To be human is to be tasked with a deluge of decisions each day. And from the lesser, more menial ones to the larger decisions that are far more consequential, we all want to settle on and make the choices we feel are the correct ones.

For those working in the upper echelons of government — the prime minister, seniors ministers, and sometimes senior civil servants — committing to the right (read: the most beneficial and positive) decisions for the country, its environment, and its population can be a far more difficult task, however.

Such important choices shouldn’t be made without sufficient information and intelligence, nor should they be based purely on feeling. And Laura Gilbert, the Chief Analyst and Director of Data Science at 10 Downing Street, knows what these high-ranking, governmental decision-makers should have to help them land on the best decisions possible: data.

During the DIGIT Leader 2023 summit — held at Edinburgh’s Dynamic Earth on April 27th — Gilbert talked through how the 23-person-strong team she directs is bringing data-led decision-making to the heart of UK government, and how the team is working to ensure that a data-appreciative culture permeates more widely throughout government, not just its centre.

Here’s what Laura Gilbert’s insightful, interesting keynote — attended by over 300 tech leaders and audience members — covered.

Joining a historic institution — and bringing more data, tech, and innovation into it

At the beginning of her talk, Gilbert — or rather, Dr. Gilbert, seeing as she holds a Ph.D. in Particle Physics and GRID Computing from the University of Oxford (while also having a wealth of industry experience, such as serving as CTO in a medtech startup) — explained that working at 10 Downing Street was wholly unexpected.

While searching for a new opportunity in 2020 following the successful sale of the just-mentioned startup, she noticed — and subsequently became interested in — Downing Street’s Director of Data Science position that was advertised online. But, despite Gilbert’s interest, she didn’t anticipate being hired for the role. “I threw a CV at it with absolutely no expectations whatsoever. I thought, ‘That won’t happen,’ and I was very surprised to be offered the job,” she confided.

Gilbert’s work at 10DS (“No.10 Data Science”) was cut out for her from the get-go: “I showed up on the first day and I sat down with my new line manager,” she continued. “He said, ‘I’ve got one piece of advice for you.’ I said, ‘Brilliant, what is it?’” His advice, according to Gilbert, was this: “‘You’ve got six months to have an impact.’”

To make that impact — and to aid with levelling-up governmental decision-making — it was imperative to begin changing how ministers and the like perceive, access, and apply data; an effort that must’ve been deemed worthwhile and impactful, considering that the team’s efforts on this front have only ramped up since then.

Gilbert shared that 10DS spends the vast majority of its time working on data-oriented tools — such as interactive dashboards — to help with the delivery of informative, relevant data pertinent to the policies and projects being discussed internally. “80% percent of the time we spend literally building tools. So we get data, we build models, and we present those in meetings,’” she explained.

Though, Gilbert added the important caveat that the team’s job “is not to tell [decision-makers] what to decide; we definitely would flag if we think a decision is clearly not supported by evidence, but we’re not going to tell them what to do.”

Despite the expected hurdles — such as time constraints, and the availability (or lack thereof) of data — Gilbert revealed that the purpose-built data tools ship quickly; they have to. “Our turnaround time — from being asked a question, producing a dashboarding product in Python that’s live and working and ready for a ministerial meeting — is somewhere between 24 hours and seven days. That’s the kind of timeframe we’re working to,” Gilbert divulged.

Delivering interactive, intelligent, and informative data tools

So what does an example of a data tool or model built by Gilbert’s team look like, exactly? One particular example provided by Gilbert concerned a data dashboard for HS2 — the high-speed rail line linking Scotland with Manchester, Birmingham, and London (and vice versa).

With the dashboard, users can essentially see a map of the HS2 infrastructure, and then view the various direct and indirect effects — from financial impacts to travel time — that decisions around it could potentially have. “It’s going to give you an estimate of what we think is important,” Gilbert disclosed. “So: Where does it fall in the cost spectrum? When are the benefits going to be delivered to the public? It’s an idea of value for money.”

However, Gilbert acknowledged that the value for money aspect can be somewhat tenuous when it comes to transport infrastructure. “You’ve got your direct cost, you’ve got some kind of estimate of how it’s going to impact businesses along the route and in other parts of the country, it’s going to impact your access to healthcare, it’s going to impact housing prices — all sorts of things. It’s difficult to do, but we’ll have the best stab we can and come up with at least a relative idea of which options look like better value for money and, crucially, how much faster it makes people’s journeys, which is the outcome that we’re trying to hit with this,” she said.

“[HS2] is an example of where we’re turning a lot of very complicated data into something that people — who don’t have a lot of time, and perhaps not a lot of expertise of working with numbers — are going to be able to understand and weigh up the benefits, and then go ahead and make a decision,” Gilbert explained further.

