One of the most popular methods I ever created is ritual dissent. By happy coincidence it is also one of the most effective. It avoids bland consensus in favour of increasing the range of material scanned before a decision is taken, but depersonalises intense criticism through the ritual of turning your chair around, or better still donning a mask. I've seen everything here from commissioned costume masks from Venice to Guido Fawkes masks used by software developers. An interesting choice buy the latter who, as a tribe like to pretend they are anarchists while imposing near catastrophic order on their users. I suppose I should add a smily face to that comment but maybe not.
I'm not going to summarise the technique as I've provided the link above and will therefore assume familiarity. The reason for posting today is that some questions came in via email and I promised to answer in a public forum to share the learning.
Questions like this are always welcome by the way, and I am happy to use this blog to post responses. My comments below are thoughts not definitive statements by the way. I've very interested in other people's experiences. Especially as we have a journal special edition on Cynefin techniques next year and I am committed to 5k words per day on the book over Christmas …
Questions and my responses
If I have a process that needs improving do I get each table to work on the same task?
Yes and no to that,
- If the issue is complex then you should be getting groups to create different safe-to-fail experiments to test out how the process could be improved. In that case they would be working on the same problem, but coming up with different solutions. That type of situation is one of the most effective uses as people shameless copy or partially imitate and overall you end up with more resilient solutions. If you are doing that I suggest you use the safe-to-fail control forms that are on the web site.
- It is complicated then I would be more inclined to take a reductionist approach and break the issue into manageable chunks and give each one to a different team. But after a couple of iterations I might shift team membership around – leave the spokesperson or an observer at the table for example before the next round.
If they are working on the same task and we then move onto ritual dissent – when completed do I then collate the responses from each table spokesman to deliver a final outcome?
If its complex then the end product of each table should be a safe-to-fail experiment. As those should run in parallel they should not be integrated; the final outcome is the portfolio of actions.
It is complicated, then yes do a conventional synthesis.