When it comes to IT, big is not beautiful

[daftandbarmy]

When it comes to IT, big is not beautiful

Agile, modern technology can transform public services and relieve taxpayers of bloated budgets
George Osborne, The Times (London)

http://www.timesonline.co.uk/tol/comment/columnists/guest_contributors/article5645288.ece


One of the central questions we face as a nation is how, in future, we can live within our means.  The next government will face the highest budget deficit in modern British history and the urgent task of delivering better value for taxpayers' money.  Information technology should be at the heart of this agenda.
In other walks of life, we have seen how IT has transformed customer service, expanded consumer choice and opened up new freedoms.  It has driven productivity improvements and efficiency gains across industries.  Unfortunately, this exciting potential has not been realised in government.
We all remember the individual failures in government IT: the National Health Service supercomputer shambles, the Student Ability Tests marking disaster, the catastrophe of the rural payments agency.
But even in the light of these failures, yesterday's report in The Times that the £100 billion that the Government spends on IT is running nearly £19 billion over budget was genuinely shocking.

It shocked me because it showed that these failures were not just a series of individual mistakes, but also that IT incompetence in government appears endemic.
That is why we have been working for some time on how we can harness the latest ideas to make government IT more efficient and effective.  Here's what we need to do.
First, government needs to stop thinking that when it comes to procuring IT systems, big is always beautiful.  The Government runs some of the biggest IT systems in the country, covering every single citizen.  Yet the very size of the projects too often makes them unmanageable and means only a handful of companies can compete to deliver them.  Not only does this lead to higher costs but it means that even when suppliers let the public sector down, the Government has few options but to keep handing them new contracts.  So how can we change that?
We need to move in the direction of what are known as “open standards” - in effect, creating a common language for government IT.  This technical change is crucial because it allows different types of software and systems to work side by side in government.
At a stroke it means big projects can be split into smaller elements, which can be delivered by different suppliers and then bolted together.  Because smaller projects are inherently less risky, this approach reduces the chance of cost overruns and opens up the procurement process to innovative start-ups.
Second, we need to follow the example of businesses all over the world and take advantage of “open source” technology.  Open source is a way of developing software so that the source code is made openly available to licensed users.  It started out as a communal philosophy but it's now mainstream.  It has been harnessed by companies such as Amazon and Bebo to enable them to keep down costs and more easily improve their products.  Amazon, for example, estimates that using open source has slashed its IT spending by a quarter.  And 20 per cent of online Europeans - including me - now use the open source Mozilla browser to surf the internet.  Unfortunately the Government is lagging far behind, with open-source suppliers all too often locked out of its contracts.
Last week the Conservative Party published an independent report by Mark Thompson of Cambridge University which sets out detailed proposals to create a level playing field for open source.  His report showed that the Government could save more than £600 million a year if it made more use of open source as part of a competitive procurement system.  That's the right way forward.  We're not saying that government should not use traditional licensed software - simply that open source should be used where it makes sense and can deliver better value for money.
Third, we need to see changes in Whitehall that fit the wider Conservative agenda of entrenching a new culture of financial discipline across government.  Astonishingly, there is no formal approval process for government IT schemes.  Too often, schemes are approved without thinking about how they will fit into wider objectives or how they will deliver value for money.  Ministers are often to blame for this, by changing project specifications at the last minute, which leads to spiralling costs.  That needs to change.
In the 1980s a unit based in the Treasury provided expertise to departments on handling complex privatisation projects.  Similar strategic leadership and advice is needed for large IT projects.
Getting value for money in IT goes hand in glove with delivering value for money more broadly.  At a political level that means setting clear goals and sticking to them.  And it means getting the whole government machine focused on value for money.
IT has the potential to transform the relationship between citizen and State, and deliver more efficient and effective public services.  Our plan will help us to make the most of this potential.  Under the Conservatives, government IT will be about getting value for money for the future, not adding to the bills of the past.

George Osborne is the shadow Chancellor in the UK. 

 

[c_canuk]

interesting, I wonder what information he bases his assumptions on?

