Friday, March 30, 2012

Posts From The Past About The Future

(Just found this draft from a couple of years go ... can't think why I didn't post it)

- the original IT outsource deals were driven by customers who thought that it was non-core, that they wouldn't be able to develop or retain the skills that they needed and that there was an opportunity to lower costs by having experts do it.  That all started late 80s and early 90s. Most big companies and governments have had their IT managed by outsiders for probably 20 years now; most small companies and local authorities have retained everything in house. 

-similarly,big projects were pushed to outsiders because of a need to transfer the risk.  Costs were supposedly fixed, scope was fixed and the problem given to someone else.  Name any recent IT project to see the fallacy of that effort. The expected risk transfer certainly didn't happen. 

- so everyone who tried it has been burned to some degree and now there are new questions about what is right and, finally, new answers

- any company starting from scratch today would own no hardware. In fact, desktops are the new company car; it used to be that you got one when you joined a company.  Now you won't, you'll have your own, it will meet some key security standards and you'll get to work. Even in government. 

- any company starting from scratch today would have no servers in house. Why would they need them? They rent them from amazon or google or rackspace. They know that even the busiest servers are only 10% busy so why pay 100% of the cost.  Servers in the office have been replaced by the equivalent of zipcar/streetcar. Someone owns them, maintains them, replaces them so that you don't have to. All you need to know is that when you want them, they're right there. 

- Any company starting from scratch today will do its best to own no software. Payroll is payroll, accounts is accounts, intranets are called Facebook or google circles, email is practically free etc

- any individual is used to doing their own it integration. Email is from google, calendar is from apple, billing is from freshbooks, friends are in Facebook, files are in Dropbox, phone is on windows 7 etc. 

So the question then is if all that is true for a company starting from scratch, when does it become true for an existing company with a big legacy estate?

Over the next 2-4 years, every company and government entity will move its easy stuff to shared facilities - email, intranet, payroll, development, test etc will all go shared. Suddenly the outsource company is making no money.  If they charge too much, the customer will look for a cheaper deal. 

The big it companies can't afford to go to their customers and say "you know that email service that we charge you £1,000/head/year for? Well, we have an idea and we'll only charge you £1/head/year for it".  Their corporate process won't allow it, their annual bonus won't allow it, their shareholders won't allow it. So the only people who will offer that are people who can afford to make the shift. Some big companies will make the shift because they know that they have to - because if they don't, they'll die. Most will be too dumb. 

From 4+ years what we have come to think of as complex, heavy duty back end apps will move to commodity apps in a shared facility.  Government only needs one place to pay money to people and one to receive it.  They only need one place that figures out how to buy paperclips. 

The outsourcers, if they are smart, will recognise that it outsourcing was always a bad idea. What should have been outsourced was business processes - people doing routine work.  Routine work can be got rid of through efficiency, through automation, through moving services online etc. 

Monday, March 26, 2012

Agile, Out Loud and Proud. Or Not. Universal Credit.

The hottest two ways of working in UK public sector today are "agile" - delivering in rapid, iterative ways - and "out loud" - talking (often incessantly) about what you are doing, why you are doing it and seeking feedback (on blogs, twitter and doubtless other places too) on both of those. Both of these approaches are plainly to be encouraged.

You can, of course, do (or be) one of these at a time, or both at the same time. The finest proponent of both together is the Government Digital Service - motto "we may not quite know where we are going,but we are going to get there fast and elicit your comments at every possible step so that, rather than doing the wrong thing faster, we are more likely to do the right thing" (mottos need not be agile where I come from). The gCloud team are certainly working out loud - louder and more controversially than perhaps any project in government has ever worked before - and they are agile too (slashing away at the bureaucracy that has encumbered government for the agile equivalent of eons).

But then there are the big projects in government. At least one of those is apparently working in an agile way, but is remarkably reticent about how things are going. I speak of Universal Credit - one of the single largest IT programmes in government (ever I believe, but certainly in today's land of austerity). Hundreds of millions will be spent on this programme before it is done and it is treading the path of less than successful forbears like Tax Credits. UC involves significant changes to the way government operates at both a central and local level - the management of benefits will be moved in both directions, substantial integration will need to be achieved with HMRC's Real Time Information (where real time doesn't quite have the definition that we usually give it), legislation is changing, it will rely on the creation and rollout of a new identity scheme to replace the ageing government gateway (that I helped put in place a dozen years ago), billions of pounds will be moved (or be prevented from moving in the case of attempted fraud) and up to 30m people will be affected (thankfully not all at once, but in a phased rollout) by changes to what they claim, how they claim it and how much they receive.

