back to article UK.gov: NO MORE tech deals bigger than £100m. Unless we feel like it

Tech titans that grew fat off the public sector purse should take note that IT contracts worth £100m and over are to become a thing of yesteryear - according to the government. This is one of a number of "red lines" that the Cabinet Office published today, in new regulations that are designed to loosen the grip that hefty …

COMMENTS

This topic is closed for new posts.
  1. John H Woods Silver badge

    Is this the first time that ...

    ... the lessons from 'lessons learned' have actually been learned?

    1. Anonymous Coward
      Anonymous Coward

      Re: Is this the first time that ...

      Even if it is, you can bet that any lessons will be forgotten or hidden behind that old 'Beware of the leopard' sign...

      1. Peter2 Silver badge

        Re: Is this the first time that ...

        No. Otherwise they would have learnt that a large project will indeed be split up into it's constituent components, however I rather suspect that in a couple of years we will see news stories reporting that all of those smaller chunks are awarded to the same supplier.

        1. Hollerith 1

          Re: Is this the first time that ...

          Oh yes, oh yes. I worked for an organisation that had a cap--as if a cap could stop stupid, wasteful projects from draining yet more capital--and all that happened was that huge projects got chunked into amounts that would fly under the radar. The final spend was much higher in total, I think, because all the initialising of the project and getting staff and consultants on board was repeated for every chunk, even when it tended to be the same faces. And we had the same time-delays in getting it approved, so work-hour money was wasted. The project approval board seldom twigged, because the Projects team had learned how to obfuscate by talking about benefits (jam tomorrow).

    2. Anonymous Coward
      Anonymous Coward

      Re: Is this the first time that ...

      Why do politicians need to learn so much all the time, can't they write all of their work down in a huge book and then look up what went wrong in the past before undertaking a new project?

    3. Anonymous Coward
      Anonymous Coward

      Re: Is this the first time that ...

      "... the lessons from 'lessons learned' have actually been learned?"

      Unfortunately not. I'm already on the sharp end of this and it's a nightmare.

      Let's look at two cornerstone policies. A preference for SMEs and a hard cap on contract size.

      Preferring SMEs sounds great, on paper, right? Promoting growth, helping small employers, yada yada. All nice. Except, they're SMALL. What about government is small? Today I did a public sector database migration - it's gone off without a hitch and you categorically won't be reading about it in el reg [except here, obviously]. That one database, one of hundreds in the system, has over 300 tables. I have 15 junior staffers churning out schemas and ETL code full time supporting this programme. There's another 20-odd blokes squirelled away in various locations ensuring the source and target infrastructures are in place. We've got another bunch of guys ensuring we don't spaff your data all over the internet.

      All in all there's probably about 40, 45ish full time equivalents working on this branch of the project. We stood up this team in a week. How many SMEs have a bench of literally dozens of appropriately skilled staff ready to respond on that kind of notice? None of them. By definition that's at least 10% of their staff.

      So, you end up doing what GDS have done. You contract entire SMEs. Entire companies. Then you integrate them together. The problem there is that they don't know how to work like that. Technically they're great, but organisationally they're shit, particularly when jumping into the IL3/IL5 environment that is The Government. They don't have the kit, they don't have the experience, they don't have the financial departments to manage the contracts, and it is simply stupid for the government to try and manage a dozen separate contracts for a dozen SMEs all on the same job.

      So, they end up sub-contracted through an SI. The SI sub-contracts on a cost-plus basis. SMEs are not cheap to begin with. End result? Chimera-like monstrosities with inherently inefficient contract structures.

