blog I often think that things couldn’t possibly get any worse for State Government IT operations in Australia, considering that major audit reports in both Victoria and Queensland have found over the past year that the states are broadly incapable of delivering IT services and major IT projects to their departments and agencies. But every time I think that, things do get worse. Today’s new nightmare is a bungled student management system in Victoria’s TAFE colleges. The Age tells us (we recommend you click here for the full article):
“The cost of the computer system, commissioned under the previous Labor government to manage student records, has ballooned to about $100 million, well over its estimated cost of $65 million.”
This sort of thing — a mere $35 million blowout in a key student management system in TAFE colleges across Victoria; nothing, really, when compared to a billion-dollar blowout such as happened in the Queensland Health payroll disaster — just reinforces my view that it is currently the norm in state government IT around Australia that IT projects of any kind fail. It is absolutely the norm that these projects go over budget, over-time and fail. It would be highly abnormal if a State Government was able to actually deliver a major IT project — an extreme outlier. This is the situation which Australia is currently facing in State Government IT. We must assume that any major project will fail and go massively over-budget, and plan accordingly. Assuming anything else is simply not rational.
But hell … do we really need to actually get data on student enrolments in Victorian TAFEs? You know, information on what courses they’re in, when they’re graduating, that kind of thing? It’s not really that important to the functioning of the TAFE system, is it? Right? Right??
Image credit: Timo Balk, royalty free
As a software weenie, I always find myself asking “what is so hard about the system that it blows out by so much”. Are these IT depts bad at managing schedule? Do they just have no idea what they actually want and so lump an incredibly tight but self contradictory set of requirements into the mix?
I don’t mean to sound naive, or as though I am trivialising what is possibly a complicated rollout. But $65M is a fair amount of dosh for something that *sounds* like it could be a 3rd year comp sci project. I can’t see any frontiers of comp sci being pushed in what should essentially be a slightly large database.
Ive been programming for 30 years, though only a few years commercially and i find myself asking the same question.
Software projects completion dates blow out because of software “design” methodologies. The majority of It professionals where taught their skills by people with a science and engineering background (left side brain dominates). People who manage IT professions similarly have to understand those people, so they need a similar backgrounds… sounds reasonable right ?
Problem is that although science and engineering principles are fine when your working on the fine details, they dont scale very well, its hard to see the big picture when your looking at so many little pieces. To be a good manager of anything you need to be a people person more than anything else, you need to see the big picture rather than all the pieces, and you need to be creative. These are skills that use right side of the brain, the opposite to those that make someone a good programmer.
So IT workers have trouble estimating the time it takes them to do anything because they can only estimate the logic component of their work, not the design side of it, how do you estimate how long it takes to be creative ?
IT Managers arent any help with the design process because they have to absorb all the msitakes in estimates people under them make, plus they make bigger mistakes with project planning.
To put it simply, “design is not a process, its a style”, and IT projects are designed in a procedural manner.
But having said all that, i still dont understand where all the money goes, there must be a lot of fatcat advisers and consultants.
that is nothing to do with the cost of developing these projects
Glen,
You’ve accurately stated most of the problems with IT projects and its Project Management.
These have all been researched and solved many times over.
At least once in each decade from 1970 onwards.
Agile, XP and “Design Thinking” are just the latest go-round.
The problem isn’t the finding solutions, but remembering them and getting them used by this generation and the next.
This is a failure of IT management, the Profession/Professional Bodies and Education.
It is also a gross failure of line management – allowing IT projects to keep stuffing up.
Everyone is allowed a first mistake, but after the first big or small IT disaster, an engaged and competent manager would ask two questions:
– why won’t it be different _this_ time?
– what have you changed so that it _will_ be different?
Line-management and senior mgrs love to assign blame and berate IT as incompetent and worse.
But they are the people who’ve abrogated their responsibility and not just allowed, but encouraged, on-going IT disasters.
Ask the parent of any 3-yo: what matters most, what you _say_ or what you _do_?
At some point, management have to take responsibility for their lack of oversight and process improvement.
Are they inventing NEW ways to stuff up?
Is current technology not able to meet the specs?
If neither, then:
– where are the Audit Office investigations, like Bureau of Air Safety investigations?
– where are the *penalties* for individs, mgrs and companies for repeating Known Errors, Faults, Failures?
– where are the consequences for Public Servants, PM’s and line Mgrs, for allowing stuff-ups?
After 60+ years of IT foul-ups, why aren’t those responsible for the “Checks and Balances” _preventing_ them or ensuring “It won’t happen again”???
