Job's a good 'un: Bank of England tech project wins watchdog praise

Parliament's spending watchdog has held up a successful large-scale public sector tech transformation as a rare example worth emulating, in a striking departure from the usual diet of failure and overspend. Reporting on the implementation of the Bank of England's £431 million Real-Time Gross Settlement (RTGS) system, the …

[Albert Coates]

Job's a good 'un

Well done, I suppose - they had long enough - 2016 to 2025, and all for only £431 million - but can you really imagine the BoE screwing up its global payments system, handling £790 billion a day?

[Bebu sa Ware]

"can you really imagine the BoE screwing up its global payments system"

All too easily^†… thankfully I shan't have to.

_{† recent precedents supply an abundance of inspiration to even the most unimaginative soul.}

[ChoHag]

> Among the features marking it out as something of a gold standard, is the time the organization spent understanding the scope and objectives

lol

That is all.

[Roland6]

How to nicely put the boot into “Agile”.. and to remind people of a truism that predates the IT industry….

[Rjan]

Not to support the "agile-industrial complex", but the difference is probably that a central bank payment system has an extremely stable environment (and the central bank has extremely powerful control over that environment), and the entire thing was adequately funded (almost half a billion pounds) and adequately paced (apparently 9 years).

For most parts of the private sector, or for parts of the public sector that are subservient to either the private sector itself or to the "efficiency" diktats of anti-state political vandals, the three factors of stability, funding, and pacing, would almost certainly be far more adverse.

In particular for the private sector itself, 9 years is a very long time for all but the very largest corporate bureaucracies (like global banks etc.).

One principle from Agile which remains current, is that lots of things change over time in unpredictable ways. Most private sector businesses should be wary of embarking on any software development estimated in years.

[Roland6]

>” One principle from Agile which remains current, is that lots of things change over time in unpredictable ways. Most private sector businesses should be wary of embarking on any software development estimated in years.”

Unpredictable ways…

One major transformation project with a 2~3 year delivery plan was into its second year - we had agreed and purchased hardware and software platform, architecture (both designed and deployed) and so development teams were in place and building for the target platform…

The client got purchased by a rival, who’s IT preferences was for a wholly different hardware and software platform and had made a public declaration to make £00’s M. savings by using a common IT platform… It was nice to be able to walk away - someone else’s problem…

[David Hicklin]

So, Planning, Planning , Planning - lots of Planning and getting the full scope right before setting spades to the ground. Heresy !!

Yes it tool a fair few years but they got it Right First Time - how many years have Birmingham been at it so far ???

[cdegroot]

They were replacing an existing system, so they got it more right the second time.

Still, I wonder, where is the complexity that requires half a beeellion Pound Sterling? It never ceases to amaze me, these amounts.

[Blitheringeejit]

Very Much This....

"Its replacement ... is still internally hosted.."

"The BoE has also been able to retain the intellectual property for the system, and has transferred its day-to-day operation and future improvement to an in-house team."

It may just be that the BoE has enough money to do an IT job properly at the first time of asking - but one lesson should be how much cheaper it is to do it right at the first time of asking, instead of doing a half-arsed job on the cheap and then paying through the nose to mitigate the fallout for decades.

[Anonymous Coward]

Re: Very Much This....

It may just be that the BoE has enough money to ...

... be impervious to the brown envelope virus and knowing the turf, made it a point to keep all the beancounters out of the decision making process, leaving it to the professionals.

Seems easy enough, right?

.

[Martin an gof]

Re: Very Much This....

It may just be that the BoE has enough money

And time, and (I imagine) lack of political interference. Note how the project was significantly longer than your typical political cycle and presumably they weren't dealing with changes in management and direction every time a politician went out to dinner with a persuasive CEO or lost a cabinet post in a reshuffle.

M.

[Tron]

They might want to send a copy of that report to Birmingham City Council.

Plus all the other local councils emulating Birmingham, and the HS2 folk. Hell, just email a copy to everyone with a .gov.uk address.

[elsergiovolador]

Plank of England

The technical content of RTGS is not difficult. It is a message bus, a ledger, a settlement queue, a gridlock resolver and a set of risk controls. Every one of those is a solved problem with reference implementations going back to the 1980s. SWIFT, CLS, every major CSD and every tier-one investment bank runs versions of this. ISO 20022 migration is a schema change, not a research programme. The throughput is laughable: £790bn a day across CHAPS settles in tens of thousands of messages per day, not per second. A competent team of fifteen could build the core engine in eighteen months. Two engineers and a weekend could prototype the matching logic.

Nine years and £431m to rebuild this, with annual running costs doubling to £40m+, is not a triumph. It is what happens when you route a tractable engineering problem through competitive dialogue, four replans, a prime contractor billing on time and materials dressed up as fixed price, nearly 2,000 "functional and non-functional requirements" written by committee, and a governance structure that treats every design decision as a board-level event. The Bank ended up with a system whose run cost has roughly doubled despite cloud-native technology that is supposed to make infrastructure cheaper. Somebody should ask where that delta is going.

The procurement theatre PAC is so impressed by, paying losing bidders £125k for a design exercise, the practical build-off, the competitive dialogue, is itself a symptom. Serious buyers of trading and settlement systems do not need any of this. They know what they want because they have engineers on staff who have built or operated the equivalent. They write a tight spec, they shortlist on capability, and they negotiate on price. The elaborate choreography exists to substitute process for in-house judgement, because government procurement assumes the buyer is technically illiterate and must be protected from itself by a paper trail thick enough to survive judicial review.

What PAC is actually celebrating is that a central bank, with sovereign funding, captive demand, no shareholders, no competitive pressure and the Bank of England's pick of the City's payments expertise on its non-executive board, managed to replace a mainframe ledger in only nine years and only 15% over budget. This is the institutional equivalent of praising a Michelin-starred restaurant for successfully boiling an egg, because the canteen down the road has been trying for twelve years and set fire to the kitchen twice. Studying how the egg was boiled is the wrong exercise. The question is why the canteen cannot hire a cook.

[cd]

Re: Plank of England

No mention of what it cost for an osteopath visit after they sprained their arm patting their own back.

[Rjan]

Re: Plank of England

Not quite as bad as the dreaded "manager's elbow" - caused by patting themselves on the back whilst also straining to keep their backside covered.

[JLV]

Re: Plank of England

I agree that the bar was pretty low £400M+. Still, going back to the old programmer homily: “Make it work. Then make it fast.”

i.e. once you reached the massively impressive milestone of mostly achieving somewhat-successful £100M+ pub sec IT projects, rather than quasi-guaranteed £200M+ absolute fails, improve further.

Baby steps, baby steps.

[Rjan]

Re: Plank of England

The reality is that the BoE has paid the real price, and waited the proper time, for this kind of development.

Everyone like you who says it can be prototyped in a weekend, ends up with a quagmire and a bonfire of funds. It's not like we haven't accumulated evidence of this.

If it takes an estimated 18 months for a team of 15 to build the "core engine", then taking 9 years for the whole project (including determining what that engine should do, and settling all the "non-core" aspects of its use) is well within the ordinary range of multipliers for a development project estimate.

[JLV]

Re: Plank of England

True, within some possible tuning: you’re talking 22.5 man-years @ say £200k / year x 2x => 9M for the core. 10x whole project? 20x? does it have to be 40x?

Certainly not “prototyped in a weekend” territory.

The corollary however is that, while BoE may have bespoke-ish requirements and deep pockets, not every local council can afford to think of itself as “very special” and start down a path where the “success branch” means spending multiple, multiple, tens of mils to run their payroll and financials for an entirely bespoke and specific solution.

Not when the councils next door are operating within pretty much the exact same legal and biz parameters.