      And the other one, a hard cap on contracts? That's just a headline grabber. Pure and simple. If a project involves more than £100m of work, it should probably offer more than £100m for the job. All you do in setting a cap is introduce risk for the SI. All they're going to do in response is hike their prices to account for this. Further you add uncertainty to the project. I know and you know and I know that you know that you are never in a million years going to refuse the contract extension - the bid process alone is a 7-figure nightmare, and the transition mid-project? Hah. Don't make me laugh. Once a team is in, it's in. Unless some uppity political type decides to intervene, and a hard cap on the contract length, regardless of project size, gives them that opportunity. Even if the process of extension is painless, you're still almost certainly talking weeks of disruption as people drift off-project and back on again. The costs quickly mount regardless of what happens.

      Cabinet office should be focusing less on headline grabbing nonsense like this and more about reintroducing a culture of lean, efficient and empowered management structures. A single agile scrum team should not have a daily reporting or escalation chain that ends at the senior civil servant/cabinet office level. They likewise be focusing more on product ownership. The problem isn't that these extremely large contracts go to single SIs, the problem is that contracts have since the Thatcher days not just contracted out the work, but transferred ownership of the entire system and all the associated expertise to the SI. When the contract does change hands, all the best talent are skimmed off the top to take back into the main business.

      1. Anonymous Coward
        Anonymous Coward

        Re: Is this the first time that ...

        Thanks for the long post, but you mention "stood up this team in a week". That rather implies a total lack of planning - on the part of the customer, or on the part of the supplier?

        I presume your 300-table database was only one tiny component of the system BTW?

        1. Anonymous Coward
          Anonymous Coward

          Re: Is this the first time that ...

          "That rather implies a total lack of planning - on the part of the customer, or on the part of the supplier?"

          How so? We bid for the work over the course of about 6 weeks and from the word go were into daily stand up mode within a week. Considering that involves finding physical space, moving people from all over the country to get there, building a team, contracting suppliers etc.; that's *excellent* planning on both of our parts. It more usual for public sector contracts to meander through the security issues and the politics for a good 3-4 months between a contract being signed off and the work actually starting. We would never have achieved that pace if it wasn't for the fact we're the incumbent supplier already.

          There's nothing evil about having an incumbent supplier - it concentrates and consolidates system expertise. They just need to be properly herded. Like cats.

          "I presume your 300-table database was only one tiny component of the system BTW?"

          God aye, there's another couple of hundred of those databases to go, all of them "critical" systems, all of them crippled and ancient and loaded with thousands of triggers and other monstrosities, all supposed to be rejigged and redesigned fit for the government's "digital ambition". There's months, if not years, of work here. We've got SMEs on staff, because otherwise we wouldn't have won the contract, but none of their companies could have handled this job alone. End result is we're making a tidy 8% on their extortionate day rates.

          That's bad for everyone.

          What is good, on the other hand, is that even though we're the incumbent supplier, we don't "own" this new project. If, once it's transitioned to live, the client decides they want a change, it's not a cost-for-change process exclusive to us (which is where incumbent SIs *really* make their money), it's another open, competitive tender process. Speaking as someone who does change projects rather than operational outsourcing, I am more than happy with that element of the new status quo.

          1. Frankee Llonnygog

            Re: Is this the first time that ...

            Thanks for that. I've had similar experiences with the added complication of having to deal with 'help and support' from the Government Digital service

      2. Anonymous Coward
        Anonymous Coward

        Re: Is this the first time that ...

        Good god, I thought I'd written this when I read it. Spot on.

        I work on similar large UK govt projects, circa £200M and upwards. They cost circa £200M upwards because they are:

        1. Very complicated, govt depts tend to take the throw all the requirements they think they need in. Often these requirements are contradictory, untestable, unmeasureable (e.g. "The system will perform", a concerto? To what spec? as fast as the old one? Faster, slower? We ask the question in clarifications to be told "it must perform well", oh well, we'll make a guess). I won't even talk about requirements churn and change.