Not just an IT and management failure, but systemic failure of auditors, investigators, political process and compliance enforcers:
FAILS Govt own rules for bureaucrats – “efficient, effective, ethical” use of Govt money and resource.
Software and Systems are “hard”.
But that’s NO reason to repeat or allow stuff-ups that are predictable and preventable.
Can good software/systems be written: Ask NASA, ask Airbus & Boeing.
You can *bet your life* it’s not just possible, but practicable.
Where is the outrage _and_ action (investigation, censure, retraining) by the Professional Bodies responsible and by those responsible for training Computing, Software and IT professionals?
Could you imagine _any_ of the Engineering Disciplines NOT reacting to on-going Professional Failures of this magnitude and importance? I can’t.
Can you imagine _any_ Government Agency running without IT systems these days?
They are _critically_ dependent on their IT systems, running and being correct, for daily operations.
What bit of “we can allow IT failures” is Professional or Ethical?
@Glen
Not fatcat, demonstrably _incompetent_ and negligent.
Here are the basic management and project questions:
– If we are doing this project “for a business benefit” in the same way we invest in marketing, how will we measure its _specific_ business benefits over the _full_ life of the project?
– How hard is the project, how long will it take _our_ people in _our_ environment to do it?
– this presumes a database of local IT projects completed, estimates and actual.
– and a database of the _productivity_ of each individual, team and Project Manager.
– What are the specific capabilities, both task area and degree-of-difficulty, of each IT worker that will be involved in the project?
– Do we have critical shortfalls on particular task areas or in high-degree-of-difficulty tasks?
(Do we have the folks we need to do the hardest bits properly?)
– Have we done any projects before with similar or overlapping specs?
– what do we need to repeat prior SUCCESSes?
– what do we need to do to avoid previous FAILURES?
These seem to me these are basic and necessary questions that should be answerable in any business whose operations are dependant on its IT & Systems and have been doing the work for more than a decade.
guys guys guys… these projects work..you just ant looking in the right direction – it works from the perspective of those that milk it and receive revenues (often the large global advisor firms and vendors).. That there is no actual outcome is an irrelevance…
LOL! absolutely right. It’s not about the client or outcomes…
Big Project Firms, redefining “working” since 1970 :-)
No no, the project worked well, the government just was unable to communicate its message through a hostile media.
Excuse me (The Age), but how is this ‘another example of a bungled Labor IT project’? I’ve worked in government and there is no ‘Liberal/Labor’ IT. It’s a bungled IT project full stop, a dime a dozen in government, private industry, or wherever. Thirty years in the industry and I can guarantee you no ‘politician’ made this mess. Lousy project management, lousy governance. lousy change management, lousy benefits realisation, take your pick, just don’t be lazy and blame it on a brand.
Remember, Fairfax now has Gina as a major shareholder, and editors need to keep here happy. :)
A speculative view at best. She rang the editor and advised him directly to do a story on another failed government IT project? I’d suggest it is in the public interest to do the story regardless of whatever opinion Gina held.
>>>>>>>>>>>>>
Until now, large-scale software failures have been most associated with the public sector. In 2007, a European Services Strategies Unit report identified 105 UK public sector contracts with substantial cost overruns.
However, Professor Flyvbjerg believes these problems apply equally to private firms.
“People always thought that the public sector was doing worse in IT than private companies – our findings suggest they’re just as bad.
“We think government IT contracts get more attention, whereas the private sector can hide its details,” he said.
See: http://www.bbc.co.uk/news/technology-14677143
Confirm.
NO differences between Public & Private sector.
Same Consultants,
same Firms,
same Project Managers
same technical people.
‘Same’ doesn’t mean ‘like’, it means ‘identical’ as in the same individuals cycle through all the big firms and work for public/private.
That could be why the rate of failure is high – because the industry’s and staff capability in managing complex IT projects just isn’t here in Australia, even amongst private vendors and consultants. From my own experience, not all private consultants to government know what they’re doing. They say they do, but they don’t, and departments unfortunate enough to engage them get ‘blackswanned’.
I’ve travelled enough and set out to meet IT practitioners from all over.
My evidence is this failure of IT as a Profession is the norm across _all_ the developed world.
Simple proof:
– Y2K had universal impact.
– as does malware.
Everyone is bad. I got walked from my job for pointing out that it cost my team around $395,000 to complete a project that got sold for $75,000, because it made the MD look like the incompetent little man that he was.
Apparently after I was walked, the next project, which I had assiduously avoided, much to my management’s disgust, cost us around $2m to deliver and had been sold for $150,000. I was working for one of the world’s largest finance companies at the time, and these were projects being delivered to household names.
