How can a website cost £35m? Easily.

The BBC’s Rory Cellan-Jones clearly doesn’t read this blog. His big story this morning is on the cost associated with the BusinessLink website: much as I predicted in my immediate analysis of the COI data a fortnight ago.
Rory was casting around on Twitter yesterday for interviewees: I know my name was put forward by a few people (for which I’m most grateful), but the call never came. Instead he’s gone to Sean O’Halloran from Hoop Associates, who offers a theory about ‘a big supplier, technology driven way of thinking’.
Speaking as someone who has worked as a civil servant, as a consultant on one of these mega-projects, and now as a small supplier trying to undermine them, I can speak with some authority on this. And whilst Sean’s theory isn’t wrong, it’s a little out of focus.
The simple answer to the question ‘how can a website cost £35m?’ is – because it can.
Government’s perceptions of ‘the going rate’ for website development have increased – ironically, just as the actual cost of web development has dropped due to open source and cheap hosting. In recent years, there’s been no shame in a department paying close to £1m on its corporate website – see these PQ answers from DFID and DIUS. It naturally follows that a ‘supersite’ representing multiple departments would cost a multiple of said figures. And when they asked for the money, they got it.
As soon as big money is on the table, the big consultancies swoop – in numbers. Waves of salespeople, account managers, and business analysts, which the civil service balances out with IT managers and procurement specialists. It’s a very cosy relationship, with both sides keeping each other busy, and everyone taking home a day’s pay.
It’s not unusual never to even sit down with the people doing the actual work. Instead, you find yourself in a whirlwind of meetings, documents, meetings about documents, and documents about meetings. And then there’s the stakeholders – mustn’t forget them. All of this costs money. And none of it actually generates a single line of code.
The brutal truth is that it isn’t in the big consultancies’ interests to deliver quickly, and the civil service often doesn’t know any better. Sure, government IT always runs late and over budget, doesn’t it?
How do we break the cycle? I think the forthcoming austerity measures will help. There simply won’t be the same amount of money sloshing around the system. Departments will simply have to try other, cheaper approaches – no matter what the current contracts say. And they’ll simply have to get tougher with suppliers who fail to deliver. A few new faces will also help: Tom Steinberg, Rishi Saha, the proposed skunkworks, (Lord) Richard Allan… maybe others.
None of this excuses BusinessLink costing £35m, and not being brilliant. But that’s for another day.

25 thoughts on “How can a website cost £35m? Easily.”

  1. And don’t forget all the processes required by government IT procedures. I’ve heard horror stories about teams trying to use cloud tools, but being banned by the IT gatekeepers, meaning proprietary and expensive solutions have to go through procurement.
    Seems like there’s a lot of fault in the system, before you even get to the point of a product being delivered. Your point is well made!

  2. Part of the problem is the cost of tendering itself.
    If you are a firm that wins (for example) one out of every five contracts you bid for, that fifth contract has to absorb all the costs of the other four lost tender processes.
    Presume that runs across the entire industry, and every tender process is automatically 3-4 times more expensive than it needs to be.
    Not to say that tendering shouldn’t be abolished, but there should be a way of tightening up the tender specifications and getting down to the shortlist much sooner so that firms stop wasting money on unwinnable RFP/RFIs.

  3. Perhaps it also makes the procurers/stakeholders feel more important and powerful to manage this size of project, and deal with the brand name consultancies. Not to mention the “nobody ever got fired for buying IBM” syndrome.

  4. Mirrors my experience dealing with government agencies and databases. Small and medium providers aren’t “qualified” to tackle government projects, big providers “don’t get out of bed” for less than £24m, therefore the starting price for any database or data management system is £24m, a minimum price set by the big four database providers and associated consultancies.
    Small companies like mine can deliver robust functional systems scalable to a reasonable level adequate for at least 50% (probably more) of government uses at a cost way less than £1m. Departments are so scared of being left high and dry by small providers they opt not to risk losing say £500k but instead spend the £24m+ on a large provider who still often fails to meet the actual user requirements due to a focus on process over good old fashioned software engineering.
    @JamesFirth

