Here's your coat, shut the door on your way out.
Oracle consultants implementing a £13.2m University of Glasgow student IT project have been asked to leave campus, replaced with a team of unnamed "bespoke" external experts. A university spokesperson told The Reg that the university decided its needs would best be served by continuing its long-running Student Lifecycle …
Here's your coat, shut the door on your way out.
Having worked in the public sector, I can guess a defence for Oracle here (and it really applies to most vendors) - mobile goalposts. What happens is the customer (Uni management) says "We need an over-arching system to c-ordinate X, Y and Z" because they just can't have enough control. The internal techies (Uni sysadmins and lecturers) say "yeah, no probs, we'll rope in some students, give us a few months and we'll handcraft some portal or other, in the latest trendy coding language, and with lots of features you never really wanted, and it might actually work some day far in the future". The management don't like the egghead option, even though it is cheap, because they worry that students and lecturers will build in backdoors and/or get access to stuff they don't want publicised, and there is no guarantee of success, so they turn to the "professionals". Enter, stage right, Mr Oracle Rep, who offers a proven, "off-the-shelf" solution ("which may need a bit of tailoring"). Clue here - the pain is in the tailoring!
So off the Uni goes, actually not really sure what they want to achieve, but dazzled by the 50-slide Powerpoint presentations and the case studies from other academic institutes which probably have very little in common with the Uni. Now there are Oracle consultants onsite, and they see and endless pot of consulting oportunity - tailoring! Several project re-jigs later, after the goalposts have been moved all over the place to accomodate new ideas brought in as a result of continual redesign by commitee (and all these public sector jobbies have more commitess than braincells), and the project has ramped into stellar over-budget figures, the Uni has no solution in place, and some beancounter finally steps in and says "woah, stop spending the cash!" So, what is the Uni to do? They can't start from scratch, that would look wasteful (and bruise too many egos), but how do they salvage something useful from the mess they have created? Of course, they go get more "professionals"! Up starts a new gravy train, more Powerpoint presentations, lots of new promises, and the it all starts off again!
BIG HINT - set the project objectives early, peg the goalposts, and do NOT let others change them for you! Big companies doing consulting will not object to moving goalposts because they will see more consulting money as the outcome, and there is usually little penalty for them as they can argue out of penalty clauses if you change the project definition (and that's if you're smart enough to include penalties in the original negotiations, and most don't!).
Without more info I can't really blame Oracle here (even the Oracle PMs as they would probably be under pressure to accept changes from sales), so I'm more looking at the Uni management.
Probably not too far from the truth of it.
From the information that is available on the UK authority Freedom of Information website www.whatdotheyknow.org it looks like it is one big sorry mess with University of Glasgow left to pick up the pieces having had 2 prime contractors leave the job with precisely none, zero, ziltch of the system so far delivered.
Given this I'd say that estimates for roll-out in 2010 look a little unrealistic.
Matt put it so well that there's not much else to say. Brilliant.
I've seen that happen more times I bother to count: Buyer doesn't know what they want and can't be bothered to specify it at any point.
Guaranteed failure and consultants get very rich as a byproduct. Plain stupidity from those who pay, usually the taxpayers pay but no-one asks from them.
You always get something stupid when buyer, seller and payer are three different entities. I can guarantee that.
The pubic sector have no idea about contractual relationships. And from bitter experience they know every silly childish trick in the book. I have seen a good web design house go under because part of the NHS decided to "dissapear" into another part of the NHS rather than pay for work done. And yes "renaming yourself" to avoid payment is legal if you are public sector and know what you are doing.
Typical universities- loads of money, good at complicated projects, but when it comes to the simple stuff like knowing what the year is they trip over their shoelaces!
I wonder what the reason for the switch is? Something to do with cutbacks from their government funding perhaps?
