How do we share lessons from 'failures'?

Recently I helped teach a course on public sector innovation as part of the Australian National Institute for Public Policy (ANIPP) executive short course offering with Dr Mark Matthews (Executive Director of the HC Coombs Policy Forum and fellow public sector innovation network member). There was some great discussion and I found it very useful to hear from people trying to apply innovation on the ground and the issues they faced in doing so.

One of the topics we discussed was 'how do we share the lessons from failures?' Whenever something innovative is tried, there is going to be a real risk of the innovation not going entirely according to plan. If you could guarantee that the execution of the idea was going to be perfect it would be unlikely it was really an innovation, as there has to be a degree of novelty and change and risk, and thus a chance that something unexpected will happen or that something will go wrong. Sometimes that will result in what some describe as a 'failure'.

Empowering Change: Fostering Innovation in the Australian Public Service discussed how a reluctance to discuss failures can be a barrier to the innovation process. "If failure is not discussed and analysed, the lessons of failure are unlikely to be learned and the innovation process will remain riskier than it need be."

This is not an issue unique to the public sector. Failure in taking a new idea to market in the private sector can have significant consequences for the organisation and there can be a reluctance to share the lessons there too. People may have concerns about admitting the failure occurred or in being seen as tarnished by association with a failure.

At the Harvard Business Review blog there has been some recent discussion on the issue of failure, how to learn from it and how to do it better. Some firms, such as Google, talk about how failure is a necessary part of the innovation process. They recognise that openly identifying failures and discussing them is of benefit and helps improve the translation of ideas to action.

But in the public sector discussion of failure is still very problematic. Talk of failure can result in immediate political and media attention. Discussion of failure in the public sector can also sometimes blur the line between a mistake caused by administrative incompetence or negligence, and one, for example, where the application of a new idea met unexpected or even unforeseeable difficulties.

Of course it is important for the public sector to have scrutiny, given that a failed innovation can result in significant (potentially even fatal) consequences. Public administration affects people's lives so there needs to be careful consideration of when things go wrong or deviate from the planned outcomes.

But surely that makes it even more important for there to be free-flowing discussion of failures, and how to avoid them in future? The worst outcome of a failure is if the lessons behind why it failed are not captured and shared. If it is learnt from, then that equips others to identify and hopefully avoid potential pitfalls.

Of course the public sector does have processes in place that manage this. Audit reports, external and independent reviews, traffic light reporting, SES forums, the APS 200, workshops, informal networks - all of these mechanisms and channels can help capture and diffuse lessons from failures. But can more be usefully done?

The group discussion at the course identified a need for leadership (and its  associated ability to accept, admit and explain failures), better evaluation (including during a process when the insights are being made), and cross-agency engagement through communities of practice and networks (environments of trust).

What do you think could be done to help the public service be better at learning from its innovation experiments? In a time of increasing expectations regarding disclosure, openness and transparency from the public sector, how can public servants discuss and learn from their innovation failures?
Tags: 
Categories: