The Central Office of Information (COI), the UK Government's centre of excellence for marketing and communications, has just published a report on the costs, usability and quality of selected UK Government websites in 2009-10. It's a detailed report and the data is available to download. It shows how the UK Government spent £94m …
What was that title...?
"How can UK.gov spend £35m on a website?...Here's how..."
Only you DON'T tell us how!. You just say that that's what they spent.
If you have done a lot of in-depth research - where's the detailed costings...?
Did you rad the article ?
If you had actually read this well written article you'd have seen that the costing research was not his own, but the published results for last financial year from the Central Office of Information.
This was on the first line btw, which really proves you did not read the article.
...how can i get MY customers to spend 3 megabucks or so.
... the bulk of the cash went straight into the pockets of Guardian-reading Shoreditch-dwelling muesli-munching Nu Meeja "consultants".
Obviously as in 'it's there in the figures' or obviously as in 'it's there in your head' ?
Do we have the information about which companies benefitted from this extravagant spending?
>in that over 60 per cent of all websites are based on Apache
Seriously... you should know that statistics mean nothing. 60% of all websites... and probably 95% of "all" those websites are irrelevant placeholders or total statical nonsense.
Anyway.. back to our regular bashing...
Not only that - I bet it's 60% of public facing web-sites. I'd bet that the proportions would be completely different if you were to include sites for web based apps within companies. Also - how can you assume that microsites linked from a main site would be hoted on the same architecture?
Fire the whole lot o'em!
I strongly suspect the values quoted were unfairly (and impossibly so) attributed to websites because there were no other tangible assets or services to attribute the cost of 1,400 so-called communications staff in Whitehall, who mysteriously appeared from nowhere during the previous government.
There is no fixing such a wastful culture. Fire 'em all ASAP.
It's actually 3200 "press officers", not 1400.
And Prescott had his snout in the trough, too.
Standards across sites.....
.. are also a bit naff. You would think that they could condense a lot of the information in to fewer more useful sites instead of providing small bits of information on each one and making you search for everything you need........ Sounds just like the fragmented gov agencies i have to deal with over the phone that can do bits of what you need but then give you another number or transfer you somewhere else to try and complete your needs.
Very poor showing especially now that VAT returns must be done online, how long till they screw that up :(
Interesting and well written.
My only complaint is that I don't like the segue into the pitch. It's not that it seems disingenuous or anything; more that it breaks the stride of the article. Maybe stick such things in a break out box? It would separate the article from the pitch while drawing more attention to the latter (as, in this case, the author deserves - I'm assuming he's not getting paid).
Re: Good article.
Paul Walsh posted the article on his own website. We liked it and asked if we could republish. No money changed hands.
We will put a disclaimer at the bottom.
How can UK.gov spend £35m on a website?
Quite easily, if the snake-oil merchants I have to work with had anything to do with it.
<- and a curry, after a hard sell.
Ans: meetings - lots of 'em
Let's start with a weekly progress meeting. Invite the whole team, say 30 people (incl. _both_ developers). Reckon on funny money hourly charging at an extremely cheap £50/hour. That's 3 grand for a 2 hour meeting. Over a year and you've "spent" £150k without actually doing any work. Now if each team member has to attend another 2 meetings each week, you're close to half a mil'
Since your staff spend so much time in meetings they are pushed for time to do real work. So you have to bring in consultants - lets say £1k / day each. 5 of them for a year is another £1.25M.
We all know that the more people you put on a project, the longer it takes, so a project planned for 1 year now takes 2. Double all your people costs and viola! £35M down the tubes without even trying.
Having worked on large gov projects all i can say is you nailed it....
"Free & Open Source Software won't matter when a Consultancy or Outsourcing company loads up a contract with tasks requiring many person weeks of expensive billable time."
I am both filled with real anger at the way our naive gov depts are so easily plucked by wily consultancies, and real jealousy at not having the opportunity to indulge in such easy practise myself. It is so easy to imagine the hapless civil servants being fired lots of powerpoints and buzzwords, and not having a budget to manage on pain of redundancy, signing open-ended contracts for work that would take an average coder a weekend to complete.
