I hope there is a plan b for this.
The Department for Work and Pensions has replaced the project director of Universal Credit, its £500m web-based benefits system, days after denying it's suffering an IT management crisis. Project director Hilary Reynolds will be moved sideways and her role filled by David Pitchford, Universal Credit's chief executive. …
Monday 11th March 2013 10:39 GMT IHateWearingATie
SRO not useful, but a programme director is...
In my experience, the competence of the SRO doesn't matter in the public sector (and is generally some poor DG press-ganged into the role) but the programme director is absolutely crucial.
Its not looking good... and this is coming from someone who is usually defending gov IT development on El Reg.
Monday 11th March 2013 12:37 GMT Anonymous Coward
Re: SRO not useful, but a programme director is...
The SRO is useful, really, honestly, truely.
I have rather a lot of experience of biggish public sector projects (circa £50M to £300M) over the last 10 years. Most haven't made the news as they have actually delivered more or less on time and more or less on budget. I know this is heresy, but some govt IT projects do work and deliver.
The projects that work have a good programme directory AND a good SRO. The programme director works the day to day stuff and does what a good PM does, the SRO works the politics and the organisation. Both of these roles are needed to successfully deliver. I have had good and bad SRO's and PM's, where they are both good the projects tend to do quite well. The SRO keeps the organisation on track by stopping the knee jerk reactions and holds the organisational ship steady, the Programme director knocks the stupid additional requirements dreamt up back into touch.
None of this is rocket science and there are a number of studies produced by govt that shows the value of a good SRO and good and appropriate governance.
The problem we see (and I post anonymously as I am still on a govt project) is where SRO's chop and change. They come in, they fiddle, they leave and are replaced. These are the projects that fail or have huge issues. Where the SRO is involved and there is a good programme director, the project has a reasonable chance of success. I am lucky that my current project has a good SRO and we have a good working relationship. This makes it easier for both parties to ride out the inevitable storms that happen on all projects.
Nothing is guaranteed in any project, I've seen major private sector projects fail as well but a good SRO will make a big positive difference.
Monday 11th March 2013 18:49 GMT Alan Brown
Re: SRO not useful, but a programme director is...
"I know this is heresy, but some govt IT projects do work and deliver."
The ones which don't are usually doomed from the start by woolly objectives + political interference, etc. You tend to find the same faces showing up over and over again in them too.
Monday 11th March 2013 10:57 GMT Anonymous Coward
The problem is there is no Govt IT its all outsourced by people who know nothing to massive inflexible Companies who take them for a ride. What they need to do is spend some of that £500m renting space in a decent DC and employing their own full time Dev team to write the code and be able to adapt quickly to any changes.
Monday 11th March 2013 11:43 GMT Shasta McNasty
You're over-simplfying things
For a start, you'd need more than one DC with a DR/failover plan in place and sufficient connectivity between the two. You also have to think about little issues like segregation of dev/test/prod networks, storage, hardware, backups, database design & management etc.
You also need to couple the above with vague and ever-changing knee-jerk requirements that are influenced from outside the organisation e.g. the media. The data can be non-sensitive, sensitive, locally sensitive or nationally sensitive which can mean it can even be stored on the same disks.
I admit the current set-up is far from ideal, but there isn't a quick-fix from a single dev team.
Monday 11th March 2013 12:37 GMT Anonymous Coward
Re: You're over-simplfying things
No but none of that means they cant or should not do it in house, I would never suggest they use just 1 DC thats just insane but when your talking about £500 million all thats just a given. For example we have 20 odd Devs and 10 testers and a dedicated 24/7/365 support team and our turnover is not even £100m let alone £500m.
And knee jerk requirements is exactly why they need an in house team who can be more adaptable, the problem with these outsourced contracts is each change usually requires either a contract change which is expensive or a costly change request. The price on paper might look OK but it rarely is.
Monday 11th March 2013 12:52 GMT Anonymous Coward
The problem is that your comment is total crap and bollocks written by somebody who hasn't the slightest ideas what he or she is talking about.
I know the project the article is referring to, its a big difficult project, that has lots of complexity. Its complicated because it tries to do a lot, its complicated as it has lots of security (do you want your personal details leaking out?), its complicated because there's lots of companies working on it as the govt splits work around. is this right or wrong, no idea, but I'd be wary of a single company trying to do everything, remember the train wreck that was the Connecting for Health debacle.
