DO IT projects ever work?

[Gus]

Was watching the Gerry Robinson & NHS programme last night, one year on from his initial visit. At one point he made a very telling comment. He referred to the NHS IT programme … its swallowed about £12 BILLION, enough for thousands of nurses for 10 years …and has delivered NOTHING, nadda, zilch … yet no one seems to have been hung, drawn and quartered over it. Anybody want to justify the wages of the hordes of IT against the wages of your average nurse?

Isn’t that a damming indictment of the IT profession. If you consider all the work over the last 30 years in terms of Prince 2 qualification, ISEB training, code generation software etc etc etc … yet I’ve yet to hear of a successful major IT project … well at least one that the end users consider to be successful. I’ve been working in/alongside IT for the last 15 years and have only see two minor IT projects meet original expectations.


So … given the hordes of IT professionals who inhabit this Forum …. Anyone want to refute my scurrilous allegation?

[Paul F]

As someone who worked in not one but two areas for the NHS IT national project (NE Cluster) I guess I could have a little input.

Pile of rubbish.

As usual with any large scale IT project it was a case of Chinese Whisper syndrome through middle and top-end management along with plain and simple gross incompetence.

The work that was going on at the 'real-world' level e.g. consultations, use-case developments, prototyping, end-user trials and so on was really good. I mean that. Some of the technologies and solution visions that I was privy to were increadibly useful. For example, paperless, content managed Primary Care surgeries and fully-integrated and networked Emergency Response vehicles were looking great when we rolled off. The feedback from the GP's and GP's assistants (as well as paramedics) was fantastic. The individual point-solutions worked and the mood was promising. The key thing here is that the mood was promising not just among the big-wigs but among the people that would actually be using the systems.

HOWEVER!!!!

As always the process in place to knit all of these point-solutions and strategies together just wasn't there. Now of course I dont know why as I wasnt involved in the larger picture as Im just a lowly Java Developer.

It came down to the fact that the planning just wasnt there on a larger scale. I remember that we pushed like hell to get an implementation in place only to be told that it was on hold pending not one but three other projects, most of which were dependant on other projects. Typical scenario.

Personally, I didnt think it would necessarily fail but the idea of splitting the UK up into what were termed 'Clusters' was never a logical idea. You have 5 or so companies, each conjuring up their own vision for their area. Bad news.


On the positive side, I moved away from the monolithic company to a considerably smaller one in Guildford. Im pleased to say that every single project, although not comparable in scale, has run more or less to time and budget and every single client can be referenced as a success.

[Gus]

Hey ... good response. I suppose, if I wasn't trying to stir up some response, the real question is "why do IT projects fail" and NOT "why are IT professionals so cr@p yet get paid obscene amounts of cash".

I think that I am looking forward to the Commission that will take place to ascertain why the NHS project failed so badly. A similar investigation into the London Ambulance Service debacle produced a report that is considered ti be compulsory reading for modern day IT project mangers ... shame we seem to make the same mistakes again and again.

[straycat]

If you're interested in this question, there's a superb book called The Mythical Man Month, which is required reading - it's an exploration into that exact question, by a man who's been involved in a number of major IT projects, both successful and unsuccessful.

[ducasi]

This is one of the reasons for many failures. (And one of the few times I agree with Bill Gates!)

[TurboTomato]

I can't speak for all IT projects, as I don't work in infrastructure or anything like that (I am a mere software engineer), but you are correct in some of your statement - IT projects are notorious for coming in late, and over budget.

IME you can't really lay the blame at any one persons door, it can come from any number of areas, for example:

Customer requirements not being correctly specified at the initial conception - this leads to scope creep and will inevitably delay the project.
Sales people promising features and suchlike to customers for a certain timeframe without referring to those who are going to develop it to confirm the timeframe.
Project managers not putting enough contingency into projects.
Software engineers and developers underestimating estimates.
Incompetent engineers/developers - like any other industry we have people that are crap at their jobs.
Not enough time assigned to QA and testing.

[Barry Shnikov]

In the projects that I have had knowledge of (always indirect), what I've learned suggests that the basic problem is mission creep.

It comes in to manage accounts; then order processing want it, then sales processing want it, then managment want it, then development...

Not knowing anything about programming in detail, it's always seemed to me that the method most likely to be effective is to have a number of sub-systems that will talk to each other - a true modular system. But what is requested and/or what the IT consultants recommend, is an all-singing, all-dancing all-encompassing grand super-system that will do everything anybody wants it to do.

It's also an inherent difficulty with a capitalist economic system. It isn't in the interest of any IT provider to recommend systems that are easily interchangeable with other people's products, because then they might have to share the business with somebody else. Just a thought.

[Barry Shnikov]

And of course it was exactly this sort of problem that led to IT people scoffing at Reagan's Star wars plans. It would have been the longest programme in history - 4 million lines of code, IIRC? - and it would have had to work perfectly on the first and only occasion on which it was required to do so, with limited possiblity of testing it.

People I knew in IT at the time were stunned that it could be taken as a serious proposal.