As they said in Independence Day
President: I don't understand, where does all this come from? How do you get funding for something like this?
Julius: You don't actually think they spend $20,000 on a hammer, $30,000 on a toilet seat, do you?
Same rules apply me thinks :-)
Here's why it's so expensive.
Why does a website cost £35m? Here's the answers
1) The website is possibly the main public face of the organisation. Therefore it's important.
2) Because it is important one needs to consult
3) Because it is a consultation there are many meetings, involving many high-level civil servants, costing tens of thousands of pounds in man hours
4) Because the initial planning alone cost tens of thousands of pounds there is absolutely no way the website can come to less. People paid as well as this can't be wasting there time with projects whose budget is smaller than the incidental costs.
5) The bigger the cost, the more planning, the more planning the more blame when things go wrong is diluted into a myriad of individuals. Big is safe.
6) Because the people who want the website probably don't know what it should do or how, but they do know that they should know. As a result all the expensive planning is poor.
The outcome of this process is that people like me who have traded successfully in the private sector for years have no chance in the public sector which simply refuses to believe that a website can be built by a small team, or often, an individual.
Oh, emm, consulting..
You do realise that a consultancy that sells a "self brewed" CMS is able to earn a lot more hourly fees than one that does the right thing and adjusts an Open Source CMS to perform duty? I assume you realise that exactly those consultancies are asked to provide a solution..
Late 1990s all the various websites running on a ridiculously aged and costly set of platforms were consolidated by one man (Google for "queen's webmaster mick morgan") for, well, peanuts. The man and his staff knew what they were doing, and nobody noticed it until some net survey discovered the queen's website ran on Linux.
However, with New Labour came the Microsoft friends who had to be rescued again and again from the deep holes they dug for themselves, and the reason anything worked at all as predicted was due to doubling up on consultants to drag those idiots out of their own failures. So, in summary, everyone was happily emptying the trough for all it was worth by deliberately using inappropriate technology.
It's easy to explain a lot of things if you simply follow the money..
Flawed process, but reasonable conclusion
OK, its a short article, but the assumptions stated don't work.
Just because the site says its Apache, doesn't mean its Unix/Open source all the way, maybe its front ended by Apache cache servers but has Microsoft application servers. Maybe there is a commercial CMS on there even if its unix hosted.
How are you sure there is no CMS, because I know of a few in existence on those sites that you have NOT mentioned. Just because the typical CMS signature is not there does not mean its not CMS driven (especially if its a good/heavily customised templates)
Also some of the best sites wont tell you anything as they wont leak OS information because it provides attackers useful info. It would be nice to mention them.
That said its all a lot of money and id like to see a bit (ok a lot) more open source used.
Were the Microsoft based websites more expensive than the Open Source based ones?
Being Government-run sites, it wouldn't surprise me if each was developed by a different team who have no contact with the developers of previous sites, and that every team purchased a full-price, complete set of web development "tools" from Micro$oft. Plus, probably, a full set of "Dummies" books on how to use them.
From the article
I believe the report said no - or at least to no significant level.
"Free & Open Source Software won't matter when a Consultancy or Outsourcing company loads up a contract with tasks requiring many person weeks of expensive billable time."
True, but a lot of consulting time is also spent not on techies but also on project managers, programme managers, business analysts, enterprise architects and a whole manner of surrounding areas. These guys are often there to solve business problems, not technical ones.
"If there isn't a FOSS advantage, there's still clearly a commercial off-the-shelf (COTS) advantage."
Says who? What do you base this on? As I said earlier, a lot of consulting time goes into doing tasks which would be done regardless of whether an off-the-shelf product was used - .e.g. a business analyst will still have to figure out business needs.
"One of the main purposes of these sites (apart from serving static information pages) is to provide a portal for file download."
Is that in the report or is it an assumption?
"Commercial open source software packages such as CogniDox allow you to do this in a completely secure and flexible manner. It costs thousands of pounds, not millions, and it delivers those features out of the box. And it has competitors such as Alfresco and Nuxeo that can also do the same."
