Monday, January 31, 2011

Does IT Matter (in Government)

Why do governments (in general) still do so much IT themselves? Do the real saves in IT budgets come from simply not doing IT rather than by consolidating, rationalising, clouding or sharing? I don't mean stopping doing the action itself - I wasn't advocating that if councils stop picking up rubbish they won't need the IT scheduling application for dustcarts, to be topical. I'm wondering if government has already proven that there are plenty of people who can do IT for it and so they don't need to duplicate the capability themselves.

For instance, stamp duty on share transactions is, as far as I can tell, tax raised without government touching the transaction. There are no forms, no authentication - just a cheque from the bank or broker reflecting the value of transactions undertaken (on the buy side) times 0.5%. Finding out exactly how much is raised by this tax is difficult (the figures are usually shown along with housing stamp duty but in 2003-4, Stamp Duty raised £2.6bn from share transactions alone (the total in that year was £ Is there any IT involved in government? Perhaps a reconcilement system - a general ledger - that compares published transaction data with the amounts received from the various banks and brokers and an audit function who go and check their books. Certainly no complex IT I'd like to think (if anyone knows, feel free to let me know).

So what else might work in a similar way?

Tax Discs

In the last few years one of UK government's most successful transaction has been the online tax disc renewal. I haven't had a car for a few years but I vividly remember the hassle of getting the various bits of paper together (insurance form, car registration and MOT) and waiting in the post office line. I likewise vividly remember carrying out that transaction online for the first time - certainly it counted, for me, as a transformational online service. Two things surprised me though

1) Why was there still a tax disc at all? I wonder how many tax dodgers are stopped randomly by policeman checking windscreens versus looked up via a system with auto-number plate checking systems? Is there really a need for a coloured disc in the car window?

2) Why was government involved at all? Surely the insurance company could issue the tax disc (if it was still needed) and collect the money (the MOT company couldn't as new cars don't need an MOT for the first 3 years making the insurance company the natural choice)? Having an insurance company do it would allow people to opt for various payment methods - monthly, quarterly, annually just as they do for the insurance itself. Getting more creative, as companies (such as Norwich Union) introduce GPS tracking systems to count miles driven, it might even allow tax discs to be paid on a per mile basis

If (2) holds, then government wouldn't need to be involved at all - except, as per stamp duty, reconciling and auditing the money received from the insurance companies.


From HMRC's own figures, income tax raises £144bn and national insurance a further £95bn. That's £239bn of revenue raised via PAYE and Self Assessment - perhaps two of the most involved systems in government.


Following Patrick Carter's work on PAYE, we are now only a couple of years away from every company in the country sending its PAYE data electronically. Companies buy complex accounting systems that generate the PAYE data according to rules published by UK government. When they send it to government, that data is checked according to the same rules and the data fed into a general ledger where doubtless countless adjustments are made because of changes in NI numbers, changes of job, hirings and firings, not to mention tax credits. Is there a way for government to step away from this process?

Can government, in fact, run an App Store the way Apple runs one? That is, store where government approves applications that, say, do PAYE or Self Assessment and even provides a place where they can be easily found and reviewed? And the point of an approved app would be that government would accept the data that it sent without needing to validate it again - therefore removing the need for all of the front and back end validation software that is in use, so reducing the amount of IT involved? Government would still need to operate the general ledger and apply adjustments and reconcile, but it wouldn't need to handle the data entry and validation.

Self Assessment has somewhere between 7 and 9 million customers. Perhaps half of them send the data electronically, perhaps more this year - today is the last day to send returns in so we will have new data on volumes very soon. I imagine the bulk of taxpayers use HMRC's own software - but there are others. Given the wide variety of Self Assessment scenarios - from a pensioner earning interest, through a Welsh Mac-using Vicar on to a multi-property owning landlord to the self-employed, is there a market for targeted apps that do just the essential pieces that these people need - a market that would allow UK government to step away from providing the systems and where intermediaries (perhaps accountants, maybe banks) would spring up to navigate the somewhat arcane tax world for each of those segments and ensure that all data was provided electronically and validly?

Sunday, January 09, 2011

Does The Web Need An AppStore?

I've had a new part of my daily routine for a while now ... on top of take vitamins, clean teeth ... I have "check for updates" on iPhone and iPad. I even have a sense of regret if there's nothing new. Even the news that some favourite app has added support for Chinese and Korean is enough to create a frisson of excitement. The AppStore is truly a wondrous thing - even though I may now have to check on iPhone, iPad, MBA and iMac.

Yet, on the web, there's nothing to create that sense that there's something to look at (unless your only source of content is Google Reader). The number of sites I visit just once is a far, far higher number than the number I regularly visit - the 10 most used of which are permanently pinned open in Rockmelt (and before that were faviconised in Firefox). Yet those other sites, just once visited, are doubtless spending money upgrading, iterating, improving user experience, adding content, enhancing functionality. If I only knew.

So does the web need some kind of AppStore? Not to deliver apps particularly, but to tell me when something new has happened at a site that I've visited. Where "new" means an upgrade rather than some new content.

