Transforming public policy for the 21st century: how the NSW Government Policy Lab learned to use modern design methods to create more user-friendly solutions

THE PIA REVIEW: PIA ANDREWS This is a Public Sector Pia Review — a series on better public sectors.

I have always been fascinated by ‘public policy’. It is a term that is used interchangeably for anything from a dot point in a speech, to legislation, to an operating manual, a political position, or even a well researched proposal to address a significant social or economic issue. Foundational (Big P) Policy can be found in the positions of the (political) government of the day, in legislation, in the constitution, or can even be developed by the public service to respond to or pre-empt complex social or economic matters. When you study ‘policy’ in Australia, you are taught the ‘Public Policy Cycle’, which presents a linear method that doesn’t quite reflect the reality of working in public policy, and throughout public sectors we have lost a common language for what we mean.

When I went to work in the NSW Government, I was delighted to be able to lead a digital government policy team, and together we transformed the unit into a modern policy lab to bring a more agile, user-centric and test-driven approach to the way we design policy. It was a great experience and the team was amazing. I thought it might be useful to share some of the lessons we learned and the benefits of transforming public policy for the 21st century.

This article draws on my own research and experience as well as blogs from the NSW Government Policy Lab (found on the Policy Lab tag on the Digital NSW blog licenced under the Creative Commons Attribution 4.0 licence). My thanks to that team and in particular to Tim de Sousa, who was a leader, collaborator and driver for the team as well as for providing comment to this article.

Defining and accessing policy

Firstly, let’s be clear on what we mean by policy versus Policy. There are many genuinely amazing policy experts across the public sector who understand this distinction well, but with so many policy teams increasingly tasked with reactive and operational tasks, it is critical to get back to basics. After we transformed our policy approach, I set the NSW Government Policy Lab this challenge and worked with them to get back to basics and define different types of policy more clearly, with feedback from many exceptional policy experts in NSW, and they developed this handy reference:

I recommend the blog post What’s in a Name? Deconstructing and Defining ‘Policy’ from the Policy Lab team for more of their thoughts on the different layers of policy.

“Big P” Policy should articulate strategic objective, direction and purpose, and should be a foundation for anything that follows, be it policy, program planning or delivery. Good foundational Policy should have a consistent, accessible and well understood direction and is critical to ensure all efforts are aimed in the same direction.

It is a useful exercise for policy teams across government to consider the categories of policy they do, and proportion of effort for each. Ideally, the policy professional and all policy users can see the Big P Policy reflected in every layer. There is probably a thin line between the top two layers for Big P Policy proposals to the government of the day, but it is worth noting that speech notes, draft correspondence and event briefs are not on this reference model.

In an ideal world, all policy would be as publicly accessible like legislation is, so that everyone could understand, measure and link their efforts (including operational policy) back to overarching Policies. In Canada, all mandatory policies are found in their “Policy Suite”. This would also help with public participation in and reform of public policy for better public outcomes.

Transforming how we create ‘Big P’ Policy

How you develop (or inherit) Policy can be incredibly varied. Often people adopt a ‘Command and Control’ model of policy development, where it is developed internally and upon completion, distributed to the masses. Sometimes there are lobbyist or market-led policies in a regulated space, and there are sometimes more participatory democractic models in which anyone can find, participate in and co-design policy. In my experience, command and control policy development is the most frequently used in government, but participatory models get the best outcomes.

The ‘public policy cycle’ developed by Peter Bridgman and Glyn Davis is considered the reference model for public policy in Australia:

In the first instance is it worth reflecting on the stages of the model, which implies the entire policy process is centrally managed and coordinated by the policy makers (rarely true), gives very little indication of who is involved, where policies originate, external factors and pressures, and how policies go from a concept to being acted upon. Even to just develop a position, resources must be allocated and the development of a policy must be thus prioritised above the development of some other policy competing for attention.