Renting space in a DC is meaningless, what fucking difference will it make renting space in a DC. Having their own full time Dev team is also meaningless, this is big, this isn't some crappy PHP project you're throwing together on Apache, you need a lot of people with a wide range of skills from infrastructure (WAN's, LAN's, OS's), security, configuration and management, development across multiple platforms, testing (and there's a lot of that), program management, finance, legal, PMO. The only people who have this range of people on tap are big Systems Integrators. No govt IT dept could afford to keep these people sat around for that once every ten years development this program is.
Talking about adapting to change quickly also shows you for the muppet you are. One of the reasons that large govt projects fail is that they try to adapt to changes at a late stage. On a project this size, you need a proper set of requirements, requirements you can trace as if you get this wrong, some people will lose their benefits and will physically suffer. Throwing in a last minute change could be a major disaster.
Big system development is hard, it's hard for the govt, its hard for the big boys, its hard for the small SME's. As you have so ably demonstrated it's actually hard for people to even understand, so that they avoid writing the rubbish you have just spouted.
Monday 11th March 2013 13:16 GMT Anonymous Coward
Oh calm down boy! No need for such hostility I haven't insulted your mother or something! I could answer all your 'points' but as I find the language you use so childish and insulting I am not going too. This is a public forum where all people can add their opinions just because you don't agree with a comment does not mean its wrong (or right for that matter).
This post has been deleted by its author
Monday 11th March 2013 21:16 GMT Camilla Smythe
Re: Crap.. ODFO
"I know the project the article is referring to, its a big difficult project, that has lots of complexity. Its complicated because it tries to do a lot, its complicated as it has lots of security (do you want your personal details leaking out?),"
... "other Blah".
If this is part of it,
Then you go sign up and have all your data given to Google and Webtrends. Want to use it with No-Script enabled...?
No Dole Money or other Services for you.... You will need a 'Gubberment Gateway ID'. Identity card via the back door.
Gubberment argues elsewhere that Google Analytics is 'free'. Go check out some .gov.uk and other probably associated domains where Gubberment Money is spent. Now remind yourself of the adage...
'If it is free then you are the product'.
Monday 11th March 2013 16:40 GMT Anonymous Coward
never going to happen. The outsourcing route is a very nice, straightforward way to ensure no individuals can be blamed for over-runs (whether costs or time), only the large outsourcing companies - of whom there are a limited number, and an over-run on one project for one department has no consequences whatsoever for the outsource company's bids for other public sector work. Now if you take the view that "if you over-run on this project in terms of cost, you have to swallow the loss, and if you over-run on time, you aren't allowed to even bid for anything else until you finish", and couple that with publicly named individual employees in each department having contractual and financial liability for over-runs if they were predictable (and let's face it, they usually are), then maybe things might change.
Tuesday 12th March 2013 09:28 GMT Anonymous Coward
The problem is John Major and his successors of all colours........
.........who outsourced Government IT in the first place. They sent the skills to the private sector and backfilled the civil service with hunderds of overpromoted policeman that were and remain IT illiterate, have no responsibility for delivery, don't accept risk, don't make decisions and expect immediate reaction to change. I'd warrant it costs more to run now in real terms numbers of people than it did when HMRC computing was privatised in 1984.
Add to that the ridiculously fragmented contracts that, in an effort to create an artificial market where there was no need for for one, split responsibility for Dev, Hosting & Support across multiple suppliers with nice big gaps in between that they then expect suppliers to cover because they can't manage their own contract properly.
There is JUST one common factor in failed Government IT. GOVERNMENT.
Monday 11th March 2013 11:57 GMT Magister
I would argue that although there is an element of IT in this project, it is not an IT Project; it should instead be viewed as a Business Change Project and approached accordingly.
The requirements for the hardware, software etc. are not simple, but by far the biggest issue will be getting the buyin from senior managers and the end users to required changes within their processes; without this, the project is doomed to failure no matter how good the design of the systems. It's not like rolling out a minor upgrade to a piece of software, but rather like a oil tanker changing course.
This is not a new concept, nor is it confined to the public sector.
Tuesday 12th March 2013 11:49 GMT Jamie Jones