They're being a bit silly aren't they? Never mind FOI requests - they should be making this information available automatically, It smells of a choice that doesn't bear up to scrutiny - implying they don't feel they can justify it on grounds like competence, value for money, fairness in the selection process etc. Business as usual then in this glorious nation of ours.
One puzzle though : the company taking on the job ought to be proclaiming their success. Why do they want to keep it secret as well?
Reading between the lines, "it defined public interest in this context as its potential inability to conduct negotiations and establishing and developing relationships with third parties" sounds like OCS was sacked.
I wonder if they did something really wrong, or just tried to leverage more money from the Uni? If you ask me, it sounds like some US corporate droid forgot that UK universities are funded differently than those in the US.
It defined public interest in this context as its potential inability to conduct negotiations and establishing and developing relationships with third parties.
Sorry "its inability"?
We cannot tell you because it would affect our inability?
So it would stop us failing, I know its 0900 on a Monday morning-ish but wtf!
£13.2m, 26500 users: £500 a head, give or take. That's quite a lot of money, and it would be interesting to see the business case for it. If it's about lecture theatre scheduling, you could build several dozen more for the money and be a bit less efficient. If it's about cost-reduction, it's perhaps 200 person-years of burdened costs, so over a 7-year cycle you'd want to see perhaps 30 fewer staff in the administration function. As an IT project for a single-site small enterprise, it fails the ``hmm, that sounds a lot'' test, and any subsequent fail on the part of the implementation team (and to lose one partner could be considered a misfortune, but two looks like carelessness) is in a way a side-show. Although as it will be pushing up acquisition costs, opportunity costs, later opportunities to buy more modern solutions and generally tying the enterprise in knots, that can't be totally ignored.
Oh the stories I could tell about an implementation of Campus Solutions at another institution! Well done Glasgow! Seems someone up there has an inkling of how to get things done.
...they had to recruit help from Strathclyde Uni and are too embarrassed to say so?
"Atos Origin – a specialist in integration of SAP and Oracle systems – was apparently overwhelmed." Did scope creep become scope sprint; would love to see the ProjMgmt logs.
Enter then exit Oracle. Now selected experts are having a crack...
OCS requirements are not onerous... not compared to kludge-fest complexity that is corporate IT, and I agree with Ian35 - 13.2mil is a serious bit of cash. Question - they couldn't make it work, or wouldn't make it work?
As for "potential inability to conduct negotiations and establishing and developing relationships with third parties"... was the business case based on third-party project costs, then stonewalling and scope sprinting the project to get rid of the third-parties so it can be done inhouse at a fraction of the cost, and the remainder of the budget used for pet projects that would never ever have seen the light of day otherwise?
As some people have jumped upon the "universities don't know what they are doing" bandwagon.
Campus Solutions in its current state is not fit for purpose in a UK institution. The entire system puts the student at the centre and the system revolves around them as that is the US system. In the UK, the degree is the central item and the course units and students hang off it.
Don't get me started on the assessment and progression components, we do not use the gpa in this country but that is central to the way Campus Solutions records marks and attempts to move the student to the next academic year. Woe betide any institution that has different criteria for governing academic success and failure, let alone differences within the institution itself!
This results in a massive amount of customisation of the code to try and make it fit the institution it is being hammered into place in. And as we all know, once you start changing the off the shelf solution, here be dragons.
Then there is fee calculation, ho ho! Good luck to Glasgow, they are going to need it! Plus, god help them when they try to get information out of the system. One key use of a database is storing data. Another key use is retrieving information, as I said, good luck to them.
As for the Oracle consultants, well... no, I'd better not.
Granted there isn't an off the shelf product that performs all of the tasks of UK University administration but crowbarring some products into position really isn't a good use of time and resources.
So write one and cash in yourself? KerCHING ;)
Java and PostGRES do yer mate? Call it a million quid, no questions asked, alright?
So basically the core *model* of this software is *totally* different to the way Glasgow (and presumably *all* UK) universities management is structured and *effectively* it's going to need to be completely *rebuilt* to work in the UK and deliver things UK Uni's need.
