On IT Project Failure
CFO Ben Courtanes: The first thing I'd like to talk about, Harold, is IT project failure.
CIO Harold Peeples: Well, that's a gloomy way to begin, Ben, but, OK, go ahead.
CFO: Numbers don't lie, Harold.
CIO: Spoken like a CFO, Ben.
CFO: Be that as it may, the Standish Group's numbers have shown that IT projects fail a lot, and that the bigger the project, the more likely that'll be the case. This has been true for decades. Can you blame me for being sceptical about IT?
CIO: It really depends on how you define "fail", doesn't it? I mean, just because a project misses a few deliverables, or has to be changed after it's deployed, that doesn't necessarily make it a failure. Maybe it's just less than perfect.
CFO: I'm not looking for perfection. But what you call less than perfect, the Standish Group calls challenged. Those projects end up with fewer features than you promised, they cost more than you said they would, and they take longer than you expected. That puts me in a bad position. If we pull the plug, we've wasted millions. If we don't, we risk wasting more.
CIO: That's a Hobson's Choice, Ben. I don't think we've been in that position here.
CFO: We're in that position right now with the wireless project. Your proposal promised all sorts of ROI for adopting a wireless network. You said we could save on networking costs and get more productivity because our workers could access the network from anywhere. But it hasn't worked out, has it?
CIO: OK, there've been issues. But by doing that project now, we're leapfrogging to next-generation technology. And we'll save money at the same time.
CFO: We will? When? After we've spent twice what we planned?
After the project started, you told me there was a problem because our buildings had steel in them. It interfered with the signals. So I come in one morning and we're tearing down walls. To date, that has cost an extra half-million. The whole thing's been going that way. Your support costs turned out to be twice what you said they'd be because you had to fix a security hole in the wireless protocol.
Even if this thing does save us money down the road, all these extra up-front costs have to be figured into the ROI. You can't just tack them on and pretend they don't affect us. When you say a few deliverables will be missed, I hear someone changing the business case on the fly. When support costs go up, I have to take money out of someone else's pocket. When you find a security hole, I see the risk profile changing completely.
CIO: But, Ben, this is the nature of software projects and, frankly, it's an inherent feature of any big challenge. What about when we take on a big merger? Remember when we acquired UpandComing Ltd? That started out as a happy story about synergies and complementary product lines. But after it was finalized, we started hearing about cultural differences, branding conflicts. That project didn't come off exactly as you hoped it would, either.
CFO: Point taken, Harold. Then again, we actively attempted to minimize the risk from the beginning. We created detailed risk models. We evaluated operating costs. We estimated market shares, workforce overlap.
CIO: Software projects are different. They pose hard-to-measure risks and sometimes offer hard-to-measure returns. We're dealing with an imperfect science here.
CFO: Imperfect yes, but also largely predictable. Lister and DeMarco's book on software risk, Waltzing with Bears, boils it down to the same five major risks for every project. They even built risk curves for them.
CIO: I know. I read it.
CFO: I'm not asking you to make IT projects fail-safe, Harold. That would be as foolish as you asking me to always keep the stock going up. What I'm asking for is a lot more discipline, especially up front. To measure risk. To consider all the forces that might affect a project, not just the ones from within your own silo. Think of it this way: Replace the word project with the word product. Would an automaker tolerate one recall for every three cars it made? What if half of Arnott's chocolate chip cookies crumbled before you could eat them, and another quarter didn't have any chips?
CIO: Actually, Ben, we've been trying to address your concerns on two levels. One, we've established a project management office to balance our project portfolio. We're starting to use some basic risk analysis. And two, we've adopted some agile development techniques to improve code quality and production efficiency.
CFO: That's a great start. But it's only the beginning. We need even more discipline and empirical analysis. Because all it takes is one troubled project to screw things up royally. When I got the update on your wireless project, tell you the truth, I started thinking about Nike's inventory mess. And that hospital in Boston that had its network go down for four days. I read about that in that magazine you're always dropping on my desk. What's it called? CIO? Story by some guy named Berinato? When these things go bad, you start talking about the vagaries of software. Me, I've got to account for it. Literally.
CIO: OK, Ben. I hear you.
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.