        2) They have large amounts of data. This data may be collected over the last N years when N is a number up to 30. I actually had a system that the data went back to 1st world war widower pensions. The data is normally pretty rubbish, its been migrated from paper, through various other systems and then comes to us 9 or 12th hand. Most of the data is rarely used, but no one knows which bits, so we have to migrate the lot and make a decent stab at trying to work out what it really means. So we find multiple surnames, various ID's that are very similar but there's now a couple of extra fields where its clear that there was a transcription error somewhere but no one is sure what the reals field is, so lets store all the fields and make sure the system can cancel N versions of more or less the same key. Oh, the penalty for losing ANY data is unlimited liabilities. So we spend months and months agreeing how the data is transferred and what transformations we can make on it. We all agree we should have nice, clean data but sometimes its utter shite so what do you do? You send it back to the govt to check, oh and we found that around 9% of our data had errors, so we sent back over a million queries because the strict error checking on data migration through ip that many problems. Lets see, how long does it take to check a million files by hand and tell us what the right answer is? Pick a number, any number, its a massive amount of effort. So we put various heuristics into to get the number of queries down to something manageable.

        3) The govt also wants it replicated immediately across the n data centres, not a five second delay, but all the data centres working in lock sync 24/7. Oddly enough this is quite hard to do, apparently we now have no maintenance window. So we could put another large system up to allows us to take one down so we can do rolling maintenance, security upgrades, patches, updates etc. Ah, that large system consists of over a few hundred servers spread across multiple security domains. So thats a very large chunk of hardware along with the power, air conditioning, maintenance staff real estate, Think upwards of £30M.

        4) Security. Did I mention the unlimited liabilities for losing data? Its OK if the govt loses a USB stick, someone gets a slap on the wrists, we lose a single data item and we could theoretically bankrupt the company. So to ensure that we don't lose the company we audit everything to hell and back, so we can prove what we have, we can make sure we don't lose anything at all. Guess what, that isn't free.

        I could go on and on and on. But I sit at the very, very, very sharp end of govt projects. I see the people face to face and discuss, negotiate, agree (and disagree) with the various UK govt people as well as run these sodding projects. Some are very sensible, some (quite a lot to be honest) are really nice, some are very, very competent who I would recruit in a second if I was allowed to, some are utter muppets, most are demoralised, most have the spectre of the GDS sitting over them.

        Just as an aside we use SME's, what we have found is that they are usually very knowledgeable, very expensive (more than we are), they are very small and struggle even to fill one small part of the overall team. As an example, a current project I'm working on has 60 testers alone, the defn of a SME is 250 people, most SME's are 10-15 people. As the OP said, we can take the whole of an SME, add 5-8% uplift and make good money.

        One thing that wasn't mentioned was that SME's normally refuse to take the level of liability on that the big, bad, nasty SI's take on. Now the Cabinet Office has said how they are going to take on the System Integrator role and therefore all the risks there. Now since I sit on the SI side I can be accused of bias, but I think the chances of the UK govt having enough skilled and experienced PM's capable of bringing together all of this stuff successfully, keeping the risk down to something sensible, delivering on time and on budget is about the same chance of me playing for England, scoring a hat trick in the next world cup final and then immediately heading out to be an astronaut alongside marrying a super model.

        These projects are big because govt is big. They aren't running your local tennis club website, they are looking after your pension, or your childcare or your benefits. The data and the legislation behind these are decades and decades old. They are hard to work with due to the layers of IT changes, the layers of legislative updates, what do you think the impact would be if we/the govt/somebody lost your data? People would seriously suffer, pensions or benefits might not be paid, the borders would stop working, your driving license would not be sent out, your taxes might be incorrectly calculated.

        <rant on>

        The weekend coders who keep saying how easy this is have not the slightest idea what they are talking about. Look at Universal Credit, that was an 'Agile' project, GDS 'helped' there, its been a complete disaster, the only person who can't acknowledge this is IDS, how many more utter, utter cockups will there need to be before the cretins in govt sit down and actually work out the way to engage with the people who can deliver things, an inclusive model that brings SME's in, manages risk appropriately and actually delivers. We lurch from one crisis to another with successive govts pissing around because they basically haven't the slightest idea how to work to deliver. Its never the govts fault, it's always the rip-off SI's, or the civil servants.