[David Bailey]

It's also an inherent difficulty with a capitalist economic system. It isn't in the interest of any IT provider to recommend systems that are easily interchangeable with other people's products, because then they might have to share the business with somebody else. Just a thought.

Ummm, the Soviet planning system didn't work too well either in terms of efficiency. As I understand it, the inherent contradictions of capitalism make things - overall - more efficient, than trying to plan things centrally.

As for IT projects, the reason they go wrong is the same reason most big projects go wrong - incompetence at the top. The national ID system is the obvious looming example of this as a disaster in the making, but this is not just limited to IT projects.

Look at the Scottish Parliament building - three years late and 4,000% more expensive than initially planned. Look at the British Library - 5 years late, 300% over budget. Look at Wembley Stadium. Look at the Dome. etc. etc.

One of the few truly impressive achievements from Ken Livingstone's mayoralty has been the building of City Hall in London for only £43m - on time and on budget.

But at least big buildings are visible, they last, and they can be valued and cherished by future generations. I can't see that future generations will be gawking over the beauty of the code for the NHS system, somehow...

[bigdjiver]

If you're interested in this question, there's a superb book called The Mythical Man Month, which is required reading - it's an exploration into that exact question, by a man who's been involved in a number of major IT projects, both successful and unsuccessful.

His major insight is that we should not expect to get it right first time, especially when "it" is huge, complex and has never been done before. We have to be quick to realise, and ready to admit, error, and, if necessary, start again, The key phrase is "smarting, but smarter.".

It is not just huge IT projects that fail, it is huge projects of any type. In fact, it is little projects too, but they tend not to matter so much. (I have yet to make the Columbian work in freestyle.)

[David Bailey]

(I have yet to make the Columbian work in freestyle.)

Ahhh, now you're talking...

[bigdjiver]

Hey ... good response. I suppose, if I wasn't trying to stir up some response, the real question is "why do IT projects fail" and NOT "why are IT professionals so cr@p yet get paid obscene amounts of cash"...

Business is about people, about seeing opportunities and snapping them up, having a positive attitude, "git there fastest with the mostest". Good enough is good enough. Appearances over actuality. Successful businessmen, managers, and politicians tend to these characteristics.

Writing software is about painstaking analysis, a pessimistic, nit-picking, pedantic, "what can go wrong?" attitude, Great software writers tend to be not so good as managers, and even less good at selling themselves or projects. A big smile and "I can do it" is more likely to get the job than "Well this is going to be a problem, and that is dubious, and ..."

Many projects fall into the chasm that separates the personality types.

[stewart38]

Isn’t that a damming indictment of the IT profession. If you consider all the work over the last 30 years in terms of Prince 2 qualification, ISEB training, code generation software etc etc etc … yet I’ve yet to hear of a successful major IT project … well at least one that the end users consider to be successful. I’ve been working in/alongside IT for the last 15 years and have only see two minor IT projects meet original expectations.


So … given the hordes of IT professionals who inhabit this Forum …. Anyone want to refute my scurrilous allegation?

IT evolves like clothes in my wardrobe

I’ve undertaken a number of projects to update it and often with a flash item (that im subsequently told went out of date years ago and doesn’t go with anything I have) .

Those who tell me to update or insist I update to look ‘cool’, are often fly by night people with big ideas but they don’t usually hang around and are soon of updating someone else’s wardrobe (ie ex’s)

However my wardrobe is functional and I get by but….

Look at it year to and you wouldn’t see any difference but 30yrs ago and you would

There is a small chance instead of a 4hr daily commute I might work from home

Ill have PC and internet access

E-Mail of course and all the mod cons

I see a day when 50% of people will work from home so maybe some IT money is wasted but the in the long run I think it has helped (The Forum being a good example)

[robd]

Some good responses to some typical nonsense from Gus.

No doubt many major projects come in late and over budget - in IT & many other sectors.

I think an issue for IT & the supposed failure of projects in the sector is that technological change needs to be accompanied by change in the culture, workflow, etc of the organisation and this may not be recognised (or recognised too late)

Additionally there is the issue that many IT projects need to build upon previous systems rather than starting afresh and this can compromise design ideals especially in an industry which is still nothing like mature compared to many other sectors.

[Gus]

Some good responses to some typical nonsense from Gus.

Thank you for that excellent contribution to taking the debate further forward . Clear to see how refined your analytical and social skills are. Any chance of you explaining how you come to this conclusion?

I think an issue for IT & the supposed failure of projects in the sector is that technological change needs to be accompanied by change in the culture, workflow, etc of the organisation and this may not be recognised (or recognised too late) .

Yup … change projects often focus on the IT element in isolation. The text book mantra is that change is based on People, Process and Systems. To ignore any one of these is to set yourself up to fail. However, that’s exactly what most projects do .. maybe down to a shortage of good process analysts and experienced HR change managers.

[bigdjiver]

... I think an issue for IT & the supposed failure of projects in the sector is that technological change needs to be accompanied by change in the culture, workflow, etc of the organisation and this may not be recognised (or recognised too late)

Additionally there is the issue that many IT projects need to build upon previous systems rather than starting afresh and this can compromise design ideals especially in an industry which is still nothing like mature compared to many other sectors.

