John Parkinson*, Chief Technologist for the Americas for Cap Gemini Ernst & Young, spoke at the KMPro Chicago meeting last night on the topic of The Real Time Enterprise, which was the topic of his January 2003 commentary in CIO Insight Magazine.
Real Time is essentially the ability of a system (whether that is an oil refinery, fighter jet or corporation) to sense the environment and respond to that information "quickly." In the world of a fighter jet, there are 1800 sensors and massive automation that are automatically adjusting the flight parameters within milliseconds without the pilot needing to be aware. In fact, the pilot cannot fly the jet without the assistance of the computer. "It becomes a very expensive brick with wings" without the onboard systems.
In the enterprise, real-time isn't quite this complex, but the implications are huge. Parkinson claimed that corporations can see massive jumps in their bottom line (30-40% in manufacturing asset use) by operating closer to the edge of the envelope that the real-time concept allows.
How does it work? I understand this from an engineer's perspective, and Parkinson did a good job of making that connection for me. Traditional control works to keep systems static and in mathematical "valleys" so that the system moves back to the bottom of the valley when perturbed. Real-time looks to operate the system at mathematical peaks, and the control systems have to be much more sophisticated because any perturbation makes the uncontrolled system fall off the peaks. These kinds of control systems are not simple and cannot be run by human hands alone. The control system must monitor tens, hundreds or thousands of sensors, depending on the application, along with understanding the overlying environment (is the plane flying at 30,000 ft or at ground level?) and react accordingly. Parkinson showed us a familiar video of an Airbus crashing at an airshow because the control systems weren't prepared to do a touch-and-go manuever. The systems need to be able to ignore data in some circumstances.
There are potential downsides to operating this way. Attempting to control on the peaks means that falling off will hurt much more - see video above. Parkinson argued that the benefit of doing this far outweighs the inevitable, but hopefully rare, tumble down the mountain.
Where does it apply in business? GE has their digital cockpit, where all management can view their scorecard that involves everything from defect rates (six sigma measures) to customer satisfaction. Parkinson didn't suggest that there was a computer control system on top of the cockpit, but there is a very human control system that fires up when one of your indicators turns red: You get a call from your boss (who got a call from his boss), wanting to know what is wrong and what you are doing to fix it. The smart manager has already seen the red and has a response in hand. The really smart manager knows when her metrics are heading the wrong way and corrects before her managers ever expects something is amiss.
Other applications include shipping and automated sorting and routing of packages; automatically adjusted insurance rates, based on your car's GPS data; rental car variable pricing to account for popularity, time of year, competitors' promotions; airlines overbooking flights on the assumption that there will be cancellations; ABS brakes on your car. Also, mortgage lending, fraud detection, product pricing, warehouse management. There are opportunities in areas where there are too many variables, where the rules associated with decision making are complex, and correct decisions must be made quickly.
And as always, the real-time enterprise will not create itself. There must be good understanding of the changes that the new approach will engender in the business, and these must be tackled. Parkinson also talked about being aware of the law of unintended consequences, and that there are always secondary effects that are difficult to know beforehand. "Humans don't deal well with changes in control modalities."
Parkinson is an excellent speaker, and gave us a very interesting hour for the meeting. KMPro Chicago has had an excellent variety of talks, and it is attempting to do more of the same.
* More info on the speaker: John Parkinson writes and talks on a wide variety of technology-related topics, both on his own and in his role as Chief Technologist for CGEY's North American operations (press release). He was with E&Y in a number of roles before this, including a one-year stint as the CIO which he says he would never do again. But he also says that during his tenure, they built one of the first KM systems AND that it created ten times more value than it cost. A Google on his name + CGEY turns up ~157 hits with articles and even a rant against one of his articles.