In Location, Location, Location Joe Francica, Editor-in-Chief of Directions Media writes about how location intelligence can enhance BI scenarios. The article does a nice job of discussing how adding location information and visualizations of same can really enhance the value of data and ease decision-making. However, it is rather focused on "decision support" and misses the value of location intelligence in decision automation. So how would location intelligence be embedded in a decision automation or decision management environment? What kinds of examples might you see?
Decision management is focused on operational or transactional decisions. These tend to be high volume and key to a specific transaction. Examples would be making an underwriting decision for a policy, approving or declining a loan, making a cross-sell or up-sell offer etc. Some of these could be location dependent. For example:
- A property underwriting decision could depend on whether a property is in a flood zone, too far from a fire hydrant or at risk from earthquakes
- A commercial underwriting decision could depend on the proximity of high risk businesses such as fireworks factories
- An offer could leverage the knowledge of the nearest outlet for a retailer to make a store-specific promotion
- A fraud detection decision could use the location of an ATM and its distance from the last use of a card to see how likely it was that a copy had been made
- A repair notification system could use the location of a piece of equipment that failed and the locations of engineers as part of deciding how to dispatch to fix it
- and so on
What all these examples have in common is the use of location intelligence in a business rule or set of business rules. The rule might be something like
- If thisproperty's location is in_10_year_flood_zone or thisproperty's location has distance_from_fire_hydrant > 100 Then...
- If at least one of thiscompany's locations has distance_from_risky_business < 10 Then...
- If customer's bestoffer has location_dependent=true Then bestoffer's store=customer's location.getClosestStore()
Location intelligence can also be used in analytic model development. If I am trying to build predictive models, of risk say, I can enhance every item in my development data set with additional location information. For a risk of damage model I might use factors like flood zone, elevation, distance from dangerous businesses etc. These location facts enhance the base facts and can be used in the analysis I do to build a predictive model. The modeling process will establish which facts are good predictors and these will then need to be used in the runtime decision.
Both these examples allow you to locationally enhance the decisions you are automating - Location, location, automation.
I posted previously on this Location Intelligence and EDM and there is a great webinar on the topic for Insurance folks at Location Rules! Optimizing Insurance Processes Using Business Rules Management with Location Intelligence