One-quarter of UK.gov IT projects at high risk of failure One-quarter of the UK government’s major IT programmes worth a total lifetime cost of £8bn are at high risk of failure, according to a Register analysis of the major project watchdog's annual report of 143 government programmes worth over £455bn. The Infrastructure Projects Authority report rated 39 ICT projects worth £18.6bn …
<side rant>
Look you silly git, when you have a real comment that no one in their right mind will find offensive, why hide yourself? Or do you just like the Guy Fawkes icon? :-P
</side rant>
25% is a huge number. When you hear it from the government, admitting their failure, we know that they are guessing on the lower side and that the number is much higher.
As an IT professional, I look at some of the projects that failed and I have come to realize that its more than just setting realistic expectations, but that there aren't enough highly skilled developers working out there at all levels.
<rant>
Too many IT people lack the core training that you get from a solid engineering program. Boot camps can teach you to code, but they don't train you in the theory or how to think about solving a problem.
Too many HR types hire in staff who couldn't design or program their way out of a paper bag. Too much focus on agile where its really just a set of iterative waterfall projects and isn't really all that agile because if you need to correct a previous sprint, you've upset their apple cart. RAP/RAD is now becoming a lost art.
Sorry to say, but if the UK gov were to actually spend $$$ and hire top notch guys who have mastered their craft, they would end up spending less overall on the project and come out ahead. Of course this goes back to the real problem... There aren't enough of us to go around and those who are around end up competing with people who cost 1/2 of their rate and lie on their resumes to get the job for which they are not qualified.
</rant>
I apologize for the rant but seriously, this industry is so f-'d up its insane. On both sides of the pond.
Why AC? Because I can.
25% is not a huge number. Show me an organisation where fewer than a quarter of projects are not currently under significant amounts of risk* and I'll show you an organisation that's fudging its numbers.
*This is what an A/R rating actually means. Significant risk of failure. Not a failure. Not behind schedule. Not underperforming. Not over budget. Not *likely* to fail. Significant risk. It's a risk rating.
25% failure is a high number.
At one client... I had 7 of the top 10 key feature enhancements get handed to me. 100% success.
Next project, same client... management claimed success, we all said it was a bust. Then the next 6 project on third team 100% success.
So yeah, 25% is a high number. Especially when you can easily lower it.
For the best part of my 25+ year software development career I have worked for investment banks and commodity trading firms, and by and large, they're OK at getting a decent systems delivery/write-off ratio (although the write-off level is directly correlated to the size of the organisation).
Once, however, I had a year from hell subcontracting through Schlumberger Sema for defra.
Not only did I have to learn how to stretch a day's effort into a week's worth, but also had to live the daily nightmare that is working on a project you know will never fly.
It was just after the Foot and Mouth disease problem circa 2003-ish. The system was to be a cloven-hoofed livestock tracking website. In this instance the problem was caused by a massive up front waterfall style analysis by SS's top £2K p.d. consultants producing a spec that didn't match the requirements (and was seemingly priced per KG of paper too).
Given my experience there I have no reason to believe that any other ministry's development are run better.
The astonishing thing is that I was paid similarly to the cutthroat world of capital markets where reward is meant to correlated to merit.
If you read anything by Taleb you'll know one of his key facets for success is having skin in the game. Government departments typically outsource everything IT to external consultants as an instant arse-covering measure. If, or rather when, shit goes bad they just say "it was those arseholes".
They outsource projects from specification to coding and testing. If they had to spec, supervise and deliver a project using internal resources on time, spec and budget with everyone having skin in the game i.e. bonus for great work, fired for shite, then I have no doubt Government projects would fare much better.
This is a universal problem in Government. I see the same thing here in Australia where I have to put up with some of the country's most unproductive workers constantly outsourcing shit so they can't take the blame when it goes bad or doesn't pass muster. They essentially do fuck all but they sure shout loudly when it comes to "pay and conditions" as their hands go out for the next pay rise. Currently I am producing more useful output that at minimum two of these workers and take about 1/5th of the time to produce a coded solution to an issue/requirement than they do. To think that tax payers are being fleeced like this sickens me and I know that I am situated in one of the more productive entities so it does get worse. Don't even get me started on the number of uncounted sick days I witness.