Of course, endless update notifications could drive me madder than Jack McMad of the planet Mad. But, judiciously used, I might return to visit those sites that hadn't enticed me to register properly - but had piqued my interest enough to have me leave something that okayed updates. If they abused the privilege then they're quickly deleted. But, if they don't, they might gain repeat business/

2011 Questions

To kick off the New Year, Barron's magazine runs a multiple choice annual quiz, this year titled "Reading the 2011 Tea Leaves". The prize is a subscription to Barron's and lunch with Andrew Bary in Manhattan (I'm guessing he's not going to pay my flight costs too in the very unlikely event that I win). Here's a sample of the questions:

4) The biggest financial surprise

a) S&P finishes with a 20% gain

b) Gold ends below $1,000 an ounce

c) Federal Reserve boosts fed-funds rate to 1% or higher, from near zero

d) Inflation lifts commodity prices by 15%

e) Treasury yields surge, 10 year note ends above 4.5%

13) What will happen in Washington

a) Obama's popularity surges

b) Sarah Palin emerges as GOP frontrunner for 2012 election

c) Courts block implementation of the health care plan

d) Strong economy adds two million payroll jobs

e) none of the above

14) What will happen abroad

a) Gold finishes above $1,700 on dollar and deficit fears

b) Afghanistan's Hamid Karzai is out as president

c) Frustrated Hillary Clinton quits as secretary of state

d) China's economy cruises and its stocks rise over 20%

e) EU monetary system collapses amid bailout crisis

f) none of the above

There are 17 questions in all. Such questions, properly focused, help you think through what you think is going to happen over the coming 12 months and establish a thesis for the year. In this case, it might tell you where you are going to put your money (or where you aren't going to put it) and, if some of the events happen, how you might react.

It occurred to me that a set of questions couched in the same way could help us think through how government IT will deal with the year ahead. I'm interested in crowd-sourcing and then publishing (and widely syndicating) a dozen or so questions so that we might see what everyone thinks will happen. A couple of example questions might be:

6) What will the first OGC gateway report published in unredacted form be

a) an historic NPfIT gate report

b) the last ID cards report

c) the gate 0 report for universal benefits

d) the gDigital report

e) Other ... please specify

9) Who will the new government CIO be, replacing John Suffolk?

a) Someone from outside government who has never worked in government

b) Someone from outside government, returning after a spell away

c) There won't be a replacement

d) An existing Cabinet Office employee

e) An existing major department CIO

f) Other ... please specify

13) What will the first true gCloud application be that gets widespread adoption (>20,000 users in government)

a) email

b) a project-based collaboration tool, based around huddle or sharepoint or equivalent

c) an equivalent to facebook that allows profiles, random knowledge connections and pan-government conversation

d) a VOIP service that saves government money when, as Sir Philip Green said, "it's phoning itself"

e) Other ... please specify

Be interested in what people think of this - and if it works, I'll aim to get the questions out before the end of the month, giving everyone a couple of weeks to reply.

Saturday, January 01, 2011

What Martha Should Say Next

Martha Lane Fox's report on the future of felt like a signing-off report to me but if it wasn't, and she stays on, here's what either Martha or perhaps the new CEO (digital) needs to grapple with:

1) Lubricate interactions with Government

The easy transactions are online already. Take up is mixed. PAYE and VAT are doubtless nearing their maximum potential as a result of Lord Carter's March 2006 report mandating the move to electronic channels by 2012 for all businesses. Self Assessment is perhaps halfway to the maximum despite the fact that something like 50% of returns are filed by accountants (who you would expect to be ensnared in the Carter recommendations too). Then there's "tax disc online" which is, I think, one of the few transformational services available - not only did it do away with the visit to the post office but it also joined up DVLA, car insurers and the MOT system so that you needed only one interaction to get your tax disc (I'm still not sure why they haven't gone the final step and removed the need for a physical tax disc but perhaps that's all about spot checks by bobbies patrolling the beat). Once you move away from those flagship services, though, I suspect that take up is much, much lower. The reason, in my view, is the high level of friction that government(s) generate.

Completing an online transaction requires too many steps. And the lack of cross-departmental trust means that those steps are repeated for every transaction. This makes for very high friction that discourages people from moving their complete set of interactions with government online. To break through this requires not only a better model for proving identity (The Government Gateway started this and G-Digital looks to be trying to complete the journey although I'm uncertain how) but also a rationalisation / simplification of the data items required by government. Years ago Andrew Stott (I am sure it was he - when he was in DWP rather than in his role as Director of Digital Engagement) told me that there were only 53 items of information needed by UK government to handle everything that was needed. 53 items isn't a lot, but it is more than I would have thought (once I list out name, address, last address, date of birth, tax number, national insurance number, I start to run out of things that might be needed). But it isn't hard to imagine a "page" - a secure place that you feel entirely comfortable with - holding that data where you can selectively send it to government departments that request it. Once one department has endorsed the validity of the data you gain credibility with other departments who then start to trust your data more and more. If you are paying money to government rather than receiving perhaps the trust level is higher. Back in 2001 I used to call this the "Green Shield Stamps" authentication model - the more stamps in the book, the higher the trust and the more transactions you were able to carry out. Perhaps the "page" for the data is held by someone you trust - your bank, Tesco, your home insurance company ... even Amazon?

