Monday, December 31, 2012

G-Cloud - Still A Hobby? Or More Fundamental?

It's been an intense ride for G-Cloud in 2012 - a small team, woefully under-resourced, has accomplished much:

- Two iterations of the framework in the bank and a third underway
- Hundreds of suppliers (many new to government) represented
- Its first IL2 assured services (from Memset)
- Its first IL3 accredited services (both IaaS and email from SCC)
- An internal review of frameworks that killed off several other frameworks that were in flight
- A change in leadership navigated as Chris Chant retired and Denise McDonagh took over 

Early purchases are perhaps disappointing - 95 separate purchases totalling a little under £4m (through the end of November).  Even if you factor in that many of these purchases are, anecodotally, being made at prices that are 50-90% cheaper than government has previously achieved, it is easy for critics to argue that G-Cloud is, at best, a hobby for government.

That said, the Strategic Implementation Plan for the ICT Strategy, published in March 2011, had an idea for moving it from a hobby to a fundamental part of ICT delivery in government:


This metric has rarely been commented on but it's interesting, to me, for several reasons:

1) Four Years

It suggests a roughly four year journey for cloud to become mainstream in central government.  There are no interim metrics available and, one year in, we are, any way you cut the numbers, not yet 1/4 of the way to achieving this.  In OeE days, when the target was "100% online by end 2005" we saw many departments peg their services as likely to be online by Q4 2005 - a classic "it will be all right on the night" hockey stick pattern.  Doomed to fail in other words.

What's needed are interim targets coupled with plans for their realisation.  What percentage of new ICT spend will be transitioned to public cloud computing services by the end of 2013? And by the end of 2014? 

Suppliers who have entered the market or who are thinking about entering it would massively benefit from seeing these targets made more granular.  Departmental customers wondering how do make the transition happen would have more choice and would be able to work with other departments, in concert, to achieve the targets.

Even more aggressively, then, where will we be by the end of June 2013?

2) New ICT Spending

As far as I know, no one has ever measured or defined what is "new" spending or ever defined what the opposite of new spending is (one assumes "maintenance" but I could be wrong).  We know that any proposals for spend in excess of £1m or £5m (depending on the category) go to Liam Maxwell's ICT Reform team for approval but it's not clear if that is new spending - it could, for instance, be £5m of disk drives to support a legacy system.

Departments that I have spoken with suggest that their maintenance spending accounts for some 65-80% of total spend.  Given the heavy spend controls, it might be even higher for some departments.  Does that mean that 20-35% of total central government spend could be classified as "new"?  That would be something like £1.2-2bn on the basis of a £6bn total (I could make up any number for that total, going as high as £13bn even).

But what if new spend turns out to be far less than that?  Or what if it turns out to be far more?  MoJ are out to market for a suite of new suppliers to manage their ICT - some of those suppliers will be new to MoJ; Does that mean that spend with those suppliers will count as new? Or because they are looking after legacy will it get counted as maintenance?  As I said, the definition of "new" needs to be made clear.

On the bright side, putting up to £2bn into public cloud over the next 4 years would certainly count as more than a hobby. 

3) What Kind of Spending

Government has at least two kinds of money - capital and operating.  Traditionally - simplifying massively - ministers like to spend capital on delivering major policy commitments.  Many departments are thus capital heavy in their spending (and all the more likely so in ICT given that buying hardware & software or developing systems usually results in having an asset on the balance sheet). 

Cloud purchases are, though, on a pay as you go basis - they are operating expenditure.  Switching £2bn from capital to operating will create quite a shift in the way government departments think about their money.  It's not clear to me that many people are planning for that shift - departmental budget settlements still look capital heavy at least for this upcoming year.

ICT moving from a drain on capital funds to a routine operating cost is a necessary consequence of the move to buying services (there will also be interesting consequences on VAT recovery).

4) Public Cloud

There's a clear statement of intent here and it is that Government isn't interested in private cloud services (or cloudwashed services as the more acerbic commentators might label them - that is existing capabilities that are rebranded as cloud).  But getting to the heart of what "public cloud" really means might be interesting - if a supplier puts a service together that is available only to the public sector and is priced to compete with pure public clouds, why would government say no to that? 

Some suppliers on the G-Cloud framework today already look to be very competitive with public cloud providers - even though they have gone through more hoops to achieve the assurance and accreditation required by central government.

Also, somewhere in the middle of government, the debate about security levels continues to rage.  A change in policy is seemingly imminent.  IL3, the most common security level in central government, will disappear apparently to be replaced by T1, a simpler level of control that should open the door to far more services that already exist.  This looks, to me, to be far from done and dusted but it is an important change that would make delivering on the cloud commitment far easier.

Not so much public cloud as public sector cloud as well.

5) 50%

If 50% of new spending is to go on public cloud, where is the other 50% to go to?  And what is the aim of moving this 50%?  Underpinning the aim, I am sure, is the desire to reduce, very substantially, the spend on maintenance - that is, the 65-80% of spend that goes on keeping the legacy ship running.  

For every £1 of new spend, what reduction should be achieved in legacy cost?

6) Central Government

The Cabinet Office only aims to control (or perhaps influence) central government of course, but it's likely that local government could make the fastest and most dramatic steps in its use of cloud (be it public or public sector cloud) mostly because they are smaller and more fleet of foot, have lower security considerations (IL2 rather than IL3) and

Local government probably spends as much on ICT as central government, although it is a far more fragmented spend (with, I believe, a much higher spend maintained in house rather than through outsourced service providers).

A commitment to move an agreed amount of spending from local government ICT budgets to public or public sector cloud would be a huge boost to the emerging UK public sector cloud marketplace. I say "UK" deliberately - this would create a chance for our truly local service providers to create new offers, expand existing capabilities, create jobs in the UK and boost the economy.

Beyond A Hobby

The G-Cloud team will doubtless be thinking about how they grow usage of their framework as they ready the third version (now released from purgatory following the endorsement of G-Cloud in the frameworks review).  They need some help, though, some of which can, I believe, be provided by addressing the points above.

Happy New Year to all readers of this blog.  I hope that 2013 brings all of you success and happiness, whether you have your head in the clouds or your feet on the ground.



No comments:

Post a Comment