Cough, Birmingham, cough.

[elsergiovolador]

Re: Plank of England

The "ordinary range of multipliers" argument is unfalsifiable by design. Nine years is fine because the multiplier is six. Twelve would have been fine because the multiplier would have been eight. Whatever falls out the back end gets retroactively declared the natural pace of the work.

Look at the comparators. CLS Bank went live in September 2002 settling FX across seven currencies on day one, payment-versus-payment with no operational precedent to copy, in roughly five years from concept to production. TARGET2 replaced the entire Eurosystem RTGS in about five years from ECB decision to first migration. Australia's NPP, real-time retail with ISO 20022 and overlay services from day one, took around three years from SWIFT contract to public launch. T2-T2S consolidation, arguably more complex than RT2 because it spanned multiple central banks, took five years and three months. Three to five years for genuinely novel work. Nine for RT2.

The "we waited the proper time" line conflates two very different kinds of elapsed time. Engineering time is real and properly paid for. Governance time is the tax the institution imposes on itself through replans, dialogue rounds, requirements workshops and steering committees that exist mainly to consume calendar. Treating the second as inherent to the first is how the tax never gets challenged.

As for "evidence that prototypes fail", it does not say what it is claimed to say. It does not show that small fast teams produce quagmires. It shows that public sector pilots, staffed through the usual procurement channels, produce quagmires. Different proposition.

The big consultancies do not put their best people on a public sector discovery phase. They cannot, because the day rates the public sector tolerates are below what those people earn on commercial work. So the prototype falls to graduates two years out of a non-technical degree, career consultants who have never shipped to production, and a thin layer of architects whose real job is producing slide decks for the steering board. Of course it ends in tears.

Then there is the incentive structure, which nobody at PAC will ever put in writing. A successful prototype is the worst possible outcome for the bidder. An honest thin slice produces an honest scope, a smaller build contract, lower bench utilisation, a thinner partner bonus pool. A failed or inconclusive prototype is much better business. It justifies a longer requirements phase, a bigger build contract, a steady tail of change requests, and years of remediation work. The supplier's commercial logic is directly opposed to the prototype succeeding, and nothing in the procurement framework corrects for it.

So yes, the evidence is real. It just supports "this is how long these projects take when staffed with the wrong people through a system that pays them more when they fail". Replace either variable and the evidence base evaporates. Which is precisely why neither variable ever gets touched.

[Rjan]

Re: Plank of England

@elsergiovolador, I think the fundamental issue you're touching on, is that decades of liberal governments have been ideologically opposed to employing trustworthy expert staff in the public sector, *or* paying significantly more for private sector equivalents.

So like you say, they just end up hiring the dross from the private sector - whose failures are then blamed on the commercial incompetence of the public sector, and used as all the better propaganda for liberals who want to vandalise even more of the public sector.

As for the "ordinary range of multipliers", it isn't supposed to be falsifiable. The reality is that estimation of software projects is not a sufficiently exact science to meet the exactness that is often wanted for business reasons. Most developments are fundamentally speculative endeavours.

Also, my experience of everyday in-house development work (all in the private sector) is that things taking 10 times longer than expected at the outset is not unusual - but for in-house staff, this common experience often flies under the radar when the initial estimate is only an hour or a day.

I've also witnessed (not caused or participated in) enough failures over the years to know that development projects whose estimates are more naturally expressed in man-months rather than man-days, have a greater chance of total failure than of success.

[JLV]

Re: Plank of England

That's one way to read it.

Another is that public sector unions in many countries insist that everyone is on a level playing field and pay isn't determined overmuch by merit. The way to go up is by seniority and/or by management level.

In other words, if you had techies pulling $150K+ salaries to save tax payers hundreds of millions, the unions would likely scream to high heaven. Not least by other techies' unions.

That's not always true. For example, people have long learned that doctors can easily move to more generous locales, so cough up fairly tidy sums (at least past the internship indentured servitude phase). But it is true more often than not.