Policy development is often heavily influenced by political players and agendas. Some policies are a fait accompli — the outcome is decided and effectively handed over to the public service to simply develop and implement. The outcomes of policy development are often given to the respective Minister for consideration, who may also take the policy to Cabinet for final ratification. This means even the most evidence-based, logical, widely consulted and highly researched policy position can be overturned entirely at the behest of the Government of the day, as discussed by Dr Cosmos Howard in The Policy Cycle: a Model of Post-Machiavellian Policy Making? (The Australian Journal of Public Administration, Vol. 64, No. 3, pp3-13). This obviously creates mixed results for policy outcomes but arguably also contributes to eroding public trust in public policy. The policy cycle model does not capture nor prepare public servants for how to manage or engage effectively with this process. Arguably, the most important aspects to successful policy entrepreneurship lie outside the policy development cycle entirely, in the mapping and navigation of the unpredictable waters of stakeholders, public engagement, myriad political and other agendas, and other policy areas competing for prioritisation and limited resources.

The changing role of the public service in the 21st century is also important to consider. The proliferation of digital information and communications creates new challenges and opportunities for modern policy makers. They must now compete for influence and attention in an ever expanding and contestable market of experts, perspectives and potential policies (Howard, 2005). This is a real challenge for policy makers who used to be a trusted source of knowledge for decision makers.

The shift to open, evidence and experimentation based policy becomes a useful way to create a new policy dynamic that is not only more effective, but more trustworthy by all parties and the public.

So what would a modern public policy cycle look like?

Transforming to a Policy Lab — the journey

In the NSW Digital Government Policy team, we wanted to explore what modern policy development, of all sorts, could look like. So the Digital Government Policy team worked with the Data and Information Policy team to undergo an immersive four-week training program to build new capability in policy development, with a human-centred design and test-driven focus to problem solving. The teams were taught contemporary and innovative policy development techniques to then simultaneously apply to these policy sprints. The typical program for the week included two days of intensive training followed by three days of applying new tools and skills in the respective policy sprints. The team employed quite a different policy model to the traditional one:

In the first week, they had a crash course in human-centred design and an introduction to two test policy projects (Legislation as Code and Data Policy). They examined the policy intent for the respective projects by exploring what success would look like and mapping out the key drivers and constraints they thought would affect the policy in the future. To really understand the policy users so they went on a customer safari (direct stakeholder analysis) which helped to identify the different types of users and map out their different personas. This allowed the teams to step into their users’ shoes and examine the policy intent from their perspective, as well as understand the pains and gains that the segmented user would go through to get their job done.

Through designing policy with the user in mind and then in the room, the teams were able to identify insights about the users they didn’t know themselves. The teams interviewed many policy stakeholders and heard firsthand the experiences of existing users, which was highly inspiring as a deep dive into the facts and realities. The teams were really surprised at the take-up rate of interviews which demonstrated high user engagement. They also received lots of compliments on how refreshing it was for government to openly listen.

After a week of intensive user research, the teams had collected a large number of notes and anecdotal quotes. These were summarised with key learnings from each interview captured and displayed on several walls in the ‘war room’. They looked for significant clusters and themes, grouping them into categories. To broaden the approach and enable the next step of ideation in a more open manner, the teams formed brief insight statements by visiting each theme through a ‘How might we?’ process. This sometimes called for further research and suggestions, but also led to generating a multitude of potential solutions. To generate even more ideas, they went through an extensive brainstorming exercise to think outside of the box to legitimise the ‘How might we?’ questions. The teams then had to shortlist the best eight ideas. The next day they selected the best ideas that could add major value to our project, and decided on what to prototype as the most innovative and likely to succeed. The most important factor was to design a policy prototype that would be easy to test with users and stakeholders, focusing on the user experience.

The teams used various ways – sketches, post-its, diagrams etc, to create a storyboard for each policy prototype to evolve. Because they could see the story of how each prototype began, evolved and ended, they could see how it helped answer the question/problem identified by the teams in the previous process. The most valuable aspect in developing an idea was getting feedback so, throughout the week the teams presented ‘in-development’ prototypes to users to test and gain feedback for the validation phase.