So it's big. It's so big that offshore workers will be involved (not for the first time but it's still rare). The technology is certainly big, but the operational changes are even bigger.

Agile they may be - though I am sceptical to be honest because it is hard to imagine hundreds of millions spent in an agile way, with vast teams from multiple suppliers working on the programme, complex dependencies that are very likely binary (it works or it doesn't) in each case and some monolithic systems that it will replace entirely (gradual replacement of such systems in the agile style seems unlikely). Another fact that counts against it being agile, for me anyway, is that at a conference late last year I heard a senior member of the team note that they were now 34% code complete for the core of the system. I don't think agile people count in percentages (who can do maths when you're sprinting?). Great is the enemy of good but if you want to be agile, you need to know when you can leave things out, switch track and maybe never come back to what you planned; that doesn't align with a precise figure on code complete.

But what I don't think UC is, in any way, is "out loud". And that's a shame. Stakes are high for this one as I outline above. The potential for a re-run of NPfIT is very high. Yet, there are no official (or unofficial that I can find) blogs, little activity on Twitter (the brilliant @pubstrat notwithstanding, though he comments only occasionally and even obliquely on UC) and no sign of an alpha, beta or demonstration release (the front end will be delivered by but they plainly have a lot of other things to do right now). Please correct me in the comments if I've missed something - google lists 37,000 pages with "universal credits" and 13,000 when I add the word "blog" to that phrase; nothing obvious appears in the first dozen pages.

The uncertainty over how it will all work has already caused one supplier to local government, Capita, to write to all of their customers and say that they won't be ready for the changeover from how things are to how they will be under UC. Other suppliers may be keeping their heads down or perhaps may not even know what is coming and whether they can deal with it.

UC, in my view, needs to be out loud and proud. At least twice as out loud as any other project - even when that project is - given the stakes involved and the vast amount of scepticism out there. And it needs to do it now.

Saturday, March 24, 2012

How Do Freelancers Really Spend Their Time?

There have been endless stories over the last few weeks about freelancers "dodging tax" through using limited companies.  Seth Godin published a lovely graph (via SwissMiss via APhotoEditor) and I thought, why not publish the same for a freelancer?  So here it is ... how the Daily Mail thinks their time is spent and how I imagine it is probably spent:

Seth's text goes like this:
"Part of the magic (and the risk) of the internet is that if you want to, you can use your access to tools, markets and media to go even further in the direction of the chart on the right. You can become your own booker, accountant, publicist and more. Hey, it's free! You get to keep all the money!  Of course, it also means you don't get to spend very much time at all doing what you set out to do in the first place, which is shoot pictures, or write music or coach or whatever it was.The other thing you can do is find the guts and resources to move even more to the left. Hire other people (at huge expense) to do all those things you certainly could do on your own, so you can actually do the work you were born to do. One thing to consider: finding and retaining a great salesperson is more difficult than you might think, since a great salesperson might very well contribute even more value than you do."

I think the same could be written for any freelancer in business.

Tuesday, March 20, 2012

The Missing Reinvention in Government IT

Two weeks ago I wrote about how IT in Government is being reinvented, driven by four main factors - real innovation, rejection of the SI model, relentless commoditisation and radical transparency.  There is, though, a missing reinvention, and that is what to do with applications?

There are essentially two categories of applications in government - those the customer sees and uses, and those that are used only by internal staff. Most departments have at least hundreds of the latter, some even have thousands.  Sometimes one application, Self Assessment for example, has one in each category.

The declaration of "digital by default" means that there is a looming shift where thousands of applications are going to be retired.  If you don't receive paper forms any more, then you don't need an internal application to process them, you don't need a rules engine inside the firewall to validate and verify what has been entered and you don't need armies of people processing that data.

This alone will certainly class as a total reinvention of government.  There are, though, plenty of issues ahead.  If the Government Digital Service is truly going to make a difference, it will soon grapple with transactions and, there, it will find that things really do start to get ugly.  Tinkering around with content and search is all well and good, but opening up the guts of government is where the fun really starts.  I say fun only in the sense that when the job is done, someone will look up and say "well, now, wasn't that fun!"

Departmental applications have not been built in a way that they can be easily exposed.  Rules are not always clear and available (some are indeed quite secret), breakpoints that would allow the citizen to see the progress being made generally don't exist (I can already visualise the screams as the application staff say "wait, you want a progress bar for your benefit claim?"), data is certainly dirty at best and really quite messy at worst and many of the applications tie in to dozens of other applications - all in all it is not so much the "citizen journey" as the "citizen meander down a long and complex path".  GDS may well find that the best advice that they get as they start work is "well, I wouldn't start from here".