That's not to say ERP gigs are guaranteed to go right in the private sector. They are not. But at the end of the day, all these businesses, whether private or public, need to learn how to a) source competent staff and/or consultants b) draw up realistic requirements and c) avoid bloodsucker consultancies. The public sector ought to find it easier to coordinate this by rotating staff amongst different projects, since they are after all the same employer while an individual bank will change ERPs every 20 years or so: practice makes perfect. Or at least, not Crapita.

[nijam]

Particularly notable is that they didn't start by saying "We need a bespoke solutionn, because our problems are entirely different from every other financial institution in the world." In contrast to every local authority in the UK, not to mention the NHS (repeatedly).

BTW, there's no such thing as the NHS, really. It is actually hundreds of dfferent organisations (just as they were before the NHS was founded) sharing little except a common logo. Oops, almost forgot - they do share fax numbers with each another, mostly.

[elsergiovolador]

The British civil service is staffed by three archetypes, and none of them can specify a system.

The first group are there because the wages are deliberately low. Low pay filters for people whose families are comfortable enough that a Whitehall salary is pocket money, and who are placed there precisely because being on the inside is the point. Influence accrues, ladders are climbed, and policy gets shaped from positions that pay less than a regional sales manager. The trade-off is that several centuries of assortative mating within a small gene pool has not, on average, produced a class noted for operational competence. These are people who have never held a job in the sense most readers would recognise: never been measured against a P&L, never had a customer who could leave, never built or shipped anything. When the button is in a slightly different place from where they imagined, they genuinely cannot find it. So the solution must be bespoke. Not because the requirement is unusual, but because the user is.

The second group are there because the private sector did not want them. Their entire professional strategy is to remain employed until the gold-plated pension vests, which in practice requires nothing more than turning up and not being caught on camera doing anything actionable. Sackings are vanishingly rare and contested for years when attempted. This group has no working model of how a competent organisation runs because they have never been inside one. They defer instinctively to the first group, partly out of class reflex, partly because at the working lunch the words "oh, don't worry, I'll get this" still carry a small magic. Between them, these two groups generate the cultural certainty that Something Must Be Done, paired with a complete inability to articulate what that something is. Hence the two-year discovery phase, the 2,000 requirements, the workshops, the design sprints, the renamed programme boards.

The third consequence follows mechanically. Wages set low enough to repel anyone with marketable engineering skills mean the actual expertise has to be bought in. Not hired, because hiring would require paying market rates and admitting that a SC-cleared principal engineer is worth much more than a Grade 7. Bought in, through procurement, from the same half-dozen consultancies who have spent thirty years learning that the buyer cannot tell good work from bad and cannot define the deliverable with enough precision to enforce it. The invoices bear no relationship to the value created because no one on the client side is equipped to assess the relationship. The consultancy runs rings around the civil servants, and the civil servants are quietly relieved, because if the project fails it is the supplier's fault, and if it succeeds they get the promotion. The supplier, for its part, does not care which way it lands. Procurement law forbids past performance from counting against them in the next tender, so the disaster on one programme has no bearing on the shortlist for the next. The same logos that delivered the last full-spectrum disaster are sitting in the dialogue room for the replacement, with straight faces and refreshed slide decks.

This is why every council needs a bespoke case management system, every trust needs a bespoke patient flow tool, and the NHS, which as you say is a logo stretched over hundreds of organisations united chiefly by their fax machines, needs a bespoke everything. The problem is that the same three archetypes staff every one of those organisations. Each instance genuinely believes its workflows are special, because the people inside have nothing to compare them to. They have never seen a generic solution work, because they have never worked anywhere a generic solution was allowed to.

As for the Bank of England, let us not get carried away. It is a less catastrophic version of the same machine, insulated by its own pay scales and a non-executive bench drawn from people who have actually settled a payment in anger. That is enough to clear the bar set by NS&I and the Home Office, which tells you more about the bar than the Bank.

