For a while, I became variously famous and infamous for shutting down one of the highest-profile ERP implementations in the US. My company at the time was Dell Computer, and the ERP software we were trying to implement was from a country not known for its sense of humour. Dell had purchased the software and started the project about six months before I was hired. Hired, I might add, to keep this project - which was about 20 percent complete and already millions over budget - on track and implement it quickly.

The simplest part of that whole episode was the actual decision to shut the project down. In short, we determined that, because of our unique business model (individually custom-built PCs) and the way the ERP software was designed (large production runs of the same PC), we couldn't process a day's worth of transactions in less than 24 hours. This was not good news, except that it made the decision to pull the plug pretty easy. That this problem wasn't discovered earlier in the project was probably a combination of wishful thinking on our part and a lack of experience on the vendor's.

In preparing to break the news, the money thing was the easiest to explain. We developed a summary of what we'd already spent and what we were likely to spend to finish this system-that-wouldn't-work. It was a blessing that much of what had already been spent could not be quantified - things such as delayed enhancements on legacy systems, long hours for stressed people and lost business opportunities. It also helped that the millions it would take to finish amounted to far more than what had been spent.

The "what do we do now?" question was much harder. It would have been irresponsible to continue spending while I figured it out. So the answer would have to be: "I don't know yet". During my presentation, when I put up a slide with just those four words on it, I looked at the audience and had the sense that I was staring at an oil painting.

By far, the toughest part of developing a message that wouldn't get me killed was remaining sensitive to how most of these nightmares got started in the first place. Here's a hint: It wasn't from the office of the CIO.

While it isn't quite so true today, a few years ago it was common to see advertising from "solution providers" targeted at CEOs and designed to shake their confidence in their IT departments. Their increasingly hyperbolic promises had the effect of ratcheting up expectations beyond anyone's capability to deliver, including theirs. Combine this with the phenomenon known as corporate herd mentality, and almost everyone was sold on the notion that complicated and innovative solutions could be delivered with the regularity and reliability of a Swiss watch.

To his great credit, Michael Dell, the executive sponsor of that project, graciously accepted my conclusions and was very supportive in our efforts to get things turned around. Things didn't go quite as smoothly with the hordes of consultants clogging up our cubicles or the software vendor who, while not disputing our conclusions, felt that we should pay for the software anyway. The worst of it came one night, at home, when I got an angry call from a member of our board of directors who also happened to be a member of the software vendor's board.

Slept pretty darn well that night!

In the end, though, we all survived (very well, I might add), and within a few months had developed an alternative approach. I could actually start breathing again. Looking back, it's absolutely clear that had we not shut that project down, our systems would have severely inhibited the explosive growth Dell enjoyed in those years. However, it would have been nice to be that sure at the time.

Whew!

Jerry Gregoire (former CIO at Dell, Pepsi and elsewhere)