And there comes the plug. I don't know about CogniDox but tools like Alresco and Nuxeo are not the types of software that any organisation will just go and install and the costs of implementation will reflect that.
Also, SharePoint also costs "thousands of pounds, not millions." So the real cost is in the deployment and integration, training, documentation, etc.
Bad article all round (as somebody who has worked on these)
This is the perspective of somebody from outside of the process - they have no idea how or why things cost so much in Gov't IT contracts and like to put the blame of the tools and developer costs - which is so wrong. I could go into a big list of hidden costs (hidden from the public) but I won't bore you. But once you get to the end it all becomes clear. The only thing missing from the article is the ***ADVERTISEMENT*** banner, and here's me thinking I blocked ads.
Not all bad
I used to work for the web "team" (there was a whole two of us) in a local government, and we were planning on re-doing the council's website. We got quoted by one contractor for about £1m, we figured for that money we could do it all in-house and employ a team of contractors at a much lower cost.
I can see why they government can spend £4m on a national web site, the contractors know a) that the people asking for the site are stupid and b) it's the government, they can over charge by 100% and get away with it. But £35m? Like one of my friends put it; "[it's] insane and no private company would be caught dead spending that kind of money".
There's a real easy solution, it's called procurement. If you procure something more than a box of tissues, you'd better follow some well thought through rules which tell you to get x bids, and other things.
I'll guarantee if you get more than 1 bid and let them know it's gone to tender, they will *not* charge you £35M for a website. They will all lower their prices to ~£1M to get the contract, and then they'll charge you £34 in change request fees once they cocked it up.
Definition of elephant
It is easy.
You have to factor in all the costs of the project.. including uncertainties..
And if you are required to tender a propeller driven flying aircraft carrier, because they heard something about a ship. flying machines and "carrier", you have to calculate your costs.. or you won't get the contract.
Ah, but do not forget..
"[it's] insane and no private company would be caught dead spending that kind of money".
Yes, but any private company would be seen profitable by RECEIVING that kind of money. And there, my friend, lies the rub. If those who advise on a certain approach stand to profit from that approach it is *extremely* unlikely said approach will not be the most they think they can get away with.
That's the core problem: there is zero correlation between the expected effort (and costs) required and the actual bill, because those who do the advising are the only ones with the skill and insight to spot the enormous padding. Why do you think so many people retired early after the ID Card setup?
Procurement? Youa re kidding me?!?
I work in an organisation with a vast Procurement department. They know squat-all about software and their 'rules' and 'knowledge' are based on going to Achilles with key words they themselves dream up, based on no IT understanding whatsoever. Then they look for 'comfort' names, such as Microsoft, IBM, Oracle, etc., as no senior management of IT person can say these are unsuitable.
To get a really good, lean, fit-for-purpose bit of software, we have to plead and beg for different keywords, we have to go to our preferred vendors and help them shape their Achilles entry, and then turn our 'evaluation' step in the Procurement process into a pretzel.
Procurement also takes about the half-life of granite to get software bought and in the door.
I wish this was just a one-off example, but sadly, no.
Old stats bad stats
That stat on the percentage of websites which are Apache is an old one and has been debunked many times. Most of the "sites" which make up Apache's statistical lead are one-pagers, domain parks, vanity or otherwise infrequently updated or visited sites. According to the article linked below, in 2007 IIS dominated Fortune 500 sites at 50% vs. 15% for Apache. The other chart shows a 50%/30% split between Apache and IIS for "Internet-based" companies, but the methodology for that number is much murkier than the other one. For instance, Google doesn't really count as Apache in the list of servers, and Amazon and Yahoo are listed as "Unknown". In any event, the statement that "Microsoft and its partners have clearly had a strong influence over UK Government procurement decisions" is clearly bunk.
Terry Pratchett got it right in The Truth....
'Special rate for government jobs,' the dwarf muttered.
'Oh, but I wouldn't dream of paying any less that other customers,' said the Patrician.
