Published on the 21/07/2015 | Written by Beverley Head

CIOs need to reframe their definition of success and how they fund software development – and it’s a particularly big ask for public sector IT professionals…

Speaking at the Gartner Application Architecture, Development and Integration Summit in Sydney yesterday, research director Darryl Carlton told delegates to; “Stop aiming for successful projects – start aiming for successful applications or products.”

He said that some Australian organisations – Carsales.com for example – had shifted their thinking to the extent that the development team had made itself accountable to the end user consumers of its systems rather than to internal business users.

Carlton said that when consulting to Queensland’s Department of Justice he had used Survey Monkey to actually test how end users felt about an application which gave IT teams clear feedback about what was working and what was not.

That approach, said Carlton, demanded a different mindset, because although an IT project traditionally has an end point and a fixed budget, successful applications don’t – they change over time and require constant work to ensure they remain useful and relevant.

He stressed that there was little point coming in on time and on budget if the end users hate the system.
Carlton said because there was no longer an end to a programme of work, instead just a first release of an application with more functions added over time, there was a need for; “A whole different approach to funding,” which he said was particularly challenging for the public sector.

“Start-ups work in this way – with venture capitalists you never get all the money upfront, they drip feed based on delivery and it’s a very good model to follow.

“When I look at failed projects – they are inevitably those with poor business cases, a lot of money upfront and didn’t understand what they were building.”

CIOs and other executive stakeholders needed to understand that; “The end of the project is just the first release… no one can design a complete solution, ever. Most applications are out of date by the time the project is up and running,” he said, adding that the average lifespan of a mobile phone application was now less than nine months.

The need for development agility did not mean however that design principles could be eschewed. The opposite in fact, because the organisation needed to have a very clear understanding of what it was trying to deliver for end users, and also understand that it should not strive to develop a “Rolls Royce when a Volkswagen would have done.”
He said that Gartner’s analysis had found that a major cause of failure was IT teams agreeing to do something that is too complicated. “Projects should be simplified,” he said.

Carlton offered the example of Queensland’s failed $1.2 billion payroll project which led to the Government suing IBM. He said that around the same time Hong Kong Health took on a similar project, but unlike Queensland which attempted to develop a system to cope with an extremely complex matrix of pay calculations for its staff, Hong Kong first negotiated a deal with staff which delivered them a 10 per cent pay rise if they agreed to accept one of just five award schemes.

“The first thing you should be doing is simplify the business outcome then architect and design for delivery,” said Carlton.