  5. @Chris I refer you to a recent tweet by my fellow e-gov agitator, Paul Clarke:
    .bbpBox15918993147 {background:url(http://a3.twimg.com/profile_background_images/56003419/3743308175_1269381662_o.jpg) #0099B9;padding:20px;} p.bbpTweet{background:#fff;padding:10px 12px 10px 12px;margin:0;min-height:48px;color:#000;font-size:18px !important;line-height:22px;-moz-border-radius:5px;-webkit-border-radius:5px} p.bbpTweet span.metadata{display:block;width:100%;clear:both;margin-top:8px;padding-top:12px;height:40px;border-top:1px solid #fff;border-top:1px solid #e6e6e6} p.bbpTweet span.metadata span.author{line-height:19px} p.bbpTweet span.metadata span.author img{float:left;margin:0 7px 0 0px;width:38px;height:38px} p.bbpTweet a:hover{text-decoration:underline}p.bbpTweet span.timestamp{font-size:12px;display:block}

    I’m actually taking the PRINCE2 bit *off* my CV. Sends all the wrong messages.less than a minute ago via dabr

  6. Simon
    Next time a journo asks me for an opinion on such matters I’ll send them your way instead. You’re clearly the man.
    Best wishes
    Sean

  7. As Ian Visits said, the cost of pitching to do these kinds of projects can be enormous. The procurement process of most government agencies is a nightmare these days and most definitely weighted towards the large agencies with a glut of consultants to throw at it.
    But, as Simon says – it cost that because it can. In the same way a house can cost £35million or £350,000 – at the end of the day, they’re both houses, and could even have the same number of rooms. Before anyone will have touched this project, they will have asked “what’s the budget?”. Then tried to hide their glee at the prospect of getting their hands on it!

  8. Rory has only quoted Sean in brief: so (with respect) it is impossible to tell from the article how “focussed” Sean really was in what he told Rory.
    In the end, you and Sean end up saying much the same thing: the projects are badly managed and purchasers are slack about how much they do cost.
    Rory breaks down the costs for the £35m a year spend – but they appear to include a lot of start-up costs. Rory asks a pertinent question about how those costs seemingly get repeated every year.
    The answer, again, just seems to be bad management and people getting away with charging exorbitant amounts of money without any proper financial or delivery auditing.
    And yes, a site can cost £35m (or any sum at all) “because it can” . . . but, as Rory points out, the next government website “only” cost £21m. The £35m website cost 70 percent more again than the £21m website. Your “because it can” statement really doesn’t help answer the question of why one site costs so much more than the other.
    You, Sean and Rory seem to share a similar view, but not sure your article brings much extra to the table other than some anecdotal observations. You seem to be having a go at Rory for not asking your opinion and having a go at Sean for not having exactly the same opinion as you and for working for websites that you don’t know about.
    Rory says in his article “But I freely admit my knowledge of website development is sketchy so I consulted a couple of experts.” Why moan because he didn’t contact you . . . if you knew Rory was looking for help via Twitter, why didn’t you pick up the phone and call him? You comment that Rory doesn’t read your blog, but what more can he get from your blog other than “because it can”?
    I think you probably have a lot of experience and you could do good work to build cost effective websites for the government, but in taking pot-shots at people like Rory and Sean – who are clearly on the same side as you – you’re just targeting the wrong people and diluting the core message. Personally, I think that you are bigging yourself up by chiding people who are pretty much on the same team as you. With respect, that just doesn’t help anyone. Of course, I don;t expect you to agree with me.
    I’m not having a pop at you: just pleading for people to work together.
    Would have been so much better if you had referenced Rory’s article, said nothing catty about Sean, and then gone on to use your experience to add a layer of knowledge to the subject matter.

  9. Fair enough Josh: reading that back, maybe that was a bit catty. Sorry Sean, it wasn’t intended to be. I’ll edit it.
    I’m absolutely with Rory on this one. Indeed, as I said at the top of the piece, I highlighted the same point myself a fortnight ago, when the first data appeared. I can’t explain exactly why BusinessLink cost 70% more than the second most expensive site… but I think my point still stands. Somebody somewhere allowed it to cost 70% more. The justification doesn’t really matter.
    However, for the record, I’m absolutely against Rory on his story earlier this week regarding iPhone apps. To quote something I put on Twitter (and didn’t blog, probably a mistake): ‘Cost of treating 1 lung cancer patient estimated at £12,000. NHS iPhone app cost £10k. If it stops 1 person getting cancer, we’re in profit.’

