosshairs might be expected in many GIS packages, it probably won't be present in the site analysis SDSS, since it has little bearing on the application and would unnecessarily congest the system's menus.

Note that while the GIS market is essentially horizontal, SDSSs appeal to verticals, i.e. they are custom-made for a specific industry. SDSS concepts extend to data -- SDSSs are unlikely to come furnished with every piece of commonly available data, but quite likely to reveal more valuable propriety information The vendors will often provide highly specialized databases. MPSI, for example, can arrange for clients to receive very extensive petroleum and financial services data. Users could acquire pricing for every gas station in the country or the locations of every car wash ... Thus, while a petroleum site analysis database might not be stocked with simple information on hydrologic features, it will be fully equipped on the gas station data most pertinent to its usage.

Conclusion

I sense that SDSSs will be quite important to GIS -- they'll be the vehicle through which those out-of-the-know and not-wanting-to-know will come to value and appreciate GIS. Without SDSSs, GIS could very well be constrained to an educated, technically able community. With SDSSs, GIS concepts can reach even those who felt comfortable with the black arts and seat-of-the-pants. SDSSs will appeal to the Einsteins who focus each day on tackling very specific business problems. It was said that Einstein couldn't remember his own phone number because he didn't bother himself with superfluous data. In a way, SDSS users will be the same -- they won't have to concern themselves with what they consider superfluous -- they can instead focus their energy on working more feverishly to conquer very specific responsibilities.