[Anonymous Coward]

You forget those of us who actually want to be in public service and are deeply frustrated by all the clowns you list so eloquently. I have several colleagues who filled their boots in the commercial world and can afford to "retire" to government employment on the proceeds. Others are content to live on air, I have no idea how they manage. There are just enough of us to keep the wreck afloat, even if way off course and the bilge pumps working overtime.

Oh, and another big reason everything gets contracted out is because when the squeeze comes down from on high, you just sack all the contractors and let everything grind to a halt until the squeeze is lifted. Can't do that to your fellow servants of His Majesty, you'd have to find them new posts. A great moment to clear the slate on all those past failures too, with the witnesses all gone.

[Doctor Syntax]

You could always apply for a job there yourself and show them how things should be done. You're obviously envious of their pay and conditions and consider yourself way more than good enough.

[elsergiovolador]

A single person cannot change a corrupt system. They get absorbed, neutralised, or quietly managed out, and the machine carries on exactly as before. The pathologies described, the procurement theatre, the deference to incumbents, the captured commissioning culture, the legal prohibition on holding suppliers accountable for past failure, none of these are individual failings that a competent hire could fix from the inside. They are the system working as designed, and the people who designed it are not interested in having it redesigned by some bloke off the internet who thinks he could do better.

The change has to come from outside, through the ballot box, and that requires an electorate that understands what is actually happening with their money. We don't have that. We have an electorate that thinks the choice is between slightly more spending and slightly less spending, while every party in serious contention is staffed at the top by the same Oxbridge PPE pipeline that produced the problem, going to the same dinners, accepting the same hospitality from the same consultancies, and rotating into the same advisory roles when they leave office. You cannot vote your way out of capture when every option on the ballot has already been captured.

The honest answer is that the situation will not improve until enough people notice, and people generally do not notice until they are personally inconvenienced by it. We are not at rock bottom yet. Council tax keeps rising, the bins still get collected most weeks, the NHS still treats you eventually, and the lights stay on. When those things stop being true, the conversation will change. Until then, the wine and steak dinners continue, and the same logos keep winning the same contracts to deliver the same disasters, and pointing this out on a tech forum is about the limit of what any of us can usefully do.

[Anonymous Coward]

Maybe one person can't change the system, but we can do our best to make sure that the "stuff" we are responsible for is done as best we can direct. The processes are "frustrating" and slow - but given the nature and importance of the project, completely understandable.

OK, ours is not the sort of team people here will be thinking of (it's serious hard-core engineering), but I can assure you that we have a lot of very capable people in the department. OK, it helps that a lot of them have a armed-forced pension so can afford the civil service pay levels (as well as actually understanding what we need to provide to the end user).

PS. The Alpha pension is not anything like as generous as Classic - which closed to new entrants something like 2 decades ago.

[elsergiovolador]

The reply boils down to "I know some competent people, therefore the criticism is unfair, and by the way please feel sorry for us about the pension." Both halves deserve unpacking.

On the pension. Alpha is a defined benefit, career-average, inflation-linked scheme with employer contributions north of 28%, no investment risk, no market exposure, and a guaranteed income for life. The private sector equivalent does not exist. Defined contribution is the norm, employer contributions are typically 3% to 8%, the employee carries all the investment risk, all the longevity risk, and all the sequencing risk. To replicate an Alpha pension privately you would need to be putting away something like 25 to 30% of salary into a SIPP every year and praying the markets cooperate for forty years. Alpha is less generous than Classic, true. So is most things compared to 1995. It is still a transfer of risk from the individual to the taxpayer that has no analogue in the wider economy, and pretending otherwise is the kind of move that only works inside a bubble where everyone has the same deal. Add in job security that is, in practice, near total provided you do not actively make the news, and the package is comfortably ahead of what comparable roles command outside.