Which, actually, may be the best way to do this half of the reinvention.  Simply start afresh.  Build a new world of applications - in the style of Egg as we would have said 10 years ago - and move people onto those, severing their links with the old world.  At some point, you get to turn off those old applications and everyone gets better service as a result.  But that sounds quite slow.

At some point I can see a decision being made about who holds the IT budget for applications.  At the moment it's all in the departments.  If it's true that the current government IT budget is about £13bn (estimates range all the way up to £20bn) and that 80% goes on BAU, then the 20% will soon be moved to Mike Bracken at GDS so that he can start the reinvention.  And then, not long afterwards, a good chunk of the other 80% (less the austerity discount that is being demanded of all departments) will go his way too.  I suspect that Aviation House isn't nearly big enough to house the team that will be needed to pull this off.  Which means that a different strategy will be called for - one that embraces the principles in the Government's own ICT strategy regarding using SME suppliers and so on.

The other half of the application reinvention is much harder to call.  What to do with all the applications that are used in house?  If all the citizen facing work goes digital then plainly there's a need for a big general ledger in the middle that counts all of the money going in and out, but what about the other systems - the ones that dispense grants, manage cases (be they criminal, probation or immigration-related), track passenger movements and so on.

The usual rallying cry for these is "shared services" but, as we've seen, everyone wants to share services provided it's their services that they're sharing and not the other way round.  The NAO's recent report on Shared Services in the Finance / Payroll / HR world show that this is still very much the problem.

The approach above - rebuilding them and migrating - isn't likely to work.  The infinite variety of business processes and the vast number of those applications would make that a very expensive job and one that would take at least a geological age.

This problem could be solved, perhaps, with the massive outsourcing that government is just starting to look at again - where, this time, it isn't the IT that goes out but the business processes.  The new providers would then be free to invest in new platforms and capabilities to make them more efficient and cheaper to operate, sharing the benefits with shareholders (be they traditional ones or the employees themselves within mutuals), with government or just keeping it all for themselves (can anyone say A4E?).

This second half is the next big problem and the reinvention that will truly transform government.  It's also the hardest reinvention.

Saturday, March 17, 2012

All Software Breaks

Using the Public Tender website just now I was shown this error message when trying to search for a tender.
I'm not sure why they think me contacting MY site administrator will help them.  Still, as I said, all software breaks, be it open source or commercial.  The trick is to spot it before your users and fix it - and, even better, anticipate the likely problem and make sure the error message is useful #fail #doublefail.  They're probably all watching the rugby.  

Wednesday, March 07, 2012

The Identity Problem

I tried to sell something on eBay the other day. I'm not a prolific user of eBay, though I've had an account for perhaps a decade. I've bought a few things but never sold anything. Crafting the description took quite some time - I wanted to be sure of a sale of course - so I worked to  figure out the right price, get a good photo and that kind of thing.

Twenty minutes after I listed it, someone had clicked "buy it now.". Rats, I thought. Too cheap. 

Five minutes later I got an email from the buyer asking for my PayPal id and some better photos - apparently mine were "blurred".  So someone wants my PayPal ID sent direct to them and some better photos … suspicious.  The eBay ID had been set up not more than 15 minutes before the email arrived.  And the name of the account came with a big fat zero results when I searched for it.   A made up name.

Plainly a scam.  Took me 90 seconds to realise that and verify my conclusion. And, contrary to eBay's excited email, I hadn't sold it, no one was happy "right now" and, oh, I hadn't pocketed the cash.

I ignored it and waited to see what happened.  Effectively, the thing I was selling was sold but I wasn't sending it to the London address I was given.  I wanted to know, though, what eBay would do about it.  If I was wrong and it was legitimate, then I'd be in trouble for not sending it.  And if it was a fraud, eBay should get in touch with me.

Five days later, eBay's Advanced Automated Anti-Fraud tools kicked in and alerted me that I was the possible victim of fraud and urged me not to send the item, or get in touch with the buyer, or provide them any more information. Five days!

It's absurd, really, that in 2012 we have not yet put in place a system for proving identities on the web - where such a proof of identity makes sense (I'm not going down the "real names" rabbit hole).  In a financial transaction, everything relies on trust.  And a higher level of trust - and so a lower level of fraud - can be achieved through such identity proving.

The Cabinet Office is pursuing this through the Identity Assurance programme, although initially for government transactions.  But, by establishing a market and seeding that market with funding (by committing that Universal Credits, in the first instance, will require such identity checking), that opens up the possibility of the private sector hooking onto that same service.  

Imagine if eBay users had an attribute that said "Trusted Identity" (along with who had verified it) ... It can't be that far off surely?