Europe's largest local authority settles on ERP budget 5x original estimate Europe's largest local authority has settled on a £108 million ($137 million) bill for its disastrous replacement of SAP with Oracle until 2026, five times the sum initially predicted and five years late. Birmingham City Council has been struggling to implement an Oracle Fusion HR and finance system, in a project that has left …
ERP systems are always going to need some tweaking to accommodate some special requirements.
Birmingham's problem is they thought they could change the organisation to fit Oracle then flipped mid-project to change Oracle to fit the organisation.
Fundamental requirements change mid project will only lead to disaster.
I've ever witnessed started with we'll change our processes to fit the system and ended up with trying crowbar the system into working how the organisation' processes work. The reality should be a little in between but never is. There's a lot of "we can change" (but don't) and usually it's because people want to do everything 'how they've always done it' and won't accept that that's not a reason to carry on that way, Change is good ffs.
'but we've always done it that way' is not a practical reason to NOT change it - you may not know how it's broken or at least bent/inefficient and you shouldn't decide not to change until you do evaluate the process properly. Pig headed people who want thing to stay exactly as they are will kill a project.
>But councils are stuck with central government regulations that say you must do it this way.
But presumably all local councils have to do it the same way.
It's not like Birmingham is required to use Pounds/shillings/pence for council tax while Leeds have to use Marks and Farthings
> Oops, and you must do it This Way now.
It is worse. Politicians print a new law defined mostly as an intent. Something like: "from tomorrow everyone must be healthy". Then downstream bureaucracy has to develop specific policies to "implement" it. So different branches define their own policies. There will be discrepancies, as it is up to lawyers and practicality HOW it is done.
I saw it in taxation. You go to the tax office, and there is nobody to answer specific questions. Because there are only a few experts in the whole country. Those being a few officials and lawyers who developed the specifics. Documentation is insufficient or lacking. But penalties are well defined and hefty :)
Solution: cut regulations from the top, the same way they were born initially. Prune regularly, once certain threshold of complexity is reached at the bottom. New laws must be implemented as DECISION TREES and passed downstream for immediate implementation AS IS. No more paper law compendiums. I am not sure how to address cases of under-specified laws, when lawyers are necessary. Yet such are exactly the cases for elimination.
But of course, there always exceptions. For example some document missing when applying for a permit because of unusual circumstances. Decision tree would fail. I have no answer to this. Such were reasons for bribes in some countries, as people have to proceed with their lives and business.
@ Ali Dodd:
"Change is good ffs."
Let me change around the menus, icons, labels, and functionality residencies on your smartphone, then tell me that again.
Re: "We've always done it that way." / "We have external requirements to do it that way.": those are stickier wickets.
Sometimes the obviously-better way is not better, but just different, with different disadvantages than the old way.
External requirements can be just as bullshittedly-wrong as internal requirements.
I don't see an effective way of fixing bad processes, because it's just too-politicised.
If one creates and empowers a process-fixing-tsar, the damage which can be done by having the 'wrong sort' of person filling that position equals the improvements which may be done by having the 'right sort' of person filling that position, and:
There are far more wrong-sort people than right-sort people. Compounding that problem is the additional problem that most of the right-sort people fail to make it past the HR- and political-animal-manager filters.
Meanwhile in Canada taxes comprise 36% of the price of a new home. Effectively, Canadians work for the government rather the other way around. https://nationalpost.com/opinion/blame-bureaucrats-for-taxes-fees-that-comprise-35-6-of-the-price-of-a-new-home-in-ontario
Bureaucracies save money by wasting your time: https://thedailyeconomy.org/article/trouble-tax-we-all-pay-a-time-price-for-bureaucratic-dysfunction/
Bureaucracies cannot be reformed to make them “efficient,” because the need to follow the rules excludes efficiency as a goal: https://thedailyeconomy.org/article/elon-musk-tigers-and-the-nature-of-bureaucracy/
Feel free to believe the brown envelope theory of local government if you want. It clearly does go on as it does in the private sector, but it isn't as prevalent as you think.
If Whitehall were to announce an ERP-for-councils unit, I'm pretty sure you'd be on here predicting it's expensive failure. Maybe you'd be right, but there's the thing, identifying problems, even accurately foretelling them is not a valuable skill. Coming up with realistic, feasible solutions to problems now there's a valuable and rare skill.
Local democracy means there's very specific limits on the extent to which national government can dictate to local government how it does things, and national government doesn't have the remit to develop a standard ERP for councils. Added to which, there's five main models of council (unitary, metropolitan, London boroughs, county, district) so one size won't fit all. I agree, a best practice model, and a "this is how you do it" homebrew ERP example would be very useful, but as things stand that's not how things CAN work. If you feel strongly, then perhaps invest some time and effort writing to Josh Goodman, CEO of the Office for Local Government, offering your insight and skills?
