Internet Explorer 11 is not supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

5 Ways to Keep a Big IT Project Out of the Ditch

Some failure along the way is inevitable. It's essential to plan for it.

For the government information-technology professional, there's probably nothing more exciting than the once-in-a-career opportunity to replace a decades-old system, leapfrogging generations of technology and bringing an agency or department into the modern age.

But the enthusiasm required to gain a green light for one of these transformation projects can sometimes produce a near-evangelism for it, blinding sponsors and project managers to a project's potential pitfalls and becoming a major liability once the project commences.

There's certainly nothing wrong with having an enthusiastic project team. But neglecting to imagine anything but a pothole-free path to success is a huge mistake. Some failure along the way is inevitable.

Here are five practical lessons I've learned, sometimes painfully, in over two decades working in government IT, including as an appointee for two governors:

1. The transformation you are embarking on is likely unprecedented for your agency, so don't assume you will make each decision correctly the first time. Plan in both your schedule and budget for some measure of re-work, adjustment and do-over. You're not sandbagging or being wasteful -- you're being practical. And you're avoiding the temptation to hide mistakes when they inevitably occur, because you've planned for them.

2. Design a system that won't break, not just one that will work well. This may seem like backwards thinking, but anyone can design a system to work the right way. You need a system that users can't break and that can handle the reality of your work. Most failed systems "work as designed." What does that tell you? It tells me that not enough thought went into how the systems would work in the real world, where data can be bad or incomplete and where humans will inevitably make mistakes.

3. Be modest in your goals; don't overreach. You may believe you can get from A to Z in one effort. Great, but I'd advise you to consider committing to A to G first and then build in a couple of planned phases to follow. This will give you a chance to both prove success and work out the kinks on an implementation pattern. It will also allow you to effectively reserve the right to either continue with or change your vendor at each phase.

4. Proper staffing of both your team and the vendors is essential. Sounds pretty basic, but many projects create a single point of failure at key positions. That's a mistake. Assume that these positions will turn over and have successors in mind. Make sure your vendor team is also not one-deep in key roles.

5. Be ready to compromise. On a large system implementation, every day is a negotiation about what can and cannot be done within the budget and schedule that remain. Be prepared with a list of what you can give up and what you can get when the negotiation happens. Getting caught flat-footed will put you at a disadvantage with the vendor.

While I hate to put a damper on the excitement of embarking on a new project, these five sober steps will help you to be better prepared. Guarding against failure can produce great success. As the top coaches say, "Defense wins championships."

A KPMG principal who oversees government information-technology projects in the Pacific region
Special Projects