I spent yesterday in London at the Government ICT conference. An early start from Bath but worth the trouble — the conference turned out to be surprisingly good. These were the highlights for me…
Alison Pritcher of GDS opened, outlining the GDS role in government digital transformation. She talked about 5 areas of activity, largely taken from the Government Transformation Strategy (which is 12 months old today I think), all of which resonated nicely with themes that emerged more broadly during the day. The areas included: Business transformation, noting that policy makers and technologists typically need to get together sooner rather than later; Skills & capabilities, in particular the DDaT roles (more later) and the fact that departments are currently at very different stages of maturity around digital and data skills and that pay, T&Cs, etc. are also very different across government — she also noted that the GDS Academy has now had 3000+ people through it; Making better use of data, emphasising that we are beginning to see use of Machine Learning in some departments (e.g. DWP) but need to be mindful of the skills-shortage, privacy concerns, security and data ethics; and Shared Platforms, focusing particularly on the desire for greater personalisation of digital service delivery.
I’d not come across the DDaT acronym before and it took me about half the day to realise what it stood for! Everyone else clearly knew because no-one bothered to expand it. Clearly, I must have been under a rock for the last 12 months
In case you also don’t know, DDaT standards for Digital, Data and Technology, referring to the critical set of skills for transforming the delivery of government services. Really interesting, and good, to see data highlighted so prominently — this was repeated in almost all the talks during the day.
Alison closed by outlining the GDS role in EU Exit — note, definitely not ‘Brexit’! — building capabilities and capacity where necessary and providing end to end support for transitional projects and programmes. She also emphasised the UK Technology Code of Practice.
Andrew Collinge of the Greater London Authority continued the data theme, talking about the importance of data to smart city initiatives and noting the economic and social benefits that should result. He questioned whether there will ultimately be more data than we can handle but also suggested that, “unstructured, unworkable data will become a thing of the past thanks to AI”. I’m not so sure. In reference to a project to uncover rogue landlords based on insights derived from rental data across 5 London boroughs, he noted that the project had to expend 5 months of effort on data preparation, just to get to the point that any analysis could be done.
Smart cities offer the opportunity of improving the citizen’s experiences of the city but bring with them massive governance issues. City data trusts need to encourage re-use of the data they hold but also need to promote themselves directly to citizens in order to make people aware of what data is being held, by who, and to whom it is being given (particularly where there is private sector involvement). We need to understand what business models will work in this space and we need to think thru the related technology, process and people issues. In London, where data is held on a borough by borough basis, we need to avoid creating 33 separate data silos — more when you consider health data as well. This is a nationwide problem and can only be solved thru standards and schemas.
Andrew closed by saying that 95% of data work needs to focus on standards, plumbing and cleansing BEFORE anyone can think about building applications. Further, that we need to be able to take, and re-use, data from the private sector, from civic bodies and from individuals.
Dave Perry of the DVLA spoke about the work that the DVLA has done to transform itself into a modern technology business. “How do we organise for success?” he asked. After in-sourcing technical roles, he was left with the question, “How do we manage ourselves?” — a significant challenge for an organisation set up previously only to manage outsourcers. He noted that to recruit the best staff into DVLA’s DDaT roles he had to revise their pay structure to compete with the private sector. He emphasised their agility and ability to move quickly. There are now 371 engineers at DVLA, using a ‘squad’ model, as per Spotify. There is an emphasis on innovation and an API-first delivery model, which has proved to be successful and become an impressive income generator.
He also emphasised the ‘succession planning’ that they have been doing, working with primary and secondary schools in Wales to build apps based on data. Pretty long term succession planning, clearly! During Q and A he noted that his personal experience is that girls tend to lose interest in tech somewhere between 11 and 14 — and that this needs to be a priority if we really want to develop a diverse tech workforce.
After lunch, I attended a workshop led by Kate Evans and others from Ofqual, talking about the progress they have made exposing data held in an Azure ‘data lake’ to staff within the business and encouraging them to make use of PowerBI to undertaken their own analysis. This kind of ‘Data as a Service’ approach has proved successful in broadening the uptake and use of available data.
Finally, Jacky Wright of HMRC talked about the transformation being undertaken by that department. Once again, data came to the fore, alongside APIs, cloud and a service-oriented approach. Echoing earlier talks, Jacky mentioned the development of their own MHRC Academy to help bring thru staff with the right skills, emphasising the fact that data scientists are still too hard to find. Asked about wider changes in civil service workforce, particularly in tech, she said, “If you want to build a diverse workforce, you’ve got to let people see a diverse workforce”. You’ve got to give young people something to aspire to and you’ve got to show them what is possible. Inspiring stuff.
As I say, overall it was a great day. In terms of themes… the main things that struck me where the much greater emphasis on data and the skills required around data than I’ve seen elsewhere and the more general theme around nurturing talent and building a truly diverse and modern workforce. Cloud, and in particular public cloud, got relatively little airtime; but it seemed pretty clear to me that much of what was being talked about, particularly around data, ML and AI, required the underpinning of hyperscale public cloud providers like AWS and Azure.
It struck me that while GDS has had a massive impact on the first D in DDaT (digital), the second D (data) has been given less airtime (with the possible exception of ‘open data’). I think GDS, or a new GDS-like body, needs to grow this area of activity — allowing, and encouraging, departments and related bodies to share their data experiences, skills and approaches in order that government can really exploit the benefits.
There are massive challenges for government around the management and use of data, challenges that go well beyond the current focus on GDPR. It would be a shame for progress to be slower than necessary because of a lack of collaboration.