The Real Story of Real-TimeBy John McCormick | Posted 2003-12-01 Email Print
Re-Thinking HR: What Every CIO Needs to Know About Tomorrow's Workforce REGISTER >
With few big success stories, don't bet the company on around-the-clock information anytime soon.
Let's get real about real-time. Sure, you want to gather a gob of data about everything your company does, pull it in around the clock, analyze it constantly and make decisions every moment. Yeah, you have the network and the technical chops to pull it off. And the result will be faster response to customer desires, better use of factories, less inventory in the pipeline and less time getting new products to market.
But is "the real-time enterprise" for real?
Few companies have made around-the-clock information networks work. McDonald's, for instance, backed off a far-reaching, five-year, $1-billion effort to keep up-to-the-second tabs on every shake, burger and salad sold at its 30,000 restaurants. It decided to just concentrate on making better products.
On the other hand, there's Delta Air Lines. The national carrier created a network that instantly links more than 30 customer and flight databases. Now, the progress of a passenger can be tracked simultaneously by ticket agents, baggage handlers and flight-control operations. The $1.5-billion system lets managers adjust flights, crews and gate assignments, literally on the fly. The system is saving the carrier an estimated $700 million a year.
This, however, is one of the few success stories. In most cases, establishing a real, real-time enterprise "is darn near impossible," says Tom Pettibone, managing partner at the consultancy Transition Partners and former information chief at New York Life and Philip Morris.
Pettibone says the main reason is cost. Linking sales forces, call centers, suppliers, distributors, customer service, and manufacturing is an expensive proposition. Then there's the cost of installing middleware. And when you get that done, there's the cost of changing old business processes and retraining staff to execute new ones.
Gartner, for instance, figures that a company with $5 billion a year of sales will spend anywhere from $80 million to $160 million every year for five years to make its business run in real time.
The message to project leaders: Be careful.
If you only need information every few hours, design accordingly. If only some parts of your business need instant information, proceed accordingly.
Landstar System is a good example of a company that's carefully approaching real-time. The $1.5-billion-a-year transportation company uses a mix of systems that deliver information in "real-time" and those that don't. One network takes information from shippers and presents it immediately on screen so truck drivers at any instant can see what freight can be picked up and where to take it. The prior, more-informal system favored drivers who had relationships with shipping agents. Now, new and old drivers alike compete on an even playing field, says chief information officer Larry Thomas.
By contrast, it's not necessary for Landstar to know every minute of every day that its drivers' licenses and permits are up to date or the exact moment a truck gets inspected. Its compliance system instead pushes such safety-related information to customers every few hours.
Even for these systems, which improve relations with drivers and customers, Landstar thought hard about potential benefits before development began.
That's a smart move. Even if you're just trying to speed up a single piece of the business, says Gartner vice president Simon Hayward, you're going to touch supply-chain systems, customer-service systems and the like. You may wind up rewriting software and you'll need to tweak the organizational structure. The costs—and impacts—can be great.
So why just do a "near real-time" system or just concentrate on a piece of the business? The real question is: Why would you do anything else?
The key, according to Bain & Co. consultant Tom Manning, is to start by being specific—real specific—about what you want to do and then stay focused on that objective.
Only then can you get real about real-time.