Universal access to on-line government is the "real" Digital Britain target.

We have much heart-searching as to what the 2 mbps universal service target means. The “answer” is to redefine it as “reliable, working, access to government’s on-line services by 2012” – particularly those of Defra, DWP and HMRC – to be assessed by the NAO. With the Audit Commission assessing the performance of Local Government in parallel.

Debate over broadband has been bedevilled by definitions and targets since the original    definition (2mbps symmetric) was quietly abandoned as too difficult to fudge, let alone deliver reliably. Yesterday I was sharply criticised by one of my directors for going into too much detail during a request for more information to support policy discussion in this area.

This morning I was at meeting where the participants ranged from ex-BT engineers running community fibre and wifi networks (often at a fraction of the costs publicly quoted for local access) to bemused parish councillors and farmers. The collective expertise round the table was phenomenal – but the service definitions were once again getting in the way.

I must have been worrying overnight at the criticism yesterday because the answer was suddenly obvious. Whatever the nominal availablity and speed the reality for “the final third” is, all too often, notspots and timed-out transactions. Meanwhile government wants us all on-line – to save money delivering services to us.

Reliable, rapid response does not entail simply addressing headline speeds in the first/final  mile and the data centre(s). It entails addressing all the bottlenecks in between.

Linking the ambitions of government to cut costs, by dealing with more of us on-line, to the availablity of the necessary local access (and the efficient working of all that lies between) should help concentrate the minds of those who matter most: HM Treasury.

Measuring the response times for end-over-end service should be used to “motivate” the providers “in between” to work together, not engage in blame avoidance games, citing liability avoidance contracts and service level agreements for their component of the service.

As the ambitions for delivering on-line public services grow over time, from transaction processing, through access to information to telemedicine etc, so the bandwidth needed to support this defnition of universal service will grow – from around 2 mbps (to support reliable transaction processing) by 2012, through 8 mbps (sufficient for low resolution image and video) by 2017 to over 20 mbps (for HDTV quality channels for telemedicine et al) by about 2020.   

In the mid-1990s I worked with one of the main market researchers to try to apply to IT the techniques used by the car and copier industries to measure delivered quality and reliablity as seen by the end user – so that they could identify and remove the problems causing disatisfaction, warranty claims and lost reputation. The only takers in the IT industry were some of those running market information services in the City of London – where the difference between 2 seconds and 15 seconds, let alone a minute or two, was worth millions. 

The approaches we planned then are now commonplace in those areas where customer service is taken seriously. Ofcom used a similar approach in its recent tests to measure delivered speed. The time has come to use them to check whether polticians’ ambitions for on-line services can be acheived, given the infrastructure available. The politicians can then decide whether to limit their ambitions or to encourage (preferably carrots not sticks) the necessary infrastructure improvements. This requires political not regulatory decision. The trade-off can only be resolved in Cabinet, with Tresury involvement. It cannot be done by BIS or DCMS, let alone Ofcom.