So predictable has this become that many of the informed IT people involved in such projects expect them to fail from the start. They naturally attribute this to the business not knowing what it wants and failing to deliver clear and consistent requirements.
For projects to succeed, consistency and clarity has to come from the top and be baked into the culture of the project.
In New Zealand, the tax department is embarking on an incredibly complex business transformation project that has so far cost NZ$83 million before a line of code was written. Overall, the program could cost up to NZ$1.5 billion.
The minister in charge of Inland Revenue is Todd McClay. Back in mid-2014, he appeared to understand the need for consistency and clarity.
At a breakfast briefing last June, he said lessons had been learned from the failed schools payroll project Novopay, which was trying to accommodate every permutation of the school payroll system in an unbelievably complex manner.
"One of the things we are doing at the forefront of Business Transformation is around simplification first," he said. "What concerns me most at the moment is getting that right for the tax system, because if we don't, we're going to have to come back and fix it."
Inland Revenue deputy commissioner in charge of change, Greg James, fleshed that out. The key, he said, is transforming and simplifying first. The department would "use the policy lever to avoid getting into a Novopay situation".
So far, so good.
However, McClay and IRD are in charge of another project, a rebuild of the child support system that has already cost at least NZ$130 million more than the NZ$30 million originally expected.
In this case, the "policy lever" helped create the problem. Worse, McClay's justification of the blowout dramatically contradicts his message of "simplification".
"We have gone from a very simple system of child support, which was inherently unfair, to a system that is far more complex but a lot fairer," McClay told Stuff.co.nz.
"The complexity of the law became more apparent the further the policy was developed -- hence the increased costs -- but the fairer system should ensure far more compliance and a reduction in the NZ$3.2 billion worth of child support debt outstanding."
Politicians are comfortable with ambiguity. In fact, it is their best friend. But the success of IT projects depends on clarity.
In the case of IRD's transformation program, that will require politicians to show policy restraint over what could be a decade of complex change. The odds of that are slim.
IRD appears (PDF) to be doing everything possible to mitigate the risk of failure. The word is the department has strong leadership in place, including James, to direct the project. The project is also not being rushed.
Leading consultancies have been employed to advise and audit the project team's work and project governance structures.
While no tender is likely this year, the signs are the department is also favouring commercial off-the-shelf (COTS) software where appropriate to further mitigate risk.
New Zealand's tax system is already one of the simplest, with a single rate consumption tax and very few income tax concessions and rebates.
Politicians are being sensible, too, at least for now. Last month, McClay announced a comprehensive review of the tax system, purportedly aimed at further simplification, well ahead of the technical phases of the transformation.
However, all of the big consultancies have been similarly involved in major project blowouts Projects based around COTS fail too, as in the case of Contact Energy.
A particular challenge, one where many projects get into strife, is likely to be the management and migration of legacy data into the new system.
And then there are those stark project failure statistics. Even with the best will, people, and processes in the world, delivering this scale of change on time, on budget, and on scope is a huge challenge.