'I wasn't going to charge you less that-'
Nice plug, if you're talking about £35k websites
This isn't an analysis. There's so much extra that goes into a project - you haven't even considered hardware costs, system redundancy and failover (it is a government website after all), staging-to-live development cycle costs, any amount of planning, bespoke integration, extra requirements that AREN'T delivered from out of the box solutions (like, er, 50% of most websites), blah blah blah.
I admit the costs are still exorbitant, but paying a few grand for your document management system isn't going to solve any of the above problems which always exist when dealing with large scale, clean, reliable web projects.
Here's a money saver for Cleggameron
Disband the COI. Do it tomorrow. Actually, scratch that, do it this afternoon.
The whole office serves no useful purpose. It was supposed to centralise communications to create efficiencies but just sprawled into a huge, stagnating, inefficient, mess.
A key role is to work with advertising and other agencies on behalf of governement departments and negotiate better prices by pooling their spend. Utterly pointless. Any (very marginal) saving is immediately offset by financing the COI bureacracy. All that's needed is an approved supplier list for advertising, websites and other big ticket items and then let the departments get on with it themselves. I'm sure the Met Police are prefectly capable of running those "Don't get your mobile phone out! Somebody will nick it!" ads all by themselves. Review supplier lists by convening and then disbanding committees on rolling two year cycles. Bargain.
I made every effort never to work on our COI account again after a meeting - which lasted over three hours - ended with a 45 minute discussion about when the next meeting should be. When we finally settled on a date, somebody who hadn't spoken through the whole debate said "I can't make that one"and they resolved to have PAs get in touch later. Timewasters. All of them.
Don't forget to visit ...
The site's down.
Newer stats better stats
Or consider data from the most recent Netcraft survey (linked below), which also says, in effect, that Apache's market share, while it declined overall, rose in the UK due to Real International Business Co. adding 1.9 million hostnames, all of which "resolve to a single IP address."
The actual Answer (if he'd looked)
The COI report contains a the breakdown of the spend of the £35m site (business link)
£6250 Strategy and Planning
£4388 Design and Build
£4661 Hositng and Infrastructure
£15229 Content Provision
£4472 Testing and Evaluation
All huge figures ill agree, but trying to flog an open source produce wont solve very much of those costs will it (im assuming the tool doesn't write content for you)
If I had my most cynical hat on, id say its just this sort of badly formulated research and "product X is the answer" approach that leads projects to go wrong in the first place.
£4.3m design and build? Who designed it for them, Norman Fucking Foster??
£4.6m hosting and infrastructure? Presumably they host this thing on solid platinum servers in a data centre housed in the Royal Penthouse Suite at President Wilson Hotel, Switzerland ($65,000 per night)
Why is BusinessLink so expensive?
So why was the businesslink website so expensive? Easy - it was developed by BT Global Services
£4.4m on testing?
That buys an awful lot of testing.
That's beyond insane
Hell, the bill for testing that site alone was more than our SMT production line cost to install *and* test. That's a piece of genuine precision engineering that's completely core to our business - and therefore really had to going properly before we could take 'delivery'.
£35 million is utterly insane for a single website - you can build a conference centre for that kind of money - That's an entire building, with complex IT systems, electrical, everything - including a website!
- Edinburgh International Conference Centre (opened 2005) cost only £38 million, Manchester Central Convention Complex (2001) was £23 million.
What the hell were they doing?
Can we fire that manager (or rather management team) who signed that contract? They're clearly incompetent.
I assume you are joking and are not that ignorant of these costs.
If somebody asked me for a ballpark figure for the infrastructure, hosting and support of a 99.999 active content govt website then £4.6M would be cheap for a 10 year contract (depends on what it's doing and what the content is). I assume that you were either joking or have no idea of large scale, high-availability systems that have to comply with 100s of compliance and security standards.
The last (medium sized) high-availability design I did had server hardware costs alone of over £3m, with another £4m for san storage, media and media libraries (data retention laws anyone). When you add the 24x7 support costs, monitoring, media handlers/drivers, security guards, 24 hour service desk, network devices AND a resilient standby site, it don't come cheap, and I haven't even added software licences.