To this day, I can never contemplate working ever again for a company with more than a dozen people.
S Kumar
Whilst ackowledging that there have been some disasterous failures in government IT projects, don’t assume that everything is squeaky clean over in the corporate world.
In the only study of its kind that I am aware of, the Standish Group in 1995 published key findings from a survey of IT executive managers from large, meduim and small companies across major industry sectors as to why IT projects fail. The total sample size was 365 rspondents and represented 8,380 applications. The findings of the report were as follows:
– Overall the success rate for IT projects was only 16.2%
– The number of projects that didn’t achieve business objectives within time and cost estimates was 52.7%
– The number of projects that were cancelled altogether was 31.1%
I would contend that failure of IT projects is not particular to government, only more visible. IT project delivery generally is very poor and I would contend that this is a sad indictment on our industry more generally.
Agree.
Did Westpac get roasted for throwing away $500MM-$1000MM on “CS90”?
Do the shareholders _know_ anything changed?
Did any of the people responsible suffer consequences?
Is CS90 used as a local example in Uni courses on Project Management?
*No* to all…
On Standish and their CHAOS report.
– created a lot of interest and controversy when it came out
-while they skipped one early year, Jim Johnson and his team have continued to release new survey data every year since, allowing change tracking.
– despite the apparently calamitous state of the Industry _nobody_ has replicated the series
– even so, the veracity & methodology of CHAOS has been roundly criticised and attacked in the literature.
What’s wrong with this picture? everything…
Those who say “it couldn’t be _that_ bad” are either uninformed or disingenuous.
The thing is, there is credible, tested research in IT, going back to the early _1970’s_ that says, not only is it _less_ work to “Do it Once, Do it Right”, it’s also quicker and cheaper.
I cannot understand why the _best_ work isn’t picked up, while the industry still endlessly chases the newest fad & fashion – with unproven or questionable results. It’s upside-down.
Gasp!!! The only way I can see to improve this situation is to implement these sort of systems as software-as-a-service. So, in retrospect (a waving a magic wand), treat one TAFE (a representative one) as the lead and implement the system there based on the best available SaaS solution building blocks in a short time-frame (6 months max and capped budget) – making pragmatic trade-offs between requirements, time frames and cost. Appoint a senior business exec to drive the project and to enforce a focus on business outcomes and a pragmatic approach. DO NOT CUSTOMISE THE SaaS SOLUTION COMPONENTS. Iterate the solution until reasonably stable for a few months.
Once the first TAFE is completed to a satisfactory working beta then the solution can be adopted by the other TAFEs in parallel as a SaaS solution shared service WITHOUT CUSTOMISATION. Each TAFE would need to take a pragmatic approach to changing their processes etc. to align with the shared SaaS solution … or to become intelligent consumers of the shared solution by waiting patiently for iterative functional enhancements of the shared solution to address their individual needs. The TAFEs that are pragmatic and business outcomes focused will get a reasonable working system pretty quickly and at low cost … the ones that aren’t will still be debating and procrastinating years later and will have nothing. Executive contracts will be renewed, or not, accordingly.
The result would be a huge step forward in several ways. (1) Time and budget boxing the development of the initial solution and focusing it on one TAFE would at least simplify things and deliver the core of a working solution as a starting point (“cloudy is as cloudy does”). (2) Making it based on SaaS solution components would enforce a genuine and sustainable shared service approach. (3) Delivering a functioning solution for the other TAFEs to adopt would get the project off on the right foot in terms of changing the dynamic from insisting on customised requirements to instead finding ways to make a shared standardised solution work. At least they have something to work with and it is a ‘real’ discussion about a ‘real’ here-and-now system not a theoretical discussion about a set of requirements and vested interests in a document. (4) Allowing the TAFEs to implement the SaaS solution in parallel creates a competitive dynamic which will add business focus and make visible those institutions that are competently managed … and those that are not.
WE NEED TO STOP TRYING TO BOIL THE OCEAN WITH THESE PROJECTS (i.e. implement a gold plated solution that meets the one-size-fits-all needs of all TAFEs) and instead focus on the speedy implementation of a ‘good enough’ shared solution for a typical TAFE which is then widely adopted by the others and improved over time as an iteratively enhanced shared service. This is the cloud services model … and it can and does work.
Its just a question of mindset. Agile thinking + cloud services platforms = innovation and productivity. QED.
hear, hear!
Your suggestion/insight, “break it down into small chunks, delivered quickly”
is _exactly_ what Standish now recommend after 15 years of on-going study and research.
Fail early, fail cheap, regroup and improve – proven in Open Source.
