Interesting article from IMD on why small companies have an edge over their large company counterparts in driving innovation. The point I found most interesting is that they point to psychological safety as a necessity to foster a culture of innovation. They define psychological safety as "the assurance that if you have an idea and explore it yet it fails, you wont be fired or penalized for it."
Most organizations understand that it is important to accept failure in innovative cultures. But how do you send a clear message to the organization about what type of failure is acceptable, and what type of failure is not acceptable? No one wants to reward people for not being thorough, screwing up, or being lax. Stating this fact seems obvious, but in practice it can become quite muddy. What do you do if an engineer perfects a new technology that has no relevance in the market? This is what can happen when companies try to encourage failure by creating sandboxes that can be tangibly measured. For example, each group or person gets a budget of $x to spend on exploration, or (as Google famously states) each person has 20% of their time to work on whatever is of interest to them. Yes, a random breakthrough may occur, but these metrics are meaningless if they are not tied to innovation goals.
I believe that the leadership of an organization bears the responsibility of defining the goals for value added innovation. By this I mean that they need to define the business the company is in based on what benefits they provide to the market. When this is very clearly stated, (as is usually the case in a small company) people know what challenges are set out before them. They will then find ways to cut through silos, form groups that consist of people with diverse skills, and be able to tell when one path needs to be abandoned for another that's more promising. These "failures" will be good failures because they won't look or feel like failure at all. The goal will be greater than any one execution which removes the burden of failure from any one task or series of events.
In my opinion, the only real failures happen when leaders do not take on this responsibility. They reorganize to remove silos, give up some money, free up some time, and sit back and wait for magic to happen in a vacuum. If this is what happens in your company, think about what you can do to set market driven innovation goals. Otherwise you risk rewarding real failure.