Dump the IT underperformers

By now we have all heard the inflated claims made on behalf of IT projects. "We are going to transform the way we do business. We will work in a highly collaborative way, adapting to changing requirements and focusing on maximising business benefit. We will deliver precisely what the business needs two months ahead of schedule while generating a significant investment return."

I've heard it many times over the years but it still makes me laugh. We all know IT projects always take longer than planned, cost a lot more than budgeted and deliver less than expected.

Not surprisingly, this common reality isn't quite so funny. The insurance industry, not uniquely, is littered with failed or failing IT projects. Indeed some would argue that the expectation of failure has almost become the norm and that success is measured by how effectively the overspend is managed or underperformance is communicated.

A recent Forrester report stated that a third of companies are unhappy with the time taken to produce poor quality applications that fail to support business needs. But what are these companies doing to change this?

The insurance industry is in many cases still allowing multi-million pound investments to rely on outdated delivery systems and then wondering why it sees no return on its investment. So why do so few people question such methods, instead blaming poor performance on the complexity of the business variables of each project.

These outdated approaches to software development mean unacceptably high levels of risk creep into projects.

Project success is about removing risk. This may sound obvious, but many insurance firms and their software suppliers ignore common sense and resulting in lengthy, uncoordinated development cycles that unsurprisingly lead to project failure.

It's time to turn the approach to IT delivery on its head, stop talking about business value and actually start delivering it.

For example, research shows that 40% of functionality is never used; developing, testing and deploying this functionality is a waste of both time and money. The single, simplest thing that organisations can do to improve the delivery of their systems is to stop building things they do not need.

Once the right decisions have been made about the functionality, organisations need to build for deployment. Easy, fault-free deployment is the key to delivering value in the short term and staying flexible in the long term. Organisations need to use business feedback instead of gathering requirements as a successful way to deliver systems. Combined with frequent software releases, this ensures that the resulting system is lean and easy to change in the future.

There is already a set of established tools, techniques and principles that deliver this level of performance, collectively known as agile software delivery (ASD). With ASD you can maximise return on investment (ROI) of systems while optimising budget management.

ASD's highly automated, customer-driven approach to delivering functionality allows the release of working features, in business value order, earlier than any other methodology and at significantly reduced cost.

So, perhaps the time has come to amend our IT expectations. "Latest insurance industry IT initiative goes live ahead of schedule and below budget. Early adopters of the system, who have already seen a higher than expected ROI, said today the project has been a complete success in all areas and marks a major milestone for IT delivery in the industry."

Suddenly it doesn't seem so funny. IT

' David York is head of UK insurance practice at ThoughtWorks