Many small independantly developed programs will:

not integrate seamlessly, and will require additional contracts between the companies to work together which is like herding cats

will make it almost impossible to determine exactly who is at fault should there be a major flaw in the system, such as a security breach. companies will pass the buck onto other companies when there are bugs (firm A " it's B's fault", firm b - "no it's c's fault", firm c -"no it's due to  D,E and F")

will require redundant inputs wasting time and increasing the potential for human error exponentially for each redundant input. - think how many places you have to update when your address changes.

will limit the potetial for code reuse increasing cost of developing due to redundant programming.

using many small companies rather that one large company increases the likely hood that you may loose support for componets as the smaller companies fold while large companies tend to stick around longer.

will greatly increase the cost of hardware as redundant processes running in parallel will be less efficient.

will greatly reduce the security of the system as entry points and people with knowledge of the internal workings will be doubled for each company involved

Will make it more expensive to obtainmaint contracts, one large company will accept a 1 million dollar contract more readily than 100 small companies will accept 10 000 dollar contracts for 5 years maint.

the only plus to this I can see is that it will spread the IT budget moeny over a lot of small IT Firms that will be struggling over the finacial downturn, which might be what they are up to.

 

[dapaterson]

On the other hand, vendor lock-in is a real problem.  DND's data is under Microsoft's control - indeed, when DND once competed a contract for a desktop operating system it was a prerequisite that the OS support MS Office 97 - hardly "fair and open".

In addition, sunk costs drive many IT decisions, so vendors may lowball initial implementation costs, knowing companies and government are loathe to walk away from an IT infrastructure where they've sunk millions of dollars.  Thus, the startup may be a loss leader, but the ongoing maintenance (proprietary system, so limited vendors to choose from) becomes a cash cow.

Adoption of open standards for data exchange can reduce some issues, and ensure that you're not abandonning old data because the formats are no longer supported.

 

[Meridian]

In addition, sunk costs drive many IT decisions, so vendors may lowball itial implementation costs, knowing companies and government are loathe to walk away from an IT infrastructure where they've sunk millions of dollars.  Thus, the startup may be a loss leader, but the ongoing maintenance (proprietary system, so limited vendors to choose from) becomes a cash cow.

Coming from the Enterprise Software Industry, I can tell you this is generally the business model the major players operate under.  Most large enterprises (public AND private) are loathe to approve large credit facilities upfront these days, and on the private sector side, this is made even worse given the financial conditions in the market today.  This is why you see such a resurgence in SaaS vendors - their whole premise is ignoring the upfront install costs and implementation / consulting benefits in favour of long-term residual commitments. 

Oracle quite commonly approaches all license sales deals from the perspective of basically offering the licenses and install costs up for free (or very near it), in favour of high maintenance fees that kick up after year 1.  They know they have a huge platform, and once you commit to it, your infrastructure and resources are all committed to it.

 

[c_canuk]

<i>On the other hand, vendor lock-in is a real problem.  DND's data is under Microsoft's control - indeed, when DND once competed a contract for a desktop operating system it was a prerequisite that the OS support MS Office 97 - hardly "fair and open".</i>

well that was a condition of DND not MS, and should DND rebid again, MS is still going to win because although Linux is free and open office is compatible with MS Office files, the cost of supporting it is many times greater as the busness desktop industry is geared around MS and Mac couldn't support the demand as it wouldn't just be DND, the whole federal gov is chained together by the treasury board now.

<i>In addition, sunk costs drive many IT decisions, so vendors may lowball initial implementation costs, knowing companies and government are loathe to walk away from an IT infrastructure where they've sunk millions of dollars.  Thus, the startup may be a loss leader, but the ongoing maintenance (proprietary system, so limited vendors to choose from) becomes a cash cow.</i>

sunk costs will be there regardless if it's one big company or many small ones, sure you could replace a componet from on small company more easily than the whole set from one big company, but keep in mind you would now have to revisit the contracts with all the other companies and redevelop the other componets to work with the replacement, or at least get the developers to work with the new guys


<i>Adoption of open standards for data exchange can reduce some issues, and ensure that you're not abandonning old data because the formats are no longer supported.</i>

mmm, I'm not convinced that Open Standards works as well as it could, there is not enforcing agency, and again you will see buck passing when standards are not adhered to, the most rigidly updated open standard I can think of is html protocol and no two browsers are completely compatible, same with WIFI.

don't get me wrong, I think it could be made to work, but not in a a government as the people making decisions and signing the cheques are not going to be the SMEs as they are in seperate disjoined departments.