Having the transactions available online will not do anything to drive take up until this data friction is overcome. At the same time, government will have to overcome the infrequent interaction problem - many people, me included, touch government only rarely. Personal tax once a year, council tax once a year, passport every 10 years etc. A second aspect of friction to overcome, then, is the infrequent interaction. The longer it is between interactions, the higher the friction to overcome to make the next transaction online - because you have to remember how to navigate the page, remember your userid/password and so on. Those things are second nature for a site you visit regularly - except when they carry out a redesign (after years of being entirely consistent, how flummoxed were we all by Microsoft's "ribbon interface" for instance? - but if you're only on a site once per year, you might just reach for the paperwork because overcoming the inbuilt friction is too difficult.


2) Are you in the mobile business or not?

Not long ago there was a lively debate about whether should be in the mobile application business (and if it was, for which platforms). Stefan Cz took this debate a step further and asked who, actually, should be building government services (which attracted some thoughtful comments, particularly from Steph Gray).

My own view is that, yes, UK government should be in the mobile business but only a limited number of platforms should receive direct support, perhaps the top 3 by UK market share - with some novel measures introduced to support both the chosen platforms and others. It should also be in the business, as Martha's report says, of publishing APIs (or schema formats) that allow others to access the data. It's a big ask for someone to string a set of transactions together in a reliable way - and a bigger ask for that someone to be outside of government (and so dependent on an as yet relatively undeveloped [very] rigorous change control and notification process for changes to such APIs and formats that would need to operate across government, rather than just in a few silos, such as HMRC).

It is, though, time to say what government's position is in the mobile market. Developer? Interface builder? Content provider? Open source advocate? And then for which platforms? iPhone, Android, Blackberry, Android, Symbian, Palm and Windows? Or some subset assessed through market share? And if only the top 3 as I suggest above, what happens to new entrants or new platforms?

Number 10 appears to be clear - they're developing and providing content feeds, for the iPhone:


I think cracking the mobile market is going to require some new approaches. Plainly government can provide content and can publish APIs that 3rd parties can support. At the same time, government should develop applications, make them available at no cost, and publish the code for others to make use of. On top of that, perhaps some kind of X-prize type competition should be started that provides cash prizes to those who develop well-reviewed applications. £5,000 for the top reviewed application that solves a given problem - the list of problems needing solutions to be crowd-sourced or identified by government representatives themselves. Those who take existing applications and make them work on other platforms (outside of the top 3) would be eligible for prizes too - this would perhaps help ensure that new platforms could also receive early support.

3) Commentary, Petitions and Forums are difficult - so what to use instead?

Long ago I learned that running forums, consultations and suchlike in government was an awful job. No matter whether we left moderation to the community or ran pre or post-moderation, it didn't work. We either rapidly deteriorated into Godwin's Law scenarios or no one showed up to comment. Pareto's Law neatly applies online although rather than 80% lurking and 20% participating I suspect it's more like 98%/2% - and the 2% tend to be opinionated and strong-minded inevitably. This week's announcement that "voters could get to shape laws which go before parliament" struck me, therefore, as an interesting move.

The Downing Street Petitions site, developed by the clever folks at MySociety, has been dormant since April 2010 when it was suspended ahead of the election. Some of the petitions setup had laudable goals although doubtless . Here were the top 8 as at the point of closure with over 3.5 million "votes" between them (those need not, I believe, be unique votes):


With over 12 million people watching the X Factor, 3.5m votes across 8 petitions isn't perhaps a great result. The responses to some of the more peculiar petitions suggest that there is a large degree of misinformation that is acted on too. For instance:

On the Red Arrows: "This allegation is not true. The Government has not banned the Red Arrows from the London 2012 Olympic Games. The organising committee of London 2012 will decide what to include in the Opening Ceremony and other celebration"

On the Mosque: "With respect to the proposal associated with a site near the Olympic development in Newham, we understand from Newham Council that there is no current planning permission or application for a mosque and Newham Council do not expect a planning application in the near future. "

I wonder how many of the nearly 800,000 people who supported those petitions checked back on the response?

Allowing comments can get difficult too - here's one from the Number 10 website


And one from the iTunes store reviewing Number 10's iPhone app:


Personally, I think government is better off leaving these campaigns to existing social networks. Let people who want to campaign organise that campaign on Facebook or wherever else, let them get PR through volume of traffic and interest. Making space available on a government website will guarantee coverage of the more obscure, irreverent and idiotic ideas. Government websites are also exactly where the people aren't. They're on social networks - perhaps increasingly just on Facebook, it's true. So government should look there if it wants ideas on policies (that said, I've worked with some of the smartest people I've ever known whilst in government, and I haven't noticed a shortage of ideas for policies - so perhaps one should be careful what one wishes for).

4) What have I missed?