But Bureaucrats don’t like the word “fail”. They perceive it as a negative and “a career limiting move”.
“Fail early, fail cheap, regroup and improve”
Now that sounds like a Lean StartUp phrase!!
As you say the big difference is that Lean StartUps call it “validated learning”, because they are actively measuring the business impact of every iteration & change they make to ensure it drives them closer to their focused business outcome. It isn’t failure if it is done fast, early and cheaply! You are actually saving money (and time/ resources/ stakeholder relationships/ careers!!)
And I think what Steve is referring to above is called building our your ‘Minimum Viable Product’, then iterating it using validated learning….
http://en.wikipedia.org/wiki/Validated_learning
Tom,
You’re exactly right, but a don’t show good knowledge of historical efforts.
This stuff is a) well-known and b) is shown to work and c) its always been _easier_ and _cheaper_ to do it right.
Von Neuman got it right and created a high-performance, high-quality culture, but this was lost. Why?
Your current understanding and language are _correct_ and I applaud you & your teams if you use them.
But they are restatement and rediscovery of previous work going back at least 4 decades.
My on-going points are twofold:
– since early 1970’s how Software can be done “better/cheaper/faster” has been studied/researched and good techniques discovered and widely disseminated,
– BUT the known, effective approaches are not widely adopted nor do known “best practices” get taken up AND made stick.
What’s going on is NOT a Profession in action, only a disorganised “Industry”.
And the deeper questions:
– Since the 1968 NATO Conference on Software Engineering, is there _any_ reason for Universities NOT to turn out graduates who aren’t trained and experienced in _known Best Practices_?
– For the last 40 years, is there _any_ reason that Management has allowed I.T. project to fail from Known Causes? It’s OK for Professionals to make mistakes, but _never_ to repeat them.
– As Government and Large Business have become increasingly dependent on Computing & ICT, is there _any_ reason that their “Checks and Balances” processes & groups _haven’t_ starting collecting basic project & outcome data or started investigating failures and imposing “consequences” on individuals and organisations repeating, or allowing, known errors, faults and failures?
I appreciate your Professionalism, insight and understanding and that you’ve dedicated yourself to Software Improvement – you should be applauded & recognised for that.
I trying to draw attention to something much larger that’s going on and that it’s been noticed and _documented_ as a massive problem for around 50 years…
As Larry Constantine observed in 1994:
There is NO Software Crisis. Crises are time-limited, this has been going for _decades_, it’s a slow train-wreck.
There is a systemic problem that keeps recreating the same poor cultures, same worst-possible management and same predicable, preventable and unavoidable failures.
This has all been said before:
so what aren’t the people that can _fix_ things, the business owners and govt ministers, taking notice?
The happy marriage of Governments and enterprise IT? http://bit.ly/12sdYbq
In recent weeks, it has struck me that things have become a bit more transparent in the intersection between governments and enterprise IT…
On the one hand, earlier in June and in Queensland’s State Parliament, the recently tabled paper (Paper #2867) detailing the Response to recommendations to the ICT Audit was immediately followed by Paper #2868 which deals with ” …. dog owners are held accountable for their animal’s actions and not able to move their animals to another area to escape punishment”, the irony was not lost on me.
Were the parliamentary papers inadvertently swapped?
These mega projects are merely a feeding frenzy for the major vendors and consulting organisations. Systemic (not technical) risks are often invisible in silo’d organisations stuck with conventional governance and risk management models. Unfortunately, as taxpayers and service recipients (directly or indirectly) we all are the losers in situations such as these.
MY glimpse in to the mad house that is government-driven I.T.: Applied for a 4-month contract with a Queensland education-related department; final interview happened 6 MONTHS after my first one, after which – nothing. Eventually found out it had been scrapped.
TL;DR: A 4 month contract position, advertised widely, was not filled 6 months after it was first advertised – and eventually the position was scrapped.
Poorly-trained, politically appointed project managers lurch through the government system, haphazardly bloating scope and rolling back any progress their predecessors may have made; and as there is little continuity in the personnel behind these projects, there’s no core vision and few repercussions for failing to meet goals; those at the helm can just point to the mess their predecessor left, and they’ll be right – as will be those that follow them, pointing back at their work.
I can only surmise that software development is so opaque to most people, the horror of the situation does not raise a public outcry – if a government-controlled project building, say, a toilet block in a park, cost $100 million and was plagued by construction problems and delayed deadlines, people would be aghast.
Do a search for “BSR” in a government owned business. 1yr and $50mil becomes 8 years and $500mil.
Comments are closed.