Dave Snowden - Thoughts on Project Management

Yesterday evening I attended a talk by Dave Snowden at the Project Data Analytics Community Meetup with the title “How do we use technology to augment human judgement in projects?“. Overall, a lot of ground was covered and the session created the messiest set of sketchnotes ever.

For me the main take aways were:

Large projects will inevitably fail. We try and do too much planning up front and don’t put in place systems and processes to deal with the compexity and chaos that large projects generate. We should look to adopt an iterative process with a base foundation of loose plans. We are commiting too early.
We are generally measuring the wrong things in projects. They tend to be lag indicator measuremens and therefore historic. We need to be measuring things like attitude towards risk and so on.
We tend to focus on the easy to measure the things that are easy to measure and typically they are not the things that cause project failure.
Measurement is stopping great work being done. People have to “cheat” the system to get work done. 500 Targets in hospotals is an example!
The further into the future something is, the harder it is to plan and measure (cynefin sits well here).
We need more abductive thinking in progect management generally. People, especially senior leaders are uncomfortable with this. We are very poor with ‘week signal” detection and doing something with these, 9/11 being a great example as hindsight does not lead to foresight. Induction is only appropriate if the past and future are symetric, which it often isn’t.
Inattentional blindness has a large impact. We miss opportunities as we scan with too much focus.
We are reducing are capability / capacity to run things well by standardising everything: tools, techniques, qualifications and even all agreeing to “visions”. All of these reduce cognitive diversity and that stops us “seeing” and doing differently.
Cognitive activation is helpful and a hindrance. Be aware of how we change habits via process changes as we may break people out of their way of getting metally prepared for a task.
We write down very little (10% of what we know) and much of our ideas are in narratve form and PM management systems don’t work with narrative or other rich media forms.
We are generally not good with exaptation but should allow more room for this and build into our way of working or at least allow it.
SEE > ATTEND > ACT is good, but we only see if we are expecting a lot of the time. Often we do not know or realise what us importsnt to attend to and Act is often stopped by politics or excessive rationality.
Continous data narrative and granular extraction to support decision making is currently state of the art.
We are shoehorning certain project management types into inappropriate projects. Obvious, Complicated and Chaos scenarios need different methodologies.

There were lots of more takeaways, but these are probably the core ones. I really enjoyed this stimulating session and it certainly got me thinking about how organisational transformations that I support get done (or not).