One well-known factoid in project management world is that 74% of all IT projects finish behind schedule. You know, I would have thought the number would have been higher.

When I worked in the software development group at Creo, I don’t think any of my projects ever finished on time (except for one), but pretty well all of them finished under budget. Actually, I don’t think I ever had a real budget for most of them. I was given a few guys and told to go do something. Then one day, I was project-managing about 10 guys sprinkled over three different teams (Note: PMing people is different than managing people. What I should say, is that one day I found myself trying to get about 10 people to do something. Which is different than being their functional manager. If you are a functional manager, you get to just tell your people to do something, and they better do it or their butts are in a sling).

Anyways, it’s no big trick to get a project completed on time, you either add more resources or de-scope some features. But note that in case one, you are spending more money, and in case two, you are going to disappoint some stakeholder. In my personal experience, it’s pretty rare to find a collection of senior management types who are willing to shell out more bucks just to meet some arbitrary calendar date. Oh sure, there are some projects which have very good reasons as to why they need to be completed by a certain date (for example, the Olympics) but it’s amazing how often a date is set just because… a project needs a end date. Whatever.

I did have one project that could not be late because the components were to be replacement for some OEM code and the contract to use that code was going to expire. Senior managment had made it know that they would rather crawl over broken glass rather than renegotiate the OEM contract. Surprise, surprise, beta ended two weeks before the expiration of the contract. When a PM has executive blessing to veto all and any scope changes, it’s amazing how everybody can adhere to the timetable.

The PMI guys have a nifty name for what is also called feature creep in the software development world: “Gold-plating.” This is term for a feature that is added on during the development cycle that wasn’t in the original project charter or project plan. Happens all the bleepin’ time. It’s very bad to allow “gold-plating” in a project, according to PMI.

So naturally, I allowed in pretty well all my projects (they made a PM at Creo before I could read the PMI stuff). To be more specific, if a request came in from the field for a feature or feature enhancement, I would take it to the developers to see how long implementing that feature would add to the schedule. On first meeting, I would usually get the brush-off. Then I would let the matter wait for about 3 days or a week. Then I would follow up. If it wasn’t too hard to implement, then we would try to do it. Naturally, the project would be a little late, but if I got happy points from the field, than who cared about deadlines?

Mind you, most of my projects were weeks late or maybe a few months. Proejcts that are massively late are usually either very high-risk projects where everybody is waiting around for some genius to design the crucial anti-gravity components or “drifty” projects that have poor management. Actually, when I say “poor” management I mean managers who are too chicken to kill a project that deserves killin’ because they are scared to lose their jobs. So they keep the projects goin’ and goin’ and hope that senior management doesn’t notice that the project that was suppose to take six months is now into year three.

Show me an experienced PM who has never taken an active part in slaying a project, and I’ll show you a bad PM.