The Cost of Large IT Project Failures

3 Dec ’05

Mary Kirwan has an interesting piece in the Globe that looks behind the scenes of large IT project failures to consider why they occur. Much has been written on this topic, but Mary has some interesting contributions, including an interview with Professor John McDermid, Professor of Software Engineering at the University of York, UK, and one of the authors of a UK Royal Academy of Engineers and British Computer Society report on project failures:

‘There are lots of good analysis tools available, and fault densities can be very low. In many commercial products, there are 30 faults to every 1,000 lines of code. They should be able to deliver, with existing tools, one in a thousand faults routinely on normal IT projects’.

So if it can be done, why isn’t it done?

‘Lots of good practices are ignored, and customer expectations have traditionally been so low that the vendors got away with it. The market has created this fault tolerance over time, and it is not an easy shift to engineer quality back in’.

As for the reason for IT project failures, Professor McDermid believes that there are ‘usually a combination of factors that contribute to a bad end’. They include ‘the complexity of large projects’, as protagonists get ‘swamped by complexity, uncertainty, changing requirements, and changes in the organisations themselves’.

Indeed, he believes the latter is a major factor, as many failed projects result from a desire to use IT to ‘force organisational change’. ‘They specify an IT system, and expect change to occur, but people can’t live with the changes, so the system fails’.

Research bears him out, as there are abundant studies that demonstrate that successful outcomes in IT projects depend only 25 per cent on the technology component- the rest is entirely attributable to organizational change aspects that are frequently ignored.

However, he also attributes failure to ‘a lack of professionalism’ in the software development field, and a dearth of competent project managers. ‘IT guys are often managing blind’, and they are ‘not good at overall risk management’. Senior managers may be expert at risk management, but ‘they don’t understand software risks’. And neither one speaks to the other.

Previous post:

Next post: