By Alistair Smith, Public Sector lead at Contino - a DevOps and cloud transformation consultancy.
Although digital innovation has repeatedly been taken up with enthusiasm by various governments and government departments at various stages over the past decade or so, attempts to adopt lean startup-style transformations have rarely delivered more than a series of ever-smarter websites.
What’s the difficulty?
The machinery of government means the lean startup model can never work. The service obligations of the public sector alone mean that lifting-and-shifting the concept is inappropriate, although this shouldn’t this be a massive surprise. That is not at all to say, however, that lean government should be immediately discounted.
The need for innovation is, in the first instance, more driven by the heartstrings than through a rational, logical approach. The feeling that things aren’t quite right, or that there must be a better way to work galvanises the call for innovation.
Within lean organisations this tends to involve the development of a series of, literally innovative, hypotheses on how to improve a given situation, followed by tests to rapidly evaluate them. Such experimentation and exploration is a fundamental element of successful innovation. By proving and disproving hypotheses, of course, comes failure, but also crucial knowledge of what works well and what doesn’t.
In my experience working with global enterprises and public sector departments, I have seen the importance of embracing an environment of controlled experimentation and capturing learning to iteratively improve delivery methods and cultures. If departments are looking to emulate the business models of Netflix, Amazon or Airbnb, they must do so whole-heartedly.
These businesses have been built on exploration, hypotheses and most importantly embracing learning through failure throughout their entire value chain from business strategy (or policy) through to delivery. The testing of business-driven hypotheses as a function of innovation (and the success or failure of that test) is not seen as an indication of business health.
But, ultimately, this room for experimentation is a luxury public sector departments unfortunately aren’t afforded. By the time work gets to the digital team’s desk - who are ready to ‘innovate’ - it has already been through a long, waterfall development process with legislators and policy-makers and is underpinned by waterfall assumptions and constraints.
Compared to startup strategy (‘business model’), governmental strategy (‘policy’) relies on a much longer incubation period, with feedback only focused on the service delivery end of the value chain (i.e. citizens). This means that bad ideas survive longer and good ideas aren’t fast-tracked. What’s more, given the time required for changes in policy to reach their intended users, there is also a tendency to bundle a number of changes together, increasing their remit and blurring the boundaries of their scope. This in turn slows their journey through the value chain.
Therefore any innovation that takes place in the public sector can often be a shiny digital veneer over the same old batch waterfall projects.
What Can We Do?
There is, however, a line between difficult and impossible and that is where effective public sector innovation sits. It’s hard to get right but is still something we should all be striving for and, if it’s going to be resigned to the ‘tried and failed’ basket, we should explore that failure to understand what we can learn for next time.
The crux of any fundamental improvement will revolve around a ‘shifting left’ of the scope of innovation, underpinned by proper mechanisms and feedback that ensure that any changes result in an improvement in public services for the citizen. Shifting the scope of innovation left entails including as much of the policy-creation process in the agile experimentation and exploration activities as possible, in order to make the machinery of government much more reactive to the needs of citizens.
In order to facilitate this kind of change, the way policy programmes are funded would have to be amended. Whilst the use of frameworks such as the Digital Service Standard has meant a change in the way service outcomes are funded, they are still directly attributed to outcomes. This means there must still be a tangible return on any investment discounting, or perhaps not appreciating, the value of learning from previous work. One of the most significant challenges that this creates is a tight coupling between outcome and funding, which allows for a clear association between funding and ROI, but limits the ability for teams to hypothesise and experiment.
The only way a more open policy environment can be truly effective is by creating clear feedback loops between policy and public sector services. While the digital government agenda has moved forward demonstrably, public sector innovation will only be disruptive if it is extended to the creation of policy and the journeys of citizens through the Whitehall machine.
Within the DevOps world, multi-disciplinary teams focus on delivering tightly-scoped packages of business and technical change from inception through to deployment. Public sector innovation could be best served by doing similar: focusing on bounded policy changes, testing the impact through hypothesis-driven feedback and monitoring it throughout the value chain.
The current approach to innovation in the public sector - innovating delivery by focusing only on the user end of the policy process - is insufficient. It’s too isolated from the core strategists of government: legislators and policy makers. They need to be integrated into the lean experimentation that typifies disruptive innovation.
This is a scary proposition because it means real change, involving structural changes, unknown outcomes, risk-taking and potential failure. But the public sector is struggling to innovate already, it might as well struggle usefully and in the right direction!