New Book Review: "The Human Side of Postmortems"

New book review for The Human Side of Postmortems: Managing Stress and Cognitive Biases, by Dave Zwieback, O'Reilly, 2013, reposted here:

The subject of postmortems should definitely get more attention in the workplace, although properly conducted agile retrospectives are more in vogue these days in the technology space because these are intended to adjust along the way rather than waiting until the end. The "retrospective prime directive" as defined by Norman L. Kerth states that "regardless of what we discover, we understand and truly believe that everyone did the best job they could, given what they knew at the time, their skills and abilities, the resources available, and the situation at hand". The problem is that some find this directive difficult to follow. In this sense, agile retrospectives can tend to have some of the same weaknesses as postmortems, although it can be argued that most would rather address project issues sooner rather than later, since waiting until the end of the project to discuss (if at all) does not address issues as they take place, leaving little room for corrective action. As a long-term technology professional, I have seen some projects avoid discussion at all costs, proving detrimental to the true personal growth of those involved.

Zwieback rightly states that many postmortem write-ups do not even hint at human factors. As he explains in his introduction, the absence of such statements are likely due, at least in part, to the social sigma associated with publicly acknowledging the contribution of human factors when things do not go as expected, resulting in many a postmortem write-up akin to docufiction. While the examples that the author provides center around outages (such as the outage of an AWS EC2 instance), which are well known, public events for all those involved, technology professionals in general will find with some consideration that relatively unknown, private events taking place throughout the course of their projects can also benefit from discussion. As the author explains, such docufiction could especially benefit in the content typically dedicated to the reason portion of the "Three Rs" (the other two being regret and remedy) from two additions: investigations into stress and cognitive biases. In other words, how can stress and cognitive biases, which form the often missing side of postmortems, be recognized, their effects mitigated?

The bulk of this short, 22-page paper offered free from O'Reilly is dedicated to a discussion of these two human factors. After discussing stress itself, a concept borrowed by social scientists from engineers in the 1930s, Zwieback discusses the impact of stress on performance, what this means from the context of simple and complex tasks, his concept of "stress surface", and how to reduce the stress surface of an event. Stress surface measures the perception of the four relative stressors: the novelty of the situation, its unpredictability, lack of control, and social evaluative threat. The author uses a modified Perceived Stress Scale to measure stress surface, which could easily be adopted for a number of project scenarios. Following his discussion on stress, the presentation turns to cognitive biases and the benefits and pitfalls of intuitive and analytical thinking, followed by a small selection of biases present in complex system outages and postmortems and how to reduce their effects: hindsight bias, outcome bias, availability bias, and other biases and misunderstandings of probability and statistics. A short, thoughtful read.