Sorry, but I am one of those contractors that gets called in to do the heavy lifting. I've worked in telco where 80% of the work was done by contractors/consultants. And this product became the cornerstone of the company. Of course every one of us contractors pushed ourselves to do our best work and we were a band of professionals.
Smaller teams of seasoned experts will always outperform the marching masses... regardless as employee or contractor.
When the contractor believes that he's only as good as his last project... he has skin in the game.
"They outsource projects from specification to coding and testing. If they had to spec, supervise and deliver a project using internal resources on time, spec and budget with everyone having skin in the game i.e. bonus for great work, fired for shite, then I have no doubt Government projects would fare much better."
If they had to deliver projects like that they'd never complete anything because they wouldn't be able to retain developers and testers with civil service salaries subject to a public sector pay cap and high risk of public humiliation for failure to deliver successfully. They are, however, moving away from monolithic SI contracts to a model where government resources project management and some analyst staff but still use contractors and consultants for the majority of analysis, delivery and QA.
Disclaimer: (Anonymous because) I currently work on one of the more ruddy-hued projects mentioned in the report.
Ranting aside, youve make the same mistakes that all naive people doing long running, comples projects.
1. Assuming you are going to get exceptional people.
Whats todays term - super stars?
English football team makes the mistake, time + time gain. Large projects are team work, some will be good, some will be bad (but you can hopefully sack those - maybe not with civil servants), most will be average. Design and plan for the average.
I blame 'Commando' comic - With my last grenade ...
2. Over focusing on the implementation/coding.
Focus on the specification, then focus on change control, then test/verification, then the implementation.
Now apply that 25% across all of government spending and you get a £100Bn+ waste bill. Not a small number, especially when you consider that they are fast running out of inventive ways to f*cking rob, defraud or bludgeon UK tax payers into paying more money.
Regarding DSAB, the contracts only went out for offer in February, so how is it overrunning as it would not even be in the PAO stats yet?
This post has been deleted by its author
"Let's use consumer grade "cloud" Office 365"
I think you will find that Office 365 is WAY more used by businesses than consumers. And it includes a local install of Office too - using the cloud is optional.
"and not LibreOffice as successfully deployed by government agencies across Europe"
Erm so the French police managed to use it on a few systems. Munich was a failure - and wants to reverse course after a decade of trying. Pretty much no one else is using it - because it's very limited in capability in comparison to MS Office - and users hate it.
"reducing both long-term costs"
MS Office has been widely shown to be cheaper than migrating to and trying to use a vastly inferior product like Libre Office. Remember that licensing is only a small part of the TCO.
"and the Microsoft macro attack vector."
Macros are locked down in corporates - and have to be digitally signed to be able to run at all. It's far more an issue with freeware like Libre Office - where a well funded attacker can just peruse the code to find the vulnerability of their choice...
"Its Microsoft fanboy mayor wants to reverse it. Politics. Or something."
Nope, users hate it, and it's widely considered a failure. The report that they got before they voted to develop a Windows 10 platform on back in February confirmed that: "between 18 and 28 per cent of users had "severe issues" related to software"
Users hate it because "it looks different". When you look at their reasons for hating it you can pretty much discount it as typical whiny user bullshit. Hell, move from 7 to 10 and "it looks different". Look at the work that is being done in the NHS (article recently on El Reg) to produce a distribution that can be used across the system that would cut license fees and standardise desktops not to mention reduce hospital shutting malware issues.
Look at the work that is being done in the NHS
AFAICT the NHS Linux distro isn't being done "in" the NHS, in the sense of it being any kind of official project. From what I can see it's a side project by some people who also contract their services to the NHS in other ways. Suppliers, GPs etc.
Pedantic? Maybe. But the comment seems to imply the NHS is doing something special here that other departments could emulate. Whereas the reality seems to be that the "innovation" - yet again, and for want of a better word - is coming from outside the system.
"seems to imply" is your inference and is incorrect. "In the NHS" in this instance was referring to "within the realm of" or "in the area of" and not "by the NHS". The point being that it is a project by whomever targeting a massive user of IT for which it sucks up christ knows how many dollars out of the theatre side of the operation. Anything more malware resilient and reliable lacking the standover capabilities of a vendor such as MS would be an improvement.
