Little (IT) Shop of Horrors

Halloween is the time for goblins, ghouls and ghost stories. But if you’re a CIO, chances are you’ve heard one or two IT terror tales. John Parkinson, former chief technologist at Cap Gemini Ernst and Young, has guided hundreds of companies through veritable nightmares. Here are his 8 favorite failures. Company names have been withheld.

A Dark and Stormy Night. A multi-company “collaborative” development project was spiraling out of control. A year late and only a third complete, the original $40 million budget had skyrocketed to $150 million—and nothing had been delivered yet. As the three separate “sponsoring” organizations argued over who was to blame for the failed project, a coven of hungry attorneys circled. What to do? Cancel the project and start the lawsuits.

Of Vice and Men. A post-merger integration review of IT operations by the acquiring company’s technology team discovered that an employee of the acquired business had been running a pornographic web business using company resources. No one knew—despite a due diligence process. Turned out the illegal business represented over 33 percent of the acquired firm’s total network traffic.

The Guillotine. A divisional CIO of a large multinational found himself stuck in the middle of a data center consolidation project. The problem: The lease on the old data center expired 90 days before the new site was supposed to be ready. In a panic, the divisional CIO moved the entire operation to the business continuity/disaster recovery center—where it ended up staying for more than a year. The company lost $1 million per month; the CIO lost his job.

Upgrade to Disaster. A mid-size company embarked on an Oracle ERP upgrade that was supposed to include all the patches applied to the old version. But when the upgrade went live, nearly a third of the patches were missing, extending the project by one business quarter.

Ghost in the Machine. A major financial institution and asset manager experienced a 5-day outage in its mainframe systems following a major OS upgrade. Because the event occurred over a holiday weekend, the escalation process with the outsourcer (intended to keep such things from happening) failed to work.

Space Invaders. A mandated upgrade from Oracle DBMS Version 7 to Version 9 required three times the physical disk space for the same volume of data-but no one noticed until the conversion was already well underway. It took three months to get the new storage delivered and installed, delaying a major rollout. Because the project was part of a corporate spin out, the new business entity was forced to remain on the old company’s systems for an extra quarter, incurring huge financial penalties.

Botched Enhancements. In efforts to reduce costs and improve operational capabilities, a large financial services firm took on a larger IT outsourcing contract for extensive and complex infrastructure enhancements. Service levels plummeted, and the substantial penalty payments were still insufficient to cover the business losses-let alone confidence among users.

Production Panic. A company using a large IBM mainframe realized that 20 percent of all “production” jobs failed on every run-and instead of fixing the problems, staffers were applying “work arounds” every time the job failed. The resulting re-runs accounted for almost 11 percent of consumed MIPS and doubled the length of the batch window.

Get the Free Newsletter!

Subscribe to Daily Tech Insider for top news, trends, and analysis.

Latest Articles