        <rant off>

        1. IHateWearingATie

          Re: Is this the first time that ...

          Excellent rant and seconded.

          Universal Credit had been floating around as an idea for years before this government gave it a shot - the policy wonks in DWP loved it but no one who knew the legacy IT systems was stupid enough to actually try and do it (massive IT failure in 2005/6 excepted, and that was with just two IT systems being shoved together) until IDS started it.

          On the war widower pensions, I'd heard tell that some of the records from the wars being used for pensions were written in ledgers created near the front line, with covers made from oil drums. Try digitizing that as part of a standard contract!

  2. Cliff

    Good.

    Next we need project contracts that are well-scoped and don't suffer from creep.

    1. James Anderson

      Re: Good.

      Will people please stop pretending that you can define the budget and scope of a project BEFORE the project has started.

      You cannot!

      Pretending that you can has lost a lot of people a lot of money and given IT "professionals" a bad name.

      Please can we initiate a well funded "scope and design study" before engaging in the actual project to uncover the hidden gotchas and get a realistic view on the amount of work involved. At least 50% of projects would not get past this stage. Which sounds bad but not nearly as heartbreaking as watching thousands of lines of your code going through the virtual shredder because the "back end" half of the project was not, and never could be, delivered.

      1. Neil Barnes Silver badge

        Re: Good.

        It's not just IT professionals; it's project management in all fields.

        I suspect more projects have gone tits up simply because no-one sat down before it started and worked out what it was the client actually wanted than for any other reason...

        It doesn't matter whether the project is spending a hundred grand or a hundred million: if no-one knows *exactly* what they want when the project starts, the project risks are automatically through the roof; you're almost certainly doomed before you start.

        The problem is, thought, as James Anderson points out, that this up-front research costs money. If the PM budgets for a research phase it makes a small project look twice as expensive; it makes a big project more expensive, too, but big projects are meant to be expensive, right? And yet - sooner or later, the project will end up spending the necessary money anyway. It can't avoid it. Spending it sooner might perhaps keep more projects from failing by stopping them before they start.

    2. Roland6 Silver badge

      Re: Good.

      >Next we need project contracts that are well-scoped and don't suffer from creep.

      Yes these (projects that are well scoped) come about by actually doing the up-front work properly - Remember the final delivery project that AC mentions above (a single public sector database migration) that is part of a much larger migration project is only really able to go ahead once the pre-work has been done.

      Yes the overall project can be broken up, but that I suggest will result in work getting far more caveated (so I can quote a 'competitive' price) and hence we get back into the situation where the government was for many years, where practically everything became a change request and they got absolutely stung...

  3. Only me!

    Shocked

    Well someone had to say something eventually!!!

    Maybe they should bring in laws like they have for banks.....if an IT supplier thinks they are ripping off the gov to almost the pont of fraund they report themsevels to the gov. What could possibly go wrong with that?

    But I would not be shocked if they did bring in a law like that! instead of walking away from suppliers they even think are ripping them off!.

  4. John Munyard

    No more tech deals over £100m... unless GCHQ want some new tech to spy on us with...

    1. J P
      Black Helicopters

      GCHQ's new spy tech won't go through the books as a tech deal though; it'll be badged as a new underground swimming pool for No 10.

  5. Anonymous Coward
    Anonymous Coward

    This has the potential to work for well-specified contracts to deliver distinct, discrete systems up to a certain size. But these are government systems we are talking about.