Seriously how the f$%k did this procurement get to this point *without* someone realising just how *different* this software works from the way UK Uni's operate?
Competitive tender anyone?
Doesn't look like it.
It's a strange one as in the minutes of Student Lifecycle Project Board Meeting on 4 Feb 2008 where evaluation and recommendations are discussed (see section 3.6 PQQ Evaluation and Recommendations of http://www.gla.ac.uk/media/media_135387_en.pdf), it is noted that 'The Board raised a slight concern that as the University are not major users of Oracle, the technical learning curve for using PeopleSoft could prove to be too big.'
Hindsight is a wonderful thing!
Oracle overcomplicated, overexpensive rubbish. Go SQL Server.
"Oracle overcomplicated, overexpensive rubbish. Go SQL Server."
It's even worse: It's overcomplicated, overexpensive rubbish which doesn't work.
This is not a high-availability solution, nor very complex and don't even have huge amount of data either, so MySQL or similar would be totally sufficient and any university having computer science in program, has enough capable programmers to do anything like this in house.
Of course, no money for administrations pet peeves then and I suspect that's the real reason why in-house wasn't even an option. As in so many other universities: Administrators are in their jobs just for filling their own pockets from universitys accounts or from bribes, nothing else.
Nothing to do with the database, its an application that's the issue here. As you don't seem to appreciate the difference, think its safe to discount your views on SQL server also.
Of course it is troll! That's why Fortune 500 companies run petabyte sized installations off Oracle kit. Not saying it's perfect, but Oracle starts getting big at around 50TB, SQLServer thinks 500GB is big.
This situation isn't unique. As someone who has been part of a team of contractors called in to get major Oracle projects completed after one (or in one case two) of the big system integrators had failed I can safely say that it happens in the private sector too.
Oddly we got the systems working more or less on time and to budget. This resulted in happy customers who even went as far as to write nice commendation letters.
Now, I'm not saying that all Oracle contractors are the dogs dangly bits, but there are a number of really good ones out there.
Surely it is time for HM Gov to pay for a common system that can be adopted by all Universities? They could even license it to other countries to recover costs? Or am I being way to simplistic?
Yes, yes, and yes.
Not that I don't agree with you entirely. I do.
Just too easy, this consulting lark. Cheques to the usual address please.
If the government stepped in and started a project, that would be public sector spending. And EVERYONE knows that the private sector is more efficient and better value for money.
UCAS and HESA are government and they don't work - that is what goes wrong with universities and the software they try to use - they have to interface with silly systems that don't work :O(
While I find SQL is orders of magnitude cheaper, and easier to put in, than Oracle, if everyone goes to it, they'll just put up their prices for the rest of us.
As for Oracle or any other consultants, well, I don't trust any affiliated team, let alone one who suggests index queries should only be used over Table Access Full operations if you're under 1% of the rowset by population, (which someone tried to push on me.) Hiring individual techies is always cheaper, and you can sack the ones who...
4. Always late,
5. you just don't like the look of.
without touching the others.
I've no idea why people hire consultancies, (although I service my own car, fix my own boiler, wire my own house, build my own space station etc so I wouldn't would I?)
Ah, so you have worked with me before ;-)
In my experience consultancies usually come cheap or next to nothing, IF you have cut a good deal with the vendor. You cough up £2M for example they may well chuck in 30 days of consultancy to sweeten the deal.
Software is just like the narcotics trade, once the dealer has you hooked on a something, it's sheer torture to get off it. If a few free consultants are thrown in, that means the software will more less stay and then the vendor can rape you for license fees every 12 months. Then just like the drugs, the price slowly rises and as you are "addicted" you have no choice but to pay.
Reason they won't say who the new company is probably because they're exclusively overseas consultants being paid a huge whack of money on an already ludicrously expensive contract?
