I’m currently in Stockholm for the Adopting Agile workshop/think tank (it has flavours of both). This is event number four in the series and I joined for event three in Hillsborough last year. Lots of interesting things over the last three days with new projects on creating an artisan model of learning and a complexity based approach to project management. Expect more on those as I have a fair amount to writing to do as a result and this blog is a good place to test out ideas.
In one discussion the topic of lessons learning came up in the context of an announcement by the Scrum Alliance of a new learning programme. I must admit to being a little depressed when people come up with 1990’s knowledge management solutions for lessons learnt databases given the lack of learning implied by adopting a largely failed process but never mind; it will probably generate the material for a book and a few cynics suggested that might be the prime objective!
Now getting irritated is wonderful for getting the creative juices flowing and I am came up with a key distinction between evaluative and descriptive approaches to learning systems. That put together a body of thinking over the years but in a novel format. I’ve always emphasised the value of seeing knowledge as a flow as well as a thing and I coined the phrase lessons learning over a decade ago to emphasis the former. It was also a key element of SenseMaker® signifier design which focuses on description rather than instrumental evaluation.
The reason this is important is that description is more adaptive than evaluation and also more easy to assimilate in a new context. Then we add in two linked concepts:
- Conceptual blending, the aspect of cognition that means we like to blend together multiple direct and vicarious memories to create a contextually appropriate solution.
- Getting the right level of granularity to allow things to combine and recombine in different and novel ways. Important for scaling a complex adaptive system and for exaptation; too chucked and nothing changes, too fragmented and there is no coherence.
Evaluation takes place at a project level so in traditional lessons learnt programmes the granularity is too high for exaptive and adaptive learning. Equally the way we describe the past is based on our perception of the present so there are significant problems with evaluation anyway. Successful teams describe the facts in very different ways after they know they succeeded from the way they do when the outcome is uncertain.
So a lessons learning database will comprise data from two sources:
- Fragmented observations are captured as the project progresses by all the actors using journals (one of SenseMaker® capabilities but there are other approaches)
- Timeline techniques such as Future Backwards shift to description of factual and imagined incidents within the overall flow of activity rather than attempting to create causal linkages to a perveption of success or failure.
The second gives us a fast start, as we set up the capability to do the former. In recall we can now ask questions that bring together descriptive fragments of memory without evaluation. That makes it easier to assimilate new ideas and ways of working and allows for partial copying of prior success rather then imitation. Another key aspect of working with complexity. SenseMaker® can handle both the recording and the recall, but even if you use traditional search engines and metadata approaches you can get benefit from this. Removing the judgement element also increases participation.