Technology vendors have taken a verbal hammering from the Australian Defence Force (ADF) after one of its top procurement chiefs blamed the industry for most of its IT project failures.
Kim Gillis, deputy chief executive officer of the ADF's procurement arm, the Defence Materiel Organisation, said vendors set unrealistic expectations in tenders -- which was usually the cause of those government IT projects failing.
Government tenders were often surrounded by "a conspiracy of optimism," said Gillis.
"Say I'm going to put in an IT system in 2000-and-whatever, and go out to industry and say 'I want you to give me this type of capability'," he told delegates at the Gartner Symposium conference in Sydney.
"And miraculously everybody who tenders comes in and says 'I can deliver that capability exactly how you specified on that day'.
"And everybody starts believing that it's a reality," he said.
DMO project managers were given a simple instruction for dealing with such companies, according to Gillis: "Don't believe it".
"Especially in the IT world, because I haven't seen in my experience in the last five years, an IT project delivered on schedule," he said.
"They do happen, but I haven't seen them."
False promises have often led to government IT project failures, according to Gillis. However, it was usually the government that wore the blame.
"The reality is the people who actually got it wrong are the industry participants who are actually providing the services," he said. "Most of the time the fault lies not with what I've actually procured but what they've actually told they're contracted for.
"At the end of the day what happens is, they've underperformed, [but] I take the hit," he said.
The DMO recently took steps to improve its procurement process by instigating the Procurement Improvement Program (PIP). It includes a series of consultations with industry and changes the tendering and contracting process.
When someone blames everyone else for failures, you can pretty well guess who's one of the biggest contributors to those failures. With quite a number of years in software project management (on both sides), I know that customers themselves can be one of the biggest hurdles.
A software development project is a very, very different beast from buying a car or a tank. When I look at a large software project to assess whether it's going to succeed or fail, the first thing I look at is the customer. Who is the sponsor? Is he full time on the project? Is the customer CEO involved and have detailed visibility and interest? Without these things, the chances of a software project failing grow by some 3 to 10 times. And that is regardless of the skill and resources of the vendor.
Sure the sales guy oversell capability, delivery and cost - but anyone who beleives that are those guys who beleive weight loss cures and other snake oils. And I admit many software project managers also believe the fairy tales spun by software team leaders. But day by day, hour by hour, minute by minute involvement and committment by the "real" end user (not consultants or proxies) right through the specification, development and deployment phases is essential. The cost of two or three staff guys is small compared to the blow-outs we see.