"Let's use consumer grade "cloud" Office 365, and not LibreOffice as successfully deployed by government agencies across Europe, reducing both long-term costs and the Microsoft macro attack vector."
What?, use LibreOffice to create something like the Rural Payments Agency application ? Seriously, there are plenty of articles you can go and peddle your productivity suite agenda ( cos users only write documents and send email right?) but this ain't the one - jog on !
Indeed. Sounds like a crap study to me ... When did ANY government, anywhere, at any point in human history, do 75% of anything as advertised? If the Ministry of Untruths and Fake News were to exist, it would be the only department telling the truth most of the time ...
"It will just cost them an extra ~ 30% on top of the outside of IR35 rates."
I don't know. If I were still working I'd be very wary of taking a caught contract even with a hefty premium for doing so, just to avoid the risk of not being able to get HMRC off my back afterwards.
What about all the "top secret" projects that no one is allowed to know about?
*
For example, Philip Hammond gave £1.9 Billion pounds extra to GCHQ last year. How much of this "top secret" development money is at risk? And then there's the basic (huge) base budget at GCHQ.
*
For example, the UK's newest aircraft carrier might qualify for "risk":
- It doesn't have any aircraft (and the software on the proposed F-35 doesn't work)
- It has lots of Windows XP instances on board
*
So maybe the headline should be "25% of the projects we're allowed to know about are at risk"!!!
"Everything is going according to plan. Soon civilizational chaos will be complete. INCREASE OUTSOURCING AND MAKE THE PROJECTS STILL MORE AMBITIOUS AND COMMITEE-DRIVEN!"
"With pleasure, Oh Dark Manglement One!"
"Oh and issue high-adrenaline articles about random IT fads in the press".
(Strokes white fluffy cat)
I feel it is only fair to point out that, IMHO, probably, 1/4 of PRIVATE SECTOR IT projects are also at "high risk of failure." Whichever "shadow government" happens to be in the shadows at any given time is only too happy to point out -any- projects (whether IT or otherwise) that fail to live up to expectations. This only serves to destroy the public's confidence in government. Think about it: The general consensus of the public in almost all Western democracies is that the "government" is inept, lazy, stupid, and wasteful. It is no more true in the public sector than in the private. The private sector simply doesn't get the open, public scrutiny that the public sector gets. Makes me sick.
"One-quarter of the UK government’s major IT programmes worth a total lifetime cost of £8bn are at high risk of failure, according to a Register analysis of the major project watchdog's annual report of 143 government programmes worth over £455bn."
There seems to be some argument over whether 25% is high or not for the proportion at risk of failure, but in terms of actual cost 8 out of 455 billion is unequivocally small; that's not even 2%. So shockingly, it seems as though there must actually be some sensible oversight in there somewhere - the big projects are not at risk, but lots of small ones are. As is often the case, there are diminishing returns; looking after 98% of your money in 75% of projects is worth the effort, doing the same for the remaining 2% of your money spread much more thinly over 25% of projects could easily end up costing more than just writing off the loss if they fail.
I agree with your analysis of the numbers*, and risk/return ratios, and as a general principle it's an aproach more people should take. But I think here the key point is in your final sentence - "costing more than just writing off the loss if they fail". If CDS fails, there'll be goods rotting on the docksides, patients dying in hospitals and factories standing idle. If MTD goes up the spout you'll have a major cashflow problem for government in the short term, followed by significant impacts on "taxpayer morale" for years to come - not to mention the direct impacts on business trying to cope with it. So on this occasion, I think it is a matter of concern that these key "enabling technologies" might fall over.
*There may well be key distinction between "government's major IT programmes" and "143 government programmes" - I've seen people caught out by the NAO press team's approach to drafting press releases before now, but don't have the time to read the full report on this occasion to check. However, reference to "39 IT projects worth £18.6bn" suggests IT projects average £0.5bn approx; the remaining 104 "programmes" are worth £436bn between them - so a risk of comparing apples to oranges if we're not careful.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