In my era I.T. was often brought in to shore up or replace adminstrative systems that were failing. All that happened was that the systems failed faster, and the computer, not built with duck and dodge abilities, got the blame.
The system I am dealing with now has cracks in its foundation stone, but the CEO laid it himself and will not admit it. The most important information for the companies future is simply not being recorded properly, and lost and corrupted, and I am out of favour for saying so.
When the articles posted by Ducasi say that communication skills are highly valued by industry, industry is not wrong. What I need is the ability to communicate the need for change and make it happen, and I do not have it. In the IT sector I am not alone.

[geoff332]

My perspective is always unique - mostly because my academic background is in organisational behaviour and theory (ie people) and my professional background is in IT.

We can talk about IT projects in general - but also projects in the NHS...

IT Projects in general tend to go over budget and time (if not fail entirely) for the simple reason that you're typically trying to do something new. Either a new way of doing something you've always done or doing something you've never done before.

From an IT perspective, a project needs to be tightly specified before it can be properly planned. And to tightly specify it, one needs to know exactly what it will look like when finished. When you're doing something new, that's
quite tricky (actually, it's largely impossible). The common complaints from IT people - poor specifications, scope creep - and so forth are the signs that no-one knows what the final product will look like.

When planning a project, one typically builds in a fair bit of fat into the proposal - to cover all the uncertainty. However, in most projects, this gets chewed up very quickly by figuring out what you're doing. Having poor functional specifications makes testing almost impossible to specify and so forth.

From an organisational perspective, the problem is being asked to describe systems and processes that have never been done before. This is, to put it bluntly, impossible. As you start seeing the deliverables, you start to get a sense of what's possible. And from that, you are finally able to clearly envision your deliverables - and start refining and improving them. This is a really common process - as our actions and interactions incrementally build knowledge and experience, the thinking behind our actions gets changed. It's a form of learning.

At the same time, organisations are not monolithic. They're made up of people who tend to form various clusters of skills, experience and power. These clusters are not going to want the same thing. Even in "good" organisations, there will be differences of opinion. In more dysfunctional places, this will devolve into overt conflict. This leads to games being played with the project. Sometimes, those games are relatively benign and geared towards producing the best outcome' other times, they will be struggles to the death. However they manifest themselves, they will effect any IT project. This is even more true if the project is touching on a sensitive area of the organisation.

IT Projects usually provide technical solutions. However, technical solutions typically have both cultural and political ramifications as well. These facets tend to be seriously under-managed, resulting in big problems. IT Change management is usually a horribly technical process, and completely fails to account for how the human behaviour and associated values and power will shift as a result of the technical change.

The NHS itself is a very large, messy and conservative organisation. It contains a number of major power-bases and is always subject to political intervention. The process of funding anything large in the NHS is just ugly - time consuming and painfully bureaucratic. Internally, it's highly politicised and has a strong - albeit not always functionally useful - culture. Trying to introduce a organisation-wide change in systems is almost certain to result in failure. This is a consequence of the organisation, as much as of the IT people who try to do work in there.

There are exceptions to these rules, but they are rare and, when one looks at them closely, they are almost always a result of either someone addressing the issues I outlined earlier or the project/organisation being a atypical enough that the issues aren't relevant.

I could, naturally enough, rant on for hours... but that's enough for now.

[bigdjiver]

... a project needs to be tightly specified before it can be properly planned...

I could, naturally enough, rant on for hours...

Actually you could rant on for the rest of your life on this subject.

It is impossible to specify most large projects in the N.H.S. because the bottom line is the value of human life, health and happiness. Everybody has a vested interest, and rarely is it a logical or consistent. Tomorrow it is your child or your mother, or even you.

Specify now for implementation in 3 years time, but the technology will be transformed in those three years. IT cannot help but deliver at least partly obsolescent systems. Where will DNA technology be in three years time? Micro-machines ? Robotics? The Internet?

[Gus]

Specify now for implementation in 3 years time, but the technology will be transformed in those three years. IT cannot help but deliver at least partly obsolescent systems. Where will DNA technology be in three years time? Micro-machines ? Robotics? The Internet?

Hold on, weren't two of the key NEEDS of the NHS at the moment well within current technology .. ie patient record management and bed availability/allocation. We've had that ability for years ... I'm no IT guru but why are these so difficult to deliver. Thats a serious question, I'm not being sarcastic .... I'd really like to understand where the difficulties lie.

[bigdjiver]

Hold on, weren't two of the key NEEDS of the NHS at the moment well within current technology .. ie patient record management and bed availability/allocation. We've had that ability for years ... I'm no IT guru but why are these so difficult to deliver. Thats a serious question, I'm not being sarcastic .... I'd really like to understand where the difficulties lie.

NHS is not my field, but I would think confidentiality and identity are major issues, and so many records go back decades and too many entries are written for the writer to read are others.

A hospital is a very dynamic environment, bed allocation would not be a picnic either. Computer systems implement rules. Would you like to specify rules for bed allocation? Put very hard choices into writing and sign your name to them?