Organisations often say they want people to experiment. They want curiosity, initiative and new ideas. They want teams to use technology differently, challenge established processes and solve problems closer to where those problems are actually experienced. The interesting question is what happens when somebody tries.
In work and research around innovation in local government, I have kept returning to the idea of permission. Not permission in the formal sense of asking a senior person to approve every small decision, but the practical sense in which people learn what they are allowed to do from the system around them.
A team may be told to innovate while every uncertain idea is treated as a risk to be controlled. A manager may encourage experimentation while the processes surrounding the team assume that anything unfinished is a failure of planning. People may be given access to capable tools but no protected space in which to learn how those tools might actually change the work. In that environment, the official message matters less than the accumulated signals.
People become very good at reading those signals. They notice which mistakes are tolerated and which become stories. They notice whether an incomplete idea can be shown without somebody immediately asking for a business case. They notice whether trying something new creates support, curiosity and useful challenge, or simply another set of approvals. Eventually the organisation develops a practical answer to the question: who is allowed to try?
The answer is often inconsistent. Some people work for managers who create room around them. They can test an idea, show something rough and work out whether it deserves to continue. Other people in the same organisation may experience experimentation as something exceptional: possible, perhaps, but only if they can find the right sponsor, navigate the right process or reassure enough people before anything has actually been learned.
That can produce a strange kind of permission culture. From the parts of the organisation where it works, it looks like empowerment. From somewhere else, it can look like a lottery.
I have written before that permission is part of the system. The more difficult implication is that it cannot depend entirely on individual leaders being unusually comfortable with uncertainty. Good managers matter, as do trust and psychological safety, but if the ability to experiment disappears when a particular manager moves on, it was never really embedded.
Structural permission is less exciting than a charismatic leader telling people to be bold, but it is more durable. It means being clear about where experimentation is genuinely welcome and where the consequences require tighter control. It means giving people safe environments, sensible boundaries around data and users, and somewhere to ask for help before a small test becomes a larger commitment.
It also means recognising that an experiment and a service are different things. A rough prototype should not need to satisfy every condition required of something people will depend upon in production. Equally, the fact that something began as an experiment should not allow it to drift indefinitely into operational use without ownership, support, privacy and security being addressed.
The useful boundary is not really between innovation and governance. It is between different stages of responsibility, and that is where organisations can make permission visible: you can try this kind of thing here; you can use these tools with these constraints; you can involve real users once these conditions are met; if the experiment begins to matter operationally, this is the route by which it becomes supported.
None of that guarantees creativity. People still need ideas worth testing. They need time, judgement and enough confidence to expose work that may not succeed. But it removes one unnecessary uncertainty: whether the act of trying is itself a breach of expectation.
When people have to infer permission, the safest interpretation usually wins. The person who is confident enough, senior enough or well connected enough may experiment anyway. Others wait. Useful ideas remain private, or are discussed repeatedly without becoming tangible enough to test. The organisation then sees little experimentation and concludes that people are resistant to change.
Perhaps some are. But sometimes people are simply responding rationally to a system that says two different things at once: be innovative, but do not create uncertainty.
There will always be areas where caution is appropriate. Public services handle sensitive information, essential systems and decisions that affect people's lives. "Move fast and break things" is not a serious operating model for that work. The alternative is not to remove control; it is to make the route to responsible experimentation clearer than the route to avoiding it.
That feels like a more useful test of an organisation's innovation culture than whether its strategy contains the word innovation. Can somebody with an unfinished idea tell, without taking an unnecessary personal risk, what they are allowed to try next?