Having seen several projects being "implemented in stages" instead of in one hit, and several vendors having cracks at the same project, I don't think there is enough info here to point to a single reason.
Past projects I have seen have slipped because,...
1. Poor or vague definition of scope or deliverables
2. Loss of one or more key Project Management personnel (on either side)
3. Poor technical analysis that misses functional issues
4. Bad costing models causing cost blowouts
5. Lack of appropriately skilled personnel to implement technology
6. Obstructive units or individuals preventing progress
Keep pressing for release of FOI data, El Reg. This might be interesting.
The phrase "implemented in stages" always sends a shiver down my spine too, especially if the requirements changes between stages.
<List of project failure reasons>
However a "Phased rollout" is not *inherently* bad and may be the only *sensible* way to do a project (most UK govt IT projects *should* have been phased rollout with feedback from the initial sites to guide the rest. How many have is another matter)
That list could apply to *any* roll out method. It would seem that for something like this setting it up at all sites in parallel with the current system (is there a current system?) for a time, evaluating the results then setting a switch off date would seem reasonable.
thumbs up for a good list. Just sad its the *same* old stuff yet again.
Were Oracle caught out Snoopy Spooky Phishing whenever they were contracted to do Phorming and Pharming?
In other words, rather than offering supply of provision with services were they leeching parasitic worms and testing trojans in higher seats of learning and academia than they are used to handling?
...they added another 8 core processor to the the server and found it cheaper the start again than pay the extra Oracle licensing.
I think my point when referring to "implementation in stages", was that the original plan was for a Big Bang, which then had to be changed to Phased Rollout, so they could at least show some progress in the original time specified and avoid admitting any form of outright failure.
Apologies for the long delay.
"so they could at least show some progress in the original time specified and avoid admitting any form of outright failure."
And I suspect this is the *biggest* problem with some IT failures. The *total* unwillingness to admit (in *any* way, shape or form) that "We were wrong. It seemed a good idea but it wasn't. We take full responsibility").
The sub text is of course "We were gullible, did not do *our* research to work out what was actually *needed* and believed *every* word the con-tractors and con-sultants told us."
But no one would come near being *responsible* or *gullible* (and given the disconnect between how UK HE establishments work with the US environment it seems to have been designed for you'd have to have swallowed *every* word to think this would work without major changes).
This is not an implementation strategy, it's a face saving strategy. Maybe the next batch of students will fair better.
... in this. I'm thinking of the OES (oracle education system) of the mid 90's. This was developed by Oracle for a consortium of about a dozen Universities and, as far as I know, only implemented by one of them (after lots of in-house redevelopment).
Year 2000 was approaching and many Uni's needed to replace quite antiquated systems so jumped ship to alternative, already in use, commercial products.
There was a reasonable number of suppliers for HE and FE student systems, but a lot of them were bought out by Capita and the choice greatly reduced.
Requirements should absolutely be set in stone before any implementation work is carried out, with any extra requirements to be delivered after the agreed deadline for the initial project, for the sole reason that it keeps things manageable and simple. Changing requirements in projects inevitably end up in horrid spaghetti code kludges which make technical architects cry themselves to sleep at night.
Yes, it's easy to agree a few changes with the architect, but then these changes have to filter down to the staff actually building the software, and you either end up with ten copies of the spec or ten appendixes of doom. And kludging things this way invariably ends up costing the customer more in the long run than doing it properly at the outset and dealing with change reactively.
...but clearly somebody wised up and got rid of them.
Some posters are a bit harsh on Glasgow. Such commissioning decisions are only for cloud-dwelling senior managers to make, they're not for academics. Senior managers don't need to deal with minutae such as where the decimal point goes in contracts, this is only for little people. Big costly projects = I am important.
No other UK Unis have put in that system without significant modifications to it - it just does not work for them. I have no idea why Glasgow is perservering with it as surely by now they should have realised that and also it looks like they've had an ideal opportunity to get out of it at this point.