On the competence argument. If the civil service is full of capable people, the obvious question is why nothing in this country works. The Chancellor extracts more revenue every year, the headline tax burden is at a postwar high, and the visible output is declining services, collapsing infrastructure, courts running years behind, NHS waiting lists measured in geological time, and councils going bankrupt one after another. Either the money is being spent on something other than service delivery, or the people spending it are not as effective as advertised, or both. Pointing at a handful of well-run engineering teams does not answer the question. It just narrows the scope of the indictment.

The one consistent winner from this arrangement is worth naming. The big four accountancy firms, the global consultancies, and the tax-optimised multinationals who provide the outsourced "delivery capability" the civil service cannot recruit for. They bill at rates that would fund three permanent hires per consultant, structure their UK operations to pay minimal corporation tax, employ former senior civil servants as advisors within months of those civil servants leaving post, and reappear at every dialogue table regardless of how the last engagement went. Civil servants of a certain seniority are remarkably warm towards these firms, attend their hospitality events, accept their secondments, and end up working for them on retirement. None of which is illegal. All of which describes a system whose incentives point in a very specific direction, and that direction is not towards the British taxpayer.

If the people inside genuinely cannot see this, that is itself part of the problem. If they can see it and have made their peace with it, that is a different problem. Either way, "we have some good engineers, please be nicer" is not a defence.

[Anonymous Coward]

Your one line summary is unfair.

When the message is "there are no competent civil servants", pointing out that there are in fact many of us is a valid response. The fact that the only people or projects you hear about are "bad news" simply proves the old adage that good news doesn't sell newspapers. "Project delivered without drama" doesn't sell papers, "project is a disaster does" - hence you hear about all the failures. That's not to say there aren't enough failures for that to be newsworthy and a concern.

My project could be considered a failure to the likes fo the Daily Wail, and to a certain extent it does have it's problems. But anyone looking with a fair and open mind would accept that the bulk of the problems are actually due to the legacy inherited from political decisions - more than that I can't say.

And I didn't say please feel sorry for us about the pension - just be honest. True, it's not bad, but bluntly you say there's nothing like it in the private sector, but I know colleagues in the private sector who do actually get something comparable - apart from the uncertainty about how stock markets might fare. True, that's not common, but similar levels of future pension do exist - and please don't forget that there are still private sector DB schemes for those who are old enough to have joined one before they closed to new entrants.

I will quote just one line "... the package is comfortably ahead of what comparable roles command outside ..." That is not true except for a lucky few. Allowing for the higher salary available for comparable roles, putting a reasonable amount into a DC pension will easily get you a similar outlook for retirement as well as more income today. Obviously, that does depend on the sort of role people are doing - as I say, ours is serious engineering, not the sort of skills you find in abundance at the job centre.

As to other departments, you'd have to ask people who know what goes on in them - I don't have any insight into those.

.

One thing I think we both agree on is contractors.

They are essential - like most businesses, the civil service (or rather, the departments they run on behalf of UK Government) have times when they need specialist knowledge/skills while the justification doesn't exist to have them on the payroll permanently. To that extent, the use of contractors for specific short term requirements makes sense.

But the mass hand-off of "stuff" to the usual suspects is a problem - which creates a self-fulfilling requirement to use them. They suck skills out of the CS, offer pay the CS can't compete with, and generally make it "hard" to have a competent CS capable of doing things in house. Hence the in-house option keeps looking less attractive, and so more stuff gets outsourced, rinse and repeat. A lot of that is also due to political decisions - such as insisting on (AIUI) an "outsource by default" policy.

I might at this point mention that our industry partners also have recruitment problems, and employ a significant number of contractors (usually via the umbrella companies) - but when they leave, they take the knowledge of how all our processes work, and a replacement then takes time to get trained up.

[ChoHag]

Get out of here with your common sense and your logic!

How is anyone supposed to rant in the face of that?

