IBM'S OLYMPIC FIASCO DEPARTMENT OF GROUNDLESS OPTIMISM
(FORTUNE Magazine) – Big Blue screwed up at the Olympics, no doubt about it. IBM invested well over $40 million in labor and equipment to fulfill its role as official Olympic sponsor and MIS department, but in the opening days of the games, part of the computer system IBM installed was impossibly slow, garbled data, and completely omitted critical information. It showed one pugilist to be 2 feet tall and another as 97 years old. It forced wire services for a few days to input statistics manually from paper forms hand-delivered by runners. The fiasco was reported all over TV, the Web, and the papers, and through it all ran IBM's multimillion-dollar Olympic ad campaign that (initially at least) touted bulletproof reliability. If self-parody were an Olympic sport, IBM would have medaled.
That said, there's more to the story.
For one thing, it wasn't a complete disaster. Most of the technology IBM installed actually worked. In fact, all except one of the seven major systems IBM built for Atlanta functioned at least passably well. Unfortunately for Big Blue, the one that didn't was the most visible--the system that was supposed to provide competition results to the one group of people who most needed accurate and timely information at the Olympics, the world's press. "We fouled up with the people who buy ink by the barrel," says IBM spokesman Fred McNeese. The press, of course, reported the story ad nauseam, even blaming IBM for things it had nothing to do with. In the aftermath of the tragic bombing in Centennial Park, for example, the Philadelphia Inquirer erroneously reported that an IBM system may have contributed to security lapses.
Whatever did go wrong, it isn't fair to blame IBM alone. The organizing committee, the Atlanta Committee for the Olympic Games (ACOG), was closely involved with IBM in the development of the technology even before Atlanta won the right to host the games, and must share some of the blame. Reports during the games suggested that the failures were a shocking surprise, but in fact ACOG knew well before the Olympics began that problems were possible. Several weeks before the opening ceremony, ACOG offered 12 news organizations a refund of one-third of their $15,000 fee and lamely suggested they use the savings to conduct their own tests. After the problems arose, ACOG refunded the entire fee.
Inadequate testing is an exceptionally egregious "don't" in the world of programming, especially when the technological and logistical task at hand is as complex as the one that faced IBM and ACOG. Bud Weydert, who managed Olympic logistics for the Associated Press, says he's not surprised the press system proved problematic. In past Olympics each news organization stipulated exactly what data format it required and contracted with Olympic organizers for a custom feed (EDS handled this last time). But this year IBM and ACOG opted for a standardized data feed that sent the same raw data to everyone, to be customized on the fly by each agency. "They significantly underestimated the task before them," says Weydert. Yet he applauds the effort: "We need a universal results format for sports events, so we don't want IBM to get a big black eye over this thing." (Too late for that.) Adding an unprecedented amount of detail about individual events while at the same time going to a universal press feed appears to have overwhelmed the capabilities of IBM's programmers. "Anyone looking for a single point of failure would be disappointed," says Bruce Sanders, a key technical executive in IBM's Olympics effort. "It was a multitude of small things, each of which on their own we deal with every day. But a confluence of those got a little ahead of us." Perhaps IBM's biggest mistake was not telling ACOG bluntly that the whole effort had become too difficult.
None of this lets IBM off the hook, of course. But in the end it probably doesn't matter. After all, the company already has contracts to manage technology for the next two Olympic games. Even the irony-defying--some say dishonest--ad campaign didn't turn out that bad, or so IBM claims. The company suspended its Olympic print ads for two days, during which it conducted focus groups and surveys. The results, company officials say, showed that most people were unconcerned by reports about the press system, so IBM resumed the campaign. Maybe it should have waited to resume the ads until it had the bugs completely worked out. Like, say, at the 1998 Winter Games in Nagano, Japan.