George Moon, CTO of MapInfo, was heard to utter a provocative thesis last May at the Location Technology & Business Intelligence Symposium, at least for the GIS-oriented crowd.Moon's simple statement was that "the map is not important."
This really should not be as controversial as it sounds. Yes, some people make maps the way others make music, as a means of artistic expression. But in a business context, a map is a vehicle for making decisions and taking action. The decision may be "map-related, but all [the decision maker] really cares about is what's the number."
Obviously, many disagreed with Moon and rushed forward in support of the map itself. Jim Stone, president of geoVue, asked, "how does one integrate human experience within the models?" The map is one of the critical links. Brady Foust of Proxix emphasized the use of maps to communicate the decision-making process and the decision itself: "If people don't know how you got there, the level of buy-in is going to be less." Adding a pragmatic voice, Michael Cosentino of Sun Microsystems told us "the map leverages information you already have: use it!"
In response to this onslaught, Moon felt compelled to say, "I didn't mean to say maps are not important." Well, my notes have him saying that just a few minutes earlier, but we know what he meant: the maps are important, but let's focus on what they are leading up to, which is the business decision.
Daily Use of GIS in Business
The successful uses of GIS and location based technologies support business decision-making in compelling ways. Bryan Fields of Millennium Retail Partners (and formerly of Home Depot) opened the second day of the conference with a case study. His focus on GIS as a decision-making tool is particularly striking. Under Fields' direction, Home Depot used GIS tools "to identify the best opportunities to deploy limited capital."
"The tool is far more important than the underlying technology," Fields stated. For it to be effective, this tool has to be intuitive (people can't afford to spend a long time figuring it out). "Optimal tools provide rich information in an intuitive manner." The tool also has to be "actionable." I take this to mean it should provide information that is specific and useful enough to let the business make decisions.
Fields offered some canny remarks about human behavior. One of his criteria for a useful technology, for instance, is that it truly becomes a part of how the business works. As he put it, "When nobody's watching [our employees], they still use their maps." That's success.
There is not enough space here to relate all the case studies presented, but two more stood out as classic applications. Kevin Quinn of Information Builders used maps to compare where a person lived with where they purchased items in order to detect food stamp fraud. [Figure?] He concluded that it was the "combination of mapping and business intelligence views" that led to his success. Michael Cosentino (Sun Microsystems) described how in 1975 (ancient history!) the USGS/BLM/BIA found 72 archeological sites within a 10,000 square mile region. [Figure?] Starting with a small number of known sites, he developed a model relating location to proximity to water, local high elevation, and so on, then conducted an overlay analysis to identify 84 likely locations of hidden sites. That's a pretty good rate of successful predictions.
Bill Hou of Siebel Systems characterized these efforts succinctly as optimization and then added, "What makes the app good is how smart" it is. Rob Stephens of SAS went even further, saying a phrase that had been in the back of my mind both days: "hypothesis testing." That's what Quinn and Cosentino and just about everybody else was doing: developing and testing hypotheses. This is no less than scientific method and statistical optimization, applied to business problems.
The Simple App
That's a significant problem, because science and statistics, modeling and optimization - "being smart" - are just not accessible to most people in most organizations. Perhaps that is what prompted Henry Morris (IDC) to inquire, "The map becomes an interactive interface into a predictive model. How do we bring this to more users?" Larry Daniel (Conclusive Strategies) and Brian Heber (Nationwide Insurance) were more blunt about it: "the available tools are too complex for many."
Several panelists independently suggested a path out of this dilemma. George Moon highlighted the interface. "The secret is making it so easy to use, [so people] can get the answers and think it's simple." Steve Walden (Group 1 Software) agreed, stating that most applications will be simple. "Our customers don't know when GIS is done to them:" a decision just pops out.
I think that again the map is going to be key. A well-made map can, as Morris said, be the user interface for a model that can be as sophisticated as needed-under the hood-to develop a good decision. It is also the mechanism to convey the model's results in a form that the decision maker can use.
In addition to needing good data and software, Brady Foust (Proxix) often adds that you need good "wetware:" the brain matter for using it well.
Jack Dangermond (ESRI) commented, "there is a lack of an academic foundation that does research; this is needed to help business grow....We need business graduates who understand ...the spatial dimension. Researchers."
Apropos this need, Susan Wachter, professor of Real Estate at The Wharton School, one of the Symposium organizers, announced a new "Wharton Geospatial Initiative." Wachter spoke of the need to "build a bridge," that is, to bring geospatial technology to business leaders and academics in order to transform business. Wharton is well positioned to do just that. Let us hope she manages to impart not just knowledge of the technology, but also to communicate a sense of the scientific methods and procedures that are so important to the effective application of the technology.
What is clear is that people who are listening closely to businesses - Wachter, Dangermond, Foust, and others - independently identify the need to improve the ability of business people to use the truly powerful analytical tools GIS offers. We have the data to make location technologies work. We have the software (although its ultimate architecture, as an "enterprise solution" or a set of "point solutions" built on top of an enterprise IT system, may be in doubt). What remains is to nurture the wetware that will cease to make "business intelligence" an oxymoron.