[Doctor Syntax]

Particularly notable is that they didn't start by saying "We need a bespoke solutionn, because our problems are entirely different from every other financial institution in the world."

Maybe they didn't explicitly say that. They just spent a couple of years designing the application and they retained the IP. It sounds like a bespoke system to me.

[J.G.Harston]

It's notable that GP Practices - that are *not* part of the NHS, but are merely *paid* by the NHS - have had this solved for decades through private sector provision of open market clinical systems, rather than throwing piles of money and years of work into one huge unified tottering ediface.

[JLV]

Didn’t Vulture Central cover this particular project before? Might be worth reminding us more prominently lest we believe Trump has become honest, cats lie with dogs and public sector IT successes are multiplying to become as numerous as honest used car salesmen.

“At a PAC hearing in March,” is a link to that Reg article but not identified as such.

[ecofeco]

Feature creep

It seems that all the wordy words just basically, said, "We eliminated feature creep and then tested and verified milestones".

You know, basic project management.

Feature creep is the number one destroyer of ALL projects. Number two is lack of testing.

[Doctor Syntax]

Re: Feature creep

Number zero is starting without any idea of what it's supposed to do.

[Anonymous Coward]

Hah......Ten Years....And No Mention Of "Agile"....

Quote1: ".....set out clear requirements for the new system...."

Quote2: "....time is taken before contracts are let to properly understand the business and technical complexity and requirements..."

We've had twenty-five years of hearing that "requirements are the problem"!!!! (See Agile Manifesto, 2001)

So in 2015-2026 the Bank Of England chooses to disagree.......but still takes ten years for a project.

Why is there ABSOLUTELY NO discussion of this rather strange story??

[P_Jamez]

Re: Hah......Ten Years....And No Mention Of "Agile"....

In my 15 years as a freelance project manager, I have only once observed a pure Agile project run in the wild, and it was a very unique situation. Every other project that claims to be agile, still has milestones and phases that need to be signed off, contractually defined deliverables etc. that mean it is some hybrid. Sure the project claims to be scrum and uses sprints but generally follows some sort of waterfall phases.

[J.G.Harston]

Re: Hah......Ten Years....And No Mention Of "Agile"....

If having project requirements is a problem, then how on earth do you specify what the project is supposed to do?

"We want a funds transfer reconciliation system"

"Bolleaux to those requirements, I've written you a video editing package"

[Anonymous Coward]

Re: Hah......Ten Years....And No Mention Of "Agile"....

@J.G.Harston

Where have you been for the last twenty years?

....so you've never seen "The Wall Of Post-It Notes"......

....and the matching set of "user stories"........................

.....and two week scrums (or sprints) and endless discussions

about "minimum viable deliverable"??????????????

.....Oh....and then there's the "Agile Product Owner"..............

.........and the bolshie development team who COMPLETELY

REJECT the idea of any "project manager", or even any "management" at all!

Where have you been for the last twenty years?

[Anonymous Coward]

Avoiding The Problem Of Estimating Bespoke Development

Once upon a time, say thirty years ago, there was (usually) a way of

avoiding the problem of estimating bespoke development.

It was called "implement an industry standard software package".

This approach worked well (most of the time) because the same

implementation process had already been done elsewhere many,

many times.

But then................SAP, ORACLE, PeopleSoft, Baan (remember them?)

and others invented "The Incomplete Software Package". The jargon

(for any CIOs reading this) was "ERP Package".

This ERP Package had three fantastic benefits for software companies

and consultants:

(1) It was incomplete....so it needed both configuration and development

(2) It was sold as though it was a "standard software package"

(3) ....so that sales people could pretend that the cost was predictable (see above)

......when actually it was far from predictable.

See one example right now in Birmingham. A 20 million pound estimate has reached

nearly 200 million.......and it might not end there.

There you have it.....the industry destroyed "the industry standard package" so that

the sales people and the consultants could baffle people with jargon (like "ERP") and

then coin it for years!