Double Loop Learning

///Double Loop Learning
Double Loop Learning 2018-02-13T12:40:22+00:00


Understand that improvement is learning

It should not be a surprise that DevOps improvement implies some kind of intervention or change to the process, and the change will be evaluated in terms of whatever improvement occurs. The evaluation that takes place adds to the knowledge of how the process really works, which in turn increases the chances that future interventions will also result in improvement. What is critical to remember is that it is a learning process, and it is crucial that improvement is arranged so that it encourages, facilitates and exploits the learning that occurs during improvement. We must as a result recognize that there is a distinction between single- and double-loop learning.

Single and double loop learning

Single Loop learning occurs when there is a repetitive and predictable link between cause and effect. Quality Assurance, for example, measures output characteristics from the development process, such as defects, adherence to requirements/acceptance criteria, etc. These can then be used to alter input conditions, such as user story/acceptance criteria quality, standards compliance, developer skill, with the intention of improving the output. Every time a development error or problem is detected, it is corrected or solved, and more is learned about the process. However, this happens without questioning or altering the underlying values and objectives of the process, which may, over time, create an unquestioning inertia that prevents it from adapting to a changing environment.

Double Loop learning, on the other hand, questions the bottom line objectives, service or even the underlying culture of the process. This type of learning implies an ability to challenge existing process expectations in a fundamental way. It seeks to re-frame competitive assumptions and remain open to any changes in the competitive environment. But being receptive to new opportunities sometimes requires abandoning existing process routines which may be difficult to achieve in practice, especially as many processes reward experience and past achievement, instead of potential at both the individual and a group level.