“ no fully costed and resourced plan in place to enable delivery of the program to completion.”
That’s the problem with Agile, you cost a number of cycles not complete delivery against a signed off requirements document….
A UK council had no fully costed and resourced plan in place to deliver a critical Oracle ERP project two years after beginning an SAP-replacement program, one which has seen years of delay with costs set to climb to 15 times the initial budget. According to an independent audit, West Sussex County Council also failed to …
Agile works well where your requirements are ill defined or changing. It allows you to deal with emerging requirements and respond quickly. The down side is that the cost and completion date are unknown, just like the requirements.
Where all the requirements are known in advance ie a signed off spec, then a waterfall approach will deliver the project on time and on budget. However, if there are new or changed requirements, then it all goes pear shaped.
Having worked with both methodologies, it is very rare that *all* requirements are known in advance, even for a 'legacy' system, such as we have here. Given that it's been in there for 20+ years, there will be many 'undocumented' quirks and 'peculiar' behaviours.
FWIW, my brother worked on a project to digitise the Registry of Births, Deaths and Marriages (in Perth, Australia). They spent 2 years documenting the existing (paper) system and writing a *huge* spec. ERP systems are several orders of magnitude more complex.
I was told by someone involved in a similar project, that their first subproject was to improve the way the paper system worked. Then having done this, and got it working, they wrote the spec for the program.
This had two advantages, the developers understood the existing system, and they did not have to improve the process and automate it at the same time which is a recipe for disaster.
Great idea to optimise the existing system.
Bad idea to customise & configure yourself into a future updates blackhole
Worse idea to try to document existing systems and create a whole new system from the ground up.
This is the perfect project for an unscrupulous IT supplier; wandering spec, plenty of public funds, and no deliverables as a result of a joint 'no fault / parting of ways' equals simple transfer of public funds into private accounts. Yes, I'll take that performance bonus as well please....
You're outlining one of the main problems with public sector projects. The PS is abysmal at defining requirements. Well that's maybe incorrect, the are unable to stick to a set of requirements. They'll start off getting the operation guys to defined it and spin up a project. Then some senior manager will want their stamp on it so will make changes so they can claim some credit for the project, repeat that up the food chain a few times, then for a council throw in elected members and for any project that spans an election make that two sets of elected members (all of whom will change their agenda multiple times).
So any large public sector project becomes a pile of unconnected Lego pieces stirred by anyone with enough authority. And now it's a massive turd that no one will want to own other than a consultant at huge expense or a manager with severe Dunning-Kruger that see this as their opportunity for stardom by turning it around. And they'll hit the road as soon as a few self-congratulatory LinkedIn posts gets them their next project to fuck up.
"(no lessons will be learned from this)"
I disagree, I work in IT for a different authority and I have learned not to attempt implimenting Oracle for pretty well anything. It doesn't help THIS situation I admit, but it won't spread here at least. Not so long as I have breath in my body, anyway.
After the number of failed UK Local Authority Oracle implementations, you'd think *somebody* would have actually identified what the root cause was. Perhaps if there was a corporate policy to reclaim sales bonuses on failed projects, a little more upfront honesty might appear...
Odds on, the council only wanted to buy an "off the shelf" product, but also wanted it to work with all of their existing processes and other legacy systems.
Oracle again, pulling the same shit they did for other councils, and laughing all the way to the bank.
This needs to be pulled into the remit of central government. That individual councils can piss away such obscene amounts of money on a bespoke "just for them" system is a national scandal.
And no, I don't buy the "it has to fit with the way we've always done things". If tens of millions are spent with nothing to show for it, how things are done is shit and heads need to roll.
These sorts of stories make my blood boil. Can't fix the fucking potholes but.......
Projects like gov.uk are actually run very well. The difference is the government is directly employing the engineers working on the project. Contractors are used but only as needed to deliver parts of the project, the overall remit remains with the relevant department.