  10. I think the wrong people in government department’s commission and project manage a lot of these websites sites and without the right technical knowledge and understanding it’s easy to get ripped off or for projects to spiral out of control. The attitude that everyone knows about websites and what’s required, just because they do a bit of online shopping sadly still prevails and many departments don’t use the expertise of their existing web teams to ensure projects are delivered well and within budget.
    Hopefully the new guidelines will put pay to the idea that every campaign has to have a microsite and that this will solve all your marketing and communications problems.

  11. And yet no one has exactly what they got for their 35 million, it could be a 10 year contract including hosting and support. The project could have involved a huge amount of business change, which is actually the expensive bit. Training people how to write documents, and setting up work flows for authorisations and compliance with eGIF and alike. Making sure the site can be accessed in several languages, a requirement for some of government is to have welsh versions of their web sites, COI has Welsh, and I’d expect that a lot of their work supports other government web sites. It’s highly unlikely that it was just implementing a basic web site. Remember a million quid isn’t a lot of money, good architects, analysts, DBAs and Project Managers all cost in excess of £100,000 a year to employ, even if they only earn half of that. If you include costs of training, 1/2 a day probably costs £1000 course cost and £100 for each person on it, plus the lost opportunity costs, paying someone else to do your job whilst you are trained. Then there are the bid costs, which can exceed the costs of some small projects. Treasury, adds all of these costs on to the final bill, it may well be that less than half the upfront costs went to a big consultancy, who by the way load government bids to cover the costs of the ones they don’t win, because if they didn’t they would go bust very quickly. Most small and agile IT companies can’t afford to bid large contracts, because one lost £35 million contract would probably bankrupt them, and so sit behind the big players as niche specialists.

  12. Respect.
    A nice edit/comment: enhances your credibility, IMPO.
    Really happy for you to edit/delete my earlier post. Didn’t want to start throwing stones.
    Good point about the iPhone apps. Think you’re right that the benefits can outweigh the costs. All these things need some level of effectiveness auditing.
    Good luck with your work.

  13. Most people who have anything to do with building websites will look at this contract – for £35m a year – and conclude that the government has been horribly ripped off. Some of the comments about bidding costs completely miss the point. Bidding costs are a tiny fraction of delivery costs and covering these need not put up prices a lot.
    The problem with this blog is that it says that what has happened is acceptable. Even the URL says “how-can-a-website-cost-35m-easily”. So I have to say I think the kind of steps suggested in the blog don’t measure up at all. You are not being nearly tough enough! No more business as usual. You should be outraged.
    We have to start by saying this is unacceptable. A good way of doing that would be to ban Serco from participating in government contracts awarded in the next three years and to make it clear that the same treatment will be handed down to any other contractors who manage in future to dupe credulous civil servants.
    The game for contractors needs to change from “how can I make the most money out of the government” into “how can I make money while doing a good job for the government”.
    The game for Whitehall needs to change from “how do I deliver this one contract without wrecking my career” into “how do we build an ecosystem of suppliers that is competitive as well as competent”.

  14. I fear, William, you’ve missed the gentle irony in my headline. In this post, I’ve tried to present an explanation for how it happens, not a justification for it happening. My final sentence explicitly states ‘None of this excuses BusinessLink costing £35m.’ And for what it’s worth, this blog is rather notorious in government circles for banging that particular drum on a regular basis. Have a peek at my archives.
    Banning Serco would certainly be, er, brave.
    But it would only happen if BusinessLink were to be considered an error. And judging by the quote in Rory CJ’s piece, HMRC consider it to be a roaring success. ‘HMRC said the businesslink.gov.uk site delivered benefits of over £800m a year to companies – a huge return on the £35m invested.’

  15. I still cannot accept even *if* there was a 10 year build/host/maintain contract *and* substantially complex requirements or peak load requirements that at £35m the government got value for money.
    Earlier this year I was working on a 3-year build/host/maintain contract for a complex system for an international organisation. The funding amounted to $1m over 3 years.

