Seizing the government data opportunity

Is data culture taking root in government?

AP: I think we’re on a good trajectory. If we gathered a group together to talk about government data a few years ago then, quite rightly, the focus would have been on data security and protection. Now, I think the focus is on utilising data effectively. We still have lots of technical, legal and administrative hurdles and constraints that we’re working with, but we are on our way.

PL: GDPR [General Data Protection Regulation] has been very helpful because it has forced the recognition that the data doesn’t belong to us in government; it belongs to the citizen. That forces conversations about how data can improve citizen outcomes. That’s a really important shift.

EC: The big gap is that most of the key opportunities to improve how the public sector works are where data flows between organisations. If we want to target resources effectively, a massive priority right now, it requires co-ordinating different teams and that means combining data sets from different organisations. We’ve seen during the COVID crisis the difficulty of getting data shared between the central government, the NHS and local governments: that’s the bit we have to nail.

MW: That’s why I like the term “data fabric” because most of the time it’s a patchwork. I think for data-sharing in government, the default should be to open unless there are strong reasons, otherwise you are putting up roadblocks.

PL: At the very beginning of the crisis, when people with clinical needs needed to very carefully shield themselves, the Vulnerable Persons Service broke down a huge number of barriers between government departments to deliver more than a million food parcels. For instance, the Department of Health and Social Care needed to share patient-level records so other government organisations could understand the broader picture on vulnerability. We used technologies, such as “data trusts” built by the Government Digital Service, and quickly unlocked a whole load of really difficult problems.

AP: I think Paul is talking about one of the best use-cases we have and we should use it as a pattern. Other use-cases have a similar feel about them, such as tackling homelessness, where you need to understand health implications, contact with the criminal justice system, geographical issues. We’re seeing, I think, a breakdown of so many of the government’s siloed operations.

EC: There was an urgent need at the beginning of lockdown to identify which families were most vulnerable and one of the indicators is whether a child receives free school meals. But if a child goes to school in a neighbouring borough, their home borough doesn’t know if they are on free school meals. We were able to put in place the information governance arrangements to share the data across London in two days, when normally it would take six months.

MW: One thing the crisis has shown us is that if you haven’t invested in people and systems to collect data and store it in a ready-to-share way, then at some stage you may have to ask frontline operational staff, who really need to be getting on with their day job, to stop doing it, to start collecting data or answering questions. They only have five days a week; how many of those days do you want them filling in data?

ST: That’s why you have to understand your data fully. Projects that may have been low burn a few months ago can become quick wins because you’ve learnt about the data. Also, the data and the team can interoperate in a good way. If you’ve held data for many years and have a lot of experience using it, you can achieve benefits you wouldn’t see if you put a new team on it.

Does government need more data scientists?

MW: Anybody who has been in an area for a number of years can be a data scientist. People get intimidated; they say it’s maths and quite complicated, but sometimes being a data scientist is just about asking the right questions. What are we trying to achieve? Is this number big in this particular context? What would happen if this went wrong? Or went right?

ST: You have people who have strong skills in analytics, they’re very good at ensuring data quality, but sometimes the timeliness of the data gets lost. But if you’re at the other end of the spectrum with people who understand real-time data, how do you make sure the quality is good enough? It’s difficult because the skillsets can be different.

AP: I think the public now expect and demand we use data effectively. Now, in this country we have a history of scrutiny and focus and making sure the right questions are being asked about data, quite rightly. So we need to be clearer about the benefits derived from the usage of data and help with a narrative for the public.

EC: We shouldn’t share every type of data and where we do, it has to be done ethically, securely and legally. But from what I’ve seen, the public’s perception of the risk and reward has shifted in favour of sharing more data.

The pace of government innovation seems to have quickened over the last ten months; can this be maintained?

ST: We’ve learnt a lot over the last few months, not just what worked, but also what didn’t. Actually we maybe learnt more from what didn’t work. We can take all this into the future and sustain progress.

MW: Another important way to keep the pace going is some central co-ordination, which doesn’t overly restrict particular departments that need to do something different.

We’ve learned a lot over the last few months, not just what worked but also what didn’t

PL: If we can continue to draw together different organisations to solve genuine, citizen-facing problems, that’s where we can end up with fantastic successes. It has forced us to think about our organisation design around data and how we can break through internal organisational and team barriers to join up data better. I think these are all positive things to take forward.