Government projects can be run well, and the private sector can do well too, but when there's no incentive to manage costs, the private sector is not the right way to deliver projects like these.
On my first ERP implementation 30 years ago one of the senior consultants gave me a rule of thumb for costing projects, it will come to about 10% of turnover when the dust settles. since then I have worked on around 20 (Oracle and others) implementations and that rule of thumb has never been far wrong.
So when someone says they can implement ERP for an £800 turnover organisation for £2.4 million that is clearly rubbish. I would also guess that the current guess of £40 million is light by a factor of 2.
Is there anything that local government do that isn't already a solved problem and in many cases a solved problem not unique to local government?
Payroll/HR? One of the first things computers were used for
General accounts? Ditto
Facilities management? People have been maintaining buildings for years, surely there have been packages for that for years. Only question, would they cover highways maintenance as well as buildings?
Social housing? I know there have been packages for that since the late 80s. Possibly overlap with facilities management and accounts to be sorted out.
Electoral register - even my lot can work that so I assume it's a long-solved problem.
Social care? Maybe this one's special but maybe someone here knws differently
Schools management? We seem to have a few readers involved in schools. What's the situation there.
Library management? Again, if my lot can run their libraries - albeit reluctantly - I assume it's a solved problem.
Is there anything which isn't a solved problem except fighting off Oracle salesdroids and coping with senior management and/or counsellors that said salesdroids have got at?
I think the problem is trying to integrate these into one humongous package. Then finding that the payroll system for schools is totally different to the one for Social Care, which is different to contracted health services, which is different to clerical staff......
And that they all have different pension record systems. Or that repairs to school buildings is totally different to repairs to council housing, which is nothing like repairs to council offices..And so on.
But you miss the point. John Deere aren't running a tractor company payroll scheme and a bank payroll scheme as well, each with different sets of scales. Let alone several schemes, all with different pension arrangements, different incremental rates, different reimbursement methods, different core hours etc.
And yet ERP systems are sold as solutions to exactly that kind of multi-business multi-national environment.
No really, that's sort of their thing. Y'know? UI-->Business rules-->DB updates
But for UK LA's I think the joker in the pack is the all-decisions-traceable-back-to-specific-council-meeting/vote.
I don't know of any business, anywhere that has that requirement.* I'm not sure if local authorities outside the UK do either.
*Unless other vultures in the committee know otherwise?
Turns out the answer is "Yes"
It was noted (WRT to Birmingham) that all major policy decisions have to be traceable to Council meeting votes.
I know of no company software that allows such tractability of major decisions, although in principle it's the equivalent to a Board vote/decision.
However that's the same for all UK councils, and maybe other countries.
Which suggests there is a market for a specific product for running LA's from parish to County size.
Maybe.
But the big savings start when multiple councils standardise on a single way of doing things without the system coping with all their existing variants. Local government exceptionalism is likely to be as difficult to work around as central government exceptionalism. "Oh we couldn't possibly use an OTS system as our requirements are sooooo unique," to recall the MOD and it's transport and logistics needs.
As has been noted by other commentators local authority requirements (and the public sector in general) are not the same as commercial companies. The rules, regulations and requirements they have to satisfy are complex and mostly created by central government and handed down from on high. However, what I don't understand is why each council goes about implementing their own system in their own unique way. Surely the requirements don't vary significantly from one LA to another (at least between councils of the same type - i.e. county, district, unitary etc).
It's a myth.
Random people on a committee, with no personal accountability.
A good start would be to have someone on it who actually knows what they are talking about.
Second option is just to keep, and advertise, the committee members. Let's know who the f'ing idiots are.
"You go high enough it always come down to one man (or woman)"
They may be the Mayor, or they may be someone on a key committee, or one of the "Officers" but for any given decision they are the person who's ordering/persuading/cajolling/browbeating people into doing things their way.
Find them. Everyone else (in their domain) is simply a minion