Another instance of an internal, purpose-built data tool is the Net Zero Map, which provides eye-catching visualisations on the impact of potential net zero-related policy decisions. “Policy around net zero is phenomenally intricate and complicated, and it’s very expensive. We have targets that we need to hit by 2050, and there are a number of ways that we can hit those targets,” Gilbert noted.

“What [the Map] would do — if you’re looking at it in real life — is, it would let you select the policies that you’re going to fund, and it’s going to give you an idea of what that costs. Then, it’s going to show you, through time, how the flow of energy in the country moves from mostly grey — fossil fuels — to mostly colourful — renewables. You can play this through time, and use it to have conversations about […] how the costs and benefits play out,” described Gilbert.

To boot, Gilbert mentioned that the team builds “targeted interventions”: tools that won’t take too much time and effort overall, but will ultimately have a profound impact.

For example, one targeted intervention is rAPId. “rAPId is literally an API system,” she said. “We’re aware that departments aren’t sharing data; they don’t necessarily know how.” In response, a lightweight API solution was built. It provides users with a browser interface where CSV files can be dragged and dropped, thereby making it simpler for everybody — no matter their department — to both share and access the most up-to-date data.

Helping wise decisions come to the fore

If 10DS allocates around 80% of its time to building accessible, informative data tools such as the dashboard for HS2 and the Net Zero Map, what’s happening for the other 20% of the time?

According to Gilbert, the 23-person team is undertaking numerous actions to help further usher in a data-appreciate culture more widely in government, ultimately bettering governmental decision-making. “We spend our remaining time thinking very carefully, strategically, about what we can do with the effort we have available that will invoke more change,” she revealed.

As part of these efforts, the team actively shows decision-makers across departments the various tools and projects they’re working on, thereby piquing interest from ministers across government. “We always go and show people what we’re doing; we never just tell them,” stated Gilbert. “People, when they see something shiny — and there’s a reason we’re using interactive dashboards — […] go, ‘Ooo,’ and want to pick it, and they want to play with it. […] We’ll show them what we’re doing, and we’ll demonstrate what they can have.”

While it may sound relatively straightforward, the act of being intentionally visible — and open and communicative to boot — is having the desired effect. As Gilbert explained, “We drive demand, because ministers from other departments and other senior stakeholders will come and see it and go, ‘Downing Street’s got this thing that talks about our data… why’s that not in my office?’ And what you’ll find is, if a minister wants something, it happens bloody quickly. So, by targeting the very senior decision-makers who have got a vested interest in doing things better, we can roll out this technology a lot more quickly.”

Recommended

• DIGIT Leader Summit 2023 | The Making of a Leader
• University of Edinburgh Alumni leaves Google and Warns against Dangers of AI
• IBM Expects to Pause Hiring for Jobs That Could be Done by AI

What’s more, 10DS is targeting how people in government think about data and evidence, as well as how they go about — and could improve — their decision-making.

For instance, Gilbert mentioned that “We have programs that train senior civil servants, and we have people that can apply to come in and give video talks about how to think about […] data and evidence.”

And when it comes to the larger backdrop of decision-making, Gilbert stated that “We also run programmes that try and encourage people to think differently in a more general sense,” helping logical decisions come to the fore, despite the myriad problems and pressures of working in government.

As she explained further, “When you are having to make a decision very quickly — and it can be a very big decision — you are heavily biassed by what’s happening at the moment. What’s in the papers at the moment? What are your pressures from perhaps more senior ministers, or from different departments, or from the Treasury who’s controlling the money — what do they want? What do they think is the right thing to do? […] It can be very difficult to untangle the best outcome and the best way to do things from the stress that people are really under.”

One way 10DS is trying to stimulate better, more rigorous decision-making is through a strategy known as the “pre-mortem.” Essentially, the pre-mortem tasks teams to imagine the categorical failure of a project. The team then works backward to determine what could’ve led to its failure to ensure that, in reality, such actions are avoided.

In a governmental context, then, the pre-mortem can assist ministers to acknowledge any pressures and biasses that could be at play, and help them to circumvent unwise decisions. And it’s but one of the various methods that 10DS is trying to deploy to bolster governmental thought, deliberation, and decision-making.

Looking towards a data- and tech-driven future in government

During the concluding section of her keynote, Gilbert made it clear that 10DS’s efforts of bringing data-led decision-making to the heart of UK government — and more widely — has by no means reached its apex. The transformation is still very much in progress, with efforts looking to be ramped up further. “We’re really wanting to draw in a lot more data and digital expertise, and to bring in new technologies,” she divulged.

To drive these efforts forward, there’s one group of professionals that government is particularly in need of: technologists. “We want to attract more technologists to come and work in government,” Gilbert continued. “We’re on the cusp of a really big change, and we need people with skill sets and drive to come and deliver some of these programs for us and make technology change really happen.”

If you happen to be a technologist yourself, here’s a(nother) decision for you to make: Will you answer Gilbert’s call, and help drive data-led decision-making in government?