If I take the total costs of £108m, and assume that (a) costs were overwhelmingly people, and (b) the average fully loaded annual cost of each bod working on it was £100k, then we're looking at a thousand man years of work to get an outcome that still doesn't do reconciliations and doesn't serve schools.
What could YOU achieve with a thousand man years of effort?
Mind you £108m is still a lot cheaper than my last private sector employer who spent a long way over £600m trying to replace a SAP CRM and still failing.
I guess after they closed down many of the circuses in the UK they had to find a home for them in local councils as leading politicians and certain placemen within their IT/infra team to define and manage these things...
Does Oracle have a "joke board" where they dare each other to find a bigger, more stupid, incredible mug of a client/contact salesperson as each time you think they've found their limit, someone comes and surpasses it.
There comes a point we you have to put in notional amounts for anticipated expenses of the project as a whole, even on a fixed price contract there are costs to be bourne by the customer and can't be predicted exactly at the outset. Let's say an additional letter needs to be sent through the post about some issue - do you want them to have to go back for a budget increase of £2 to cover the envelope and the stamp?
I remember having to use/implement 6 sigma software development. This was based on hardware testing. You test the hardware and it fails once in a million... You test it with software and it works every time, till you vary it slightly and it fails.
Our quality target was there would be 4.6 defects over the life of the product. Yes that accuracy - down to one decimal place - so no need to have any support teams.
We hit more than number of defects in the first day, when customers could not install it the product. Their systems were slightly different to ours!
Goodbye six sigma.
Oracle has its failings but as a functional ERP system, it's not entirely culpable here. It sounds as though a poorly experienced council employed an integration partner (not named here but named elsewhere as Evosys) and then failed to manage the expectation between technology change and business change. Then as it started going badly wrong, brought in multiple consultancies (KPMG, EY and PWC all seem to have been involved at some point) who undoubtedly ran up significant bills to tell the council what they should already have known.
I've not seen anywhere a breakdown of where the £108m has actually been spent, but I'd suspect Oracle themselves would have been one of the minor beneficiaries compared to the integration partner and consultancy firms.
Do large organisations actually have reliable information about what goes on day by day?
My experience of building business process documentation in large businesses is that management rarely knows much about the detail.
(1) Department A sends a regular report to Department B. When asked, Department B says the report is useless...they throw it away.
(2) Department X sends a regular report to Department Y. When asked, Department Y says they spent quite a long time revising the report every time.
(3) Director of an HR department is shocked (shocked!) to find out that there two completely separate processes for recruiting.
(4) Director from foreign owner asserts that two computer systems are "fully integrated".....and is shocked to see evidence that this assertion is completely false.
(5) A weekly financial report to directors is replaced. It turns out that the old report was wrong.....for years....no one noticed!
.....and so on......
This is what goes on.....and now we are supposed to be surprised that an SAP to ORACLE rip and replace is a disaster!
> Director from foreign owner asserts that two computer systems are "fully integrated".....and is shocked to see evidence that this assertion is completely false.
No.
Director from foreign owner asserts that two computer systems are "fully integrated".....and refuses to see evidence that this assertion is completely false.
In the book, The Psychology of Computer Programming, by Gerald M Weinberg, failing to report bad news upward toward managers/executives who shot messengers of bad news was termed "consensus on a falsehood." (Ex.: "Sure, things are good. There are some bugs, but we can fix those few things in a post-release update.")
It will always be more.
quote: any Oracle-based solution for schools would be comparatively expensive and inefficient.
Have hackers from Oracle got into El Reg's System/360 and added 'for schools' to that statement?
Any local authority planning this should instead switch back to paper and give Oracle £50m. It will work better and cost less that way.
Certainly a contributing factor. But who would suggest they could have picked instead to magically solve all of their ill-defined problems? Fujitsu? IBM? Microsoft? U4BW?
No, as much as I despise Oracle and all of the above alternative bloodsuckers, the problem was to jump into a migration without first completely understanding its organisational implications and technical requirements.
Probably, some overzealous salesperson from the systems integrator and a clueless council director are jointly to blame here
The most obvious option was to migrate from their customised SAP ECC 6.0 ERP to the SAP S/4HANA Private Cloud edition which SAP enable with migration tools and simplification items where product improvement may be sufficient to retire the custom for standard.. That provides smart organisations the opportunity to fit to standard and minimise custom costs.
Given that the SAP ECC 6.0 ERP implementation was governed to minimise custom and ensure that was upgrade safe, they actually started from a good situation. The Integration was also high quality which matters because that can easily be more than 59% of the implementation cost, so again a good starting point.
The ability to negotiate a reasonable SAP S/4HANA Private Cloud edition cost may have been beyond BCC but not implementation partners so again an opportunity apparently missed...
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
