13 August 2008

IT Project Management IS rocket science!

For years I've been telling people that delivering successful IT projects is not 'rocket science'. After all we've got a huge body of past experience to learn from, we've even got a Project Management Body of Knowledge! But still we see projects failing for reasons that get enumerated and dissected post hoc (often in this very blog).

But now it is time to admit that I was wrong. Clearly delivering successful IT projects is like rocket science, and it is time we acknowledged that it is hard to do successfully.

The recent 2006 Chaos Report from The Standish Group shows that 19 percent of development projects were outright failures, compared with 31.1 percent in 1994. This is not much of an improvement over 12 years, rather it is an embarrassment with approximately 1 in 5 projects failing. It tells me that we are completely wrong headed in saying that IT project management is easy and if everyone just followed the rules all would be well.

What we actually need to do is to admit that IT projects are hard, complicated and prone to failure. They are, in fact, just like rocket science and we need to put our best and brightest to work on them. We need to ensure that everyone working on projects is actually trained to work on projects, acknowledging that it is not a place for the gifted amateur.

Nobody would put untrained people to work on building rockets to send to the moon, only trained rocket scientists are welcome on those kind of projects. We need to become just as focused on getting people with the correct skills on IT projects. The skills development even needs to be pushed out to business owners of projects so that they understand the basic science of projects, and why certain things have to be done at certain times.

It is time for the entire industry to stop playing around and wasting huge amounts of organisational resources and capital on unsuccessful IT projects. It is time to apply discipline and skill to the delivery of IT projects. No longer should it be the preserve of the gifted amateurs within the organisation. Instead, we need to spread the word that delivering successful IT projects is exactly like rocket science and that it needs appropriately skilled and trained resources to apply the body of knowledge.

By Carruthers via Aide-mémoire


Saul Cunningham said...

Hi Kate,

couldn't agree more. I have often used the analogy of how we build any engineering product and the comparison to IT. Think of a bridge, plane, building etc. With these types of products we can say at design time with a great degree of accuracy whether it will work or not. What I mean is whether it will fall down or stand up, fly or crash etc. We can take the design documents/models and validate them - why can't we do that with IT? Now of course time and budget overruns occur with these engineering projects but they don't really result in a disfunctional outcome that simply fails.

Now why is that? I can only postulate two possible reasons:

1. We are simply not applying the same level of skills and discipline to IT projects that we apply to normal engineering projects


2. IT projects are just more complex than say building a rocket.

Now I find it hard to accept reason 2. It is difficult to believe that building your typical IT project is more complex than building a substantial engineering project where success seems to be more predictable.

I can only concur with you and say that I believe that the skills, discipline and execution of the typical IT practitioner need to increase.

IT vendors also need to work on tools that can model the IT system design, simulate its execution and give guarantees of outcome with accurate "factors of safety" in the same way a civil engineer can model a bridge or a space engineer can model a rocket.

Personally I believe that this is an area that will develop as system design starts to solidify around accepted frameworks and architectures that fall into what we now call SOA. As we move towards an environment where all major vendors applications and technologies follow this pattern proper modelling with accurate predictive outcomes will emerge.

I am writing a blog at the moment that will appear over at the Oracle Red Room http://oracle-gtmi-anz.blogspot.com that will talk about one aspect of this called Governance. It should be up in the next week.

Saul Cunningham
SOA Business Development

dw said...

Hi Kate and Paul,

My view on this is that the world of IT is changing insofar as it applies to business processes - as opposed to rocket building where completion is the object itself.

For those of us principally engaged in outwardly - client - facing IT business commerce systems, the Web has now upturned our world. Permanent Beta is the reality we face. There is no end in itself, for the end now is in fact the journey.

Business value expressed as a spec for a build at the beginning of a project is ipso facto now, no longer the outcome a client will expect. The building process itself informs/refines the anticipated business value outcome.

The dilemma this unpredictability presents can only be overcome by substantial improvement in the flexibility with which project contributers interact with each other. Particularly in the marketing platform builds that predominate HotHouse's core business activities, continuous iterative shifting of the goal posts is now de rigueur.

Are you familiar with Alistair Cockburn's perspective. I'm drawn to it by virtue of its emphasis of the human interactive element - 'methodologies meet people'.

If you are familiar with his approach I'd much appreciate any feedback you may be able to pass on regarding its efficacy in practice.


Dion Weston
HotHouse Interactive

Craig Brown said...

Hi Kate

The Standish group pin the marginal improvements ont hings like better PM education and smaller projects.

These explanations not only don't sound right, but they aren't really relevant.

The big issue is that success rates shuld be up over 50%.

What are we doing to radically change the landscape these days?

Agile? Scrum? XP?

There is no real solid evidence i have been able to find that says these processes will stand up once the early adopters ave been replaced by the bulk of project pracitioners.

Any ideas?