By week three, the teams were starting to embrace uncertainty as they familiarised themselves with new tools and new ways of working. What differentiates a good idea from a great one is the ability to validate whether the idea is truly executable and will make a measurable difference. The insights gained from users after testing policy prototypes really helped define the dimensions of the prototype, from implementation to effectiveness. There were several successes (not bad after three weeks) but some ideas had to be let go, which was also good because policy design is an iterative process! Not only did they learn a lot from user testing but sometimes discovered new insights meant that they had to go back and build on existing ideas or head in a different direction entirely. They gathered learnings and revisited the policy intent to see whether the intent had changed. They then looked at how to scale the promising policy prototypes with an adaption of the popular Business Model Canvas. This helped to keep the connection to the fundamentals of design by assessing whether the policy was desirable for stakeholders, feasible for delivering and viable? They asked when would each prototype be ready and could be communicated to stakeholders and when could a decision on scale be made? Like NASA’s Technology Readiness Level (TRL) and Steve Blank’s Investment Readiness Level (IRL), they were provided with the Policy Readiness Level (PRL) to understand the journey of policy creation to scale.

In the last week, they focused on recapping on the process, the tools and the art of storytelling. So, what makes a good story? To be meaningful, your story should convince your audience, and the audience for this kind of storytelling is twofold: 1) to convince decision makers of the right move, and 2) to convince and engage the public, ideally both in the development of good policy and of the legitimacy and trustworthiness of the policy position. The teams developed storytelling canvasses and learnt about some of the hacks of good storytelling and creating those ‘a-ha’ moments for the audience. They each presented their story about the policy developed to the other team to gain feedback and insights. This was a rich and satisfying experience at the culmination of four weeks hard work that rewarded both teams and also the external participants from other agencies that joined for parts of the journey. It also gave the team their big ‘a-ha’ moment: that policy development can be more effectively done using modern design methods. It can be done more innovatively and efficiently and it can be more meaningful in addressing the needs of all users. There is still a place for the conventional policy development lifecycle, however the use of these more innovative policy development techniques with human-centred design really help to underpin the needs of all users, and not just the ones that write a submission to a green or white paper.

Was it worth it?

Well, in four short weeks a traditional policy team had a new toolkit and new identity. They moved through a lot of stages in that month, from curiosity, confusion, a little fear and finally, confident enthusiasm, to be an effective Policy Lab for the NSW Digital Government agenda (now the Department of Customer Service). The policy work done by the team is better quality and more user-friendly, and they have been able to drive key areas of new public policy development.

Evidence-based versus experimentation-based policy

All of the above was a journey of exploring the intersection of evidence-based and experimentation-based policy development. Some people see these approaches as being in opposition, but my experience is that they are necessarily complementary. Evidence-based approaches are great at identifying issues or areas for focus but do not extend to establishing great solutions. We need experimentation-based approaches to design and test novel policy positions and solutions, equitably co-designed with communities. By working with policy consumers and the communities affected, you can create sustainable and effective policy and other solutions. Evidence-based solutions often are normative but experimentation-based solutions are formative, so we need both evidence and experimentation based policy making, combined with system thinking and public engagement to make a real difference to public outcomes.

Too many people forget that they are not their users, and creating experimental policy with subject matter experts and those affected by policy as well as those who will actually need to consume and implement the policy simply gets better policy outcomes through better realisation of policy intent.

My final thought on transforming policy is that if we want people to trust our policies, services and legislation, we need to embrace open approaches to public policy development and implementation. Public discovery and access to public policies (especially Big P Policy), data, models, and traceable and accountable policy decision making are necessary pre-conditions to enable public participation in policy, and are also necessary to ensure implementation is traceable back to formative Policy direction. Fundamentally, citizen-centric practices need to expand beyond service design and into the realm of policy development.

After all, aren’t citizens the real end users of government policies?