    Department heads measure their power by the size of their systems and don't count lots of little ones in the same way as one big, huge mega-system. Also some department and nationwide systems tend to be large by their very nature. Large systems can often be delivered as a series of smaller systems integrating into a whole, but now as well as being well-specified distinct and discrete they have to be well-integrated and have their phased delivery integrated in a well-managed way, something that I have yet to see in a long and bloody career delivering government IT systems. You also multiply inter-company and intra-company interfaces and lines of communication which is another cause of inefficiency, friction and delay, and that's before you factor in communications with the commissioning department and whichever other departments you will need to interface with in these pan-governmental eGovernment times.

    Keep things simple and this approach will work well, but my hopes aren't high.

  6. James 100

    Promising ... in theory

    It's a nice theory, but I'm sure the government will manage to screw it up somehow in order to keep the cash flowing into the pockets of the failure-factories as it has for years - chop that billion pound fiasco into a dozen £99m pieces, so it just ends up costing a bit more for the same lousy result.

    The root problem of civil servants wanting to buy a fire-breathing monster truck when a Transit van would do the job fine will take a lot more than just a spending cap, though - more like a wholesale cultural change. Having seen the process up close, people with no understanding get to write specifications and sign the cheques: demanding that a hosted service be written in a particular language (or, on one project I worked on, demanding that the web application's appearance conform to both of two conflicting templates, since they were re-organising at the time and had no idea which one actually applied!) ... at least they eventually moved off IE 6 though. That alone was adding a hefty chunk to the development costs on one project, between the extra workarounds and the testing involved.

  7. Anonymous Coward
    Anonymous Coward

    The big problem was awarding a huge contract to one company who couldn't possibly deliver everything themselves.

    Thus resulting in a quote that included two profit margins, one for the company winning the contract and another for the other companies they would sub-contract the work to.

    1. Bilious

      Administrative megalomania

      These governmental database systems are administrative databases. The real content - whatever you say - must be rather simple, even though the entire structure is complex.

      I've seen a number of IT disasters from below - and I've seen the providers conning the bosses into requiring huge and complex systems for quite simple tasks. The systems are written for the bosses, and the lay users are always forgotten. They are let down with substandard tools for doing their jobs.Ever after they have to follow rigid and non-intuitive routines to mellow the aggressiveness and unforgiving nature of the final solution which could have been made much simpler, cheaper and better.

  8. Anonymous Coward
    Anonymous Coward

    If it weren't for the fact that I'm a tax-payer, I would be rather looking forward to the prospect of a major river road crossing contract being split up into £100M chunks. Or HS2 being let to 400 different direct suppliers.

    Presumably the Civil Service will use its legendary project management and integration skills to make a coherent whole out of the disparate component deliverables.

    1. Roland6 Silver badge

      Re: Or HS2 being let to 400 different direct suppliers.

      You are absolutely right my thoughts were also around how the government was going to handle HS2 - given it is just an IT project with a train set attached...

  9. John Smith 19 Gold badge
    Unhappy

    The usual problem is getting "A proper top."

    That is a system that can be "built out" from.

    That's tricky.

  10. SVV

    So with all those public school and Oxbridge educations....

    sat in government, it's only taken them how many decades and billions of ££s of public money wasted to realise that something needs to change?

    Next stop, reform the ideologically driven bidding process so it's not just awarded to the lowest bidder, but favour suppliers who have a track record of successful deliveries (and don't let companies who have milked you for millions whilst failing ever bid again).

    Finally, try and do some research on successful large projects in industry (they do exist you know) and learn from them. Read The Mythical Man Month. And finally start small with proofs of concept and scale up and build out as the project progresses, rather than trying to manage everything as one monolithic project from the start.

  11. Anonymous Coward
    Anonymous Coward

    It's nice...

    ...to see some comments from people who actually know what goes on for a change. But because the truth is complicated, and therefore the comments are long, I suspect they won't be read.

    I will confine myself to pointing out that unless the government procurement processes are reformed, replacing n procurements with y*n procurements will increase procurement costs by a factor of y. Guess who pays?

This topic is closed for new posts.

Other stories you might like