As part of two of Bloomberg Philanthropies’ initiatives dedicated to helping cities harness the power of data, The City Data Alliance and What Works Cities Certification, Public Digital is helping cities to become more data-informed, digital organisations. The essence of this transformation is to shift cities from thinking of data as something they might publish simply for transparency, towards implementing what we call “Data as a Service”.

One approach that we are encouraging cities to adopt is the development of a Data Service Standard. This blog post outlines why service standards are useful and how we have adapted the UK Government Service Standard to apply it to cities.

Service standards explain what good looks like in the design, delivery and quality of a service and set expectations on ways of working. As such, they are often more like a holistic set of policies than technical standards.

Service standards have proven to be a cheap, powerful lever to change and align behaviour. They help organisations publicly demonstrate leadership in their work, reduce the complexity of their service portfolio and prevent continued expenditure on wasteful or ineffective services.

The process of developing a service standard is important in and of itself. It helps to align expectations and reach an agreed vision of what teams want to achieve and how they should function.

First developed by the UK’s Government Digital Service (GDS), service standards are now common across the world. Examples include the UK Government Service Standard, Ontario’s Digital Service Standard, California’s Digital Crisis Standard, Singapore’s Digital Service Standards, and the US Digital Services Playbook.

However, these service standards have largely focused on transactional services, where users are citizens and the “whole service” might span across multiple channels including calls, letters and texts. Data services are a little different, particularly because their users are more specialised and interact with data services differently, so we think Data Service Standards are going to look a bit different too.

Every service standard is unique because it must reflect the context, priorities and values of the team that create and implement it. Nevertheless, everyone has to start somewhere, so we have created a sample Data Service Standard to act as a springboard for cities aiming to develop their own.

We have taken the UK’s Government Service Standard as a starting point for a sample Data Service Standard. We added some points, and removed or reworded others, to make them better reflect the needs of data and analytics services. Some of the most important changes include:

• Transactional services often have an obvious purpose: for example, people need to be able to tax their car or register a business. The purpose of a data or analytics service often isn’t so clear, and without it there’s a risk of being led by what data is available rather than an impactful goal. We think teams should begin by establishing their purpose and publishing data with this in mind.

• Understanding users and their needs is equally vital for data services as for any other service. Teams developing data services need to consider not only the direct users of the service (for example, a developer using data to create an application) but also their users’ users (for example, a citizen using that application to make a decision). Teams need to consider all end users who consume and are affected by the information. Therefore, in data services, understanding users becomes more like understanding an ecosystem.

• The particular target for our Data Service Standard is a (primarily US) city context in which mayors have an important role. We want to both recognise the importance of getting the high-level buy-in of the mayor, but also reflect the risks of the politicisation of data. Neutrality is important for both trust in data and for long term sustainability of services. The point “Include democracy as a user need” is designed to reflect this.

• Data users and data communities are typically closer to service teams than users of transactional services. Additionally, data services aren’t typically ones that people have to use, but ones which they should be encouraged to use. Our standard therefore emphasises the need for data services to “Inspire data users” in the work they do.

• We know that data and analytics services can be harmful as well as beneficial, and that people are concerned about the ways in which data is collected, used and shared. As a result, our standard includes points to emphasise safe, equitable and accountable outcomes, such as “Make information accessible by everyone”, “Create a secure service that protects sensitive data” and “Be transparent and accountable”.

• Finally, one of the key problems for reusers of data are data services that make unexpected changes, such as changing filenames or formats of data from month to month, or services that go offline without warning. Unlike transactional services where changes to the design often improve the experience for users, changes to data services can be to the detriment of those who rely on them. We therefore emphasise the need to “Make the service consistent and simple to use”, and encourage caution with regards to service change through the points: “Operate a reliable service” and “Iterate and improve frequently”.

What have we missed? Are there other data service standards, or other guidance, that we should point to or reference? We would love to have your comments on the draft sample Data Service Standard, which is available here.

If you’d like to discuss any of these themes further, please get in touch with our Program Director Amanda.