Just getting a tool to make static content web-site design easy won't help with all of that. And has the author even heard of e-GIF?
PS. If you can build, host and support the above for less then put a bid in.
> infrastructure, hosting and support of a 99.999 active content govt website then £4.6M would be cheap for a 10 year contract
WTF should the frickin business link website require 5x9 availability? This bog standard public advisory CMS website could be hosted on a standard co-location VPS for a few thousand pounds a year including support.
We are talking about hosting a website that serves a set of leaflets advising on howto setup a business not the Inland revenue VAT mainframe.
You are part of the problem, not a part of the solution.
You've already got all that probably.
If you're in the business of providing infrastructure / hosting, then you've probably got 24*7 security staff and a building/data centre with all the relevant bits and bobs already.
(Like most companies do these days).
If you want to build a whole new building, just to house a couple of servers for a website, then yes, 4mill isn't excessive, but if you're just slapping their servers in a data centre with everyone else's, then it's daylight robbery.
If you;re building a whole new datacentre just for a single website then you should be first against the wall come the revolution, unless you're getting mega traffic, which some of these figures show quite clearly isn't the case. It sounds like they're getting about the amount of traffic that my PC in the garage could handle.
All I need is a mate with a PC in his garage and round robin DNS and a handful of UPSs and we're set to rake in 4M....
Aha, that explains..
"a data centre housed in the Royal Penthouse Suite at President Wilson Hotel, Switzerland ($65,000 per night)"
That explains the humming in the cabinet. Bast*rds, now I know why they were sniggering. A good thing I paid with a government procurement card..
Money spent unwisely
Those wondering where £35m could go on a website are assuming that you should be able to look at the final product and see where your money went.
While I certainly wouldn't defend the undoubtedly gluttonous consulting costs, I have to say that from even my limited experience of working for govt. organisations (mostly LSC but some others too) I can say with some certainty that at least 50% of that cost would have come from project over-run caused by various combinations of poor briefing, mind changing, lack of understanding, lack of comprehension, pointless decision-by-committee processes that never actually decide anything, redundant projects, duplicate projects, requirements that cease to exist mid-development, requirements that never existed in the first place but which no-one had bothered to check on, and finally jobs that manage to make it through all of that intact and simply don't get rolled out.
Amen to that
The worst part is that it also ups your own ops costs because you have to document everything to the nth degree to avoid being blamed for the failure on top of it all.
I rather enjoyed running a gov project once - I was busy rescuing a group of, well, let's not be coy, complete f*cking morons who had plonked around on a 2 year project. 8 weeks before the end, someone wakes up and realises they have no chance in hell hitting the deadline with what they had so far demonstrated as rate of "progress" (think snail on a salt train stuck in reverse), so they called in some other people to blame it on.
That was fun - either they stood back and let us work, or interfered and we'd be able to hand them the deserved blame right back. The funny thing was that the "standing back" gave us just the space we needed to fix the job - so we did in 8 weeks what a bunch of "Microsoft approved" nitwits weren't capable of doing in almost 2 years. But hey, they got paid just the same - we just made them look like the oinks they were..
I think this is what I'll remember of the Blair era: the fantastic amount of bullshit wrapped around an almost perfect vacuum of knowledge with a gentle coating of blabbering ignorance and topped with extreme vacuous arrogance, sucking the money out of the UK at a rate that genuinely belongs in the Guiness Book of Records. And people *still* voted for them, proving that nuking the educational system is a good way to go if you depend on marketing instead of skills.
But I digress. Did I mention I never liked New Labour? :-)
- +Analysis Microsoft: We're making ONE TRUE WINDOWS to rule us all
- Apple: We'll unleash OS X Yosemite beta on the MASSES on 24 July
- Pics It's Google HQ - the British one: Reg man snaps covert shots INSIDE London offices
- White? Male? You work in tech? Let us guess ... Twitter? We KNEW it!
- Apple fanbois SCREAM as update BRICKS their Macbook Airs