  16. Years ago, when I was working for a big government IT supplier, I was sent along to a meeting at DTI to discuss a new web site project. It apparently had a budget of £20m, so we were interested. I asked how they intended to spend that kind of money, and was told that at present they had no idea. So I asked why they intended to spend that kind of money, and the answer was that the government thought e-government was a good thing and had therefore decided to spend lots of money on it.
    So I went along to the meeting. There were about 25 consultants around the table, some representing potential suppliers, some representing the client. And it was quickly clear that there were no requirements. So the first phase would be a requirements-gathering exercise. Ah, some of the consultants knew some good requirements-gathering methodologies. Which one should the project use? Clearly, they had to be evaluated before one could be chosen. So there had to be a phase of evaluating requirements gathering methodologies. And of course, that phase had to be planned. So the only outcome of a meeting costing about £20K was an action for someone to write a plan for evaluating methodologies for gathering requirements for the government web-site. I think the site was the one that became BusinessLink, and after that meeting I was in no doubt that the government would have no difficulty at all in spending its £20m budget. By comparison, we had just built an online banking system for a major retail bank for under £5m.

  17. Websites of this type have to maintain content that has to be legally correct, approved by multiple civil servants annually or more often, and reworked into something the public can understand and act on, often over multiple cycles. Not like a simple throwaway comms site and not cheap. Techies seem to think stuff like this comes for free.
    And there’ll be a huge bureaucracy required to serve the governance process imposed by the owning Department. Ironically the governance is there to ensure that money is wisely spent, but ends up being one of the biggest overheads.

  18. As a freelance, I supplied content to the Business Link site.
    This is how the sub-contracting chain worked: Business Link > Serco > Sweet & Maxwell > A small ltd co content provider > Me.
    Everybody in the chain needed their cut. This is one reason it cost so much.
    But I’ve also worked as an employee of a direct sub-contractor on a different gov.uk site. The team was enormous – at least 120. There were writers and editors, chief editors, a search team, a build team, a development team. Furthermore, everyone was paid roughly twice what they would have earned in the private sector. We’re talking 40 to 50k for just uploading content.
    Despite a veritable army operating the site, major changes were behind schedule, there was extensive turf war between contractors and employees, and nobody ever mentioned, let alone thought about, the users attempting to glean information from the site.
    It wasn’t just a gravy train. It was legalised plunder of the public purse. It was piratical in scope, and shameless in execution.
    The bottom line? Salaries are where the money for Business Link went.

  19. Having to deliver to Government contracts on many occassions, I can point to the blind leading the blind on many IT projects. The more grandiose the scheme the better it feels for the Civil Servants leading the project – that usually means bigger budgets too. And there is so little focus on the “vision” – the real benefit from the programme – but lots of detail that is added in: contractual, financial, project management, governance, quality, security, safety, standards etc etc. A small company doesn’t have much hope of covering all these bases.
    These are REAL factors that make Government projects longer and harder. Plus each project is “self-contained” and doesn’t share data – that only comes along at vast extra cost.

  20. Shame on the BBC for not reading your blog !
    I always come here first, then the BBC as a poor second….

  21. I know this is slightly off topic, however, I would just like to say that as someone who has been involved with small businesses (under 500k turnover), I have consistently found the business link website to be a very useful source of information. yes it could be improved and on the face of it, no it doesn’t justify £35m.
    A small company with hardly any cashflow can access a lot of helpful tips and advice without being robbed by some overpriced consultant. I don’t know how many businesses make use of the site but useing your i-phone apps analogy of saving one person from cancer, I feel sure there are a number of businesses that have saved money, adjusted their strategy, or expanded their business after making use of the information on this site.

  22. According to the COI figures, the content accounts for (roughly) 50% of the total cost. It’s a lot of money, but content requires people to write, edit and check it; and people cost money.
    To be perfectly honest, I find the content a bit hit-and-miss. All too often, it doesn’t deliver the cut-and-dried answer I’m hoping for. Mind you, that’s probably due to the over-complexity of the tax/legal system, rather than the content’s deficiencies.

  23. Great post Simon. These folks should be shot – the procurement dept and the project managers. Meantime, the new COI open source directives ought to help speed up the process….

Comments are closed.