Review: The Fifth Discipline

The Fifth Discipline is Peter Senge's management book about building learning organizations. I first read it in the 1990s, and recently read the new edition again. Re-reading it again nearly 20 years later is definitely an experience that's different from the first time.

As a writer, The Fifth Discipline is verbose, meanders all over the place, repeats itself frequently, and name-drops obscure people that you'd never have heard of. These properties makes it a difficult and frequently frustrating read.

As a manager, however, the fifth discipline encodes some ideas about leadership that I've found nowhere else, and hammers home certain ideas in ways that not only make sense, but have you excited about putting them in place.

The central premise of the book is that human organizations are dynamic living systems which have non-linear behavior in response to events and change. This includes several properties that make leadership challenging:

• Many incentive systems improve performance in the short term but decrease performance over the long term.
• Many feedback cycles are extremely long, far beyond what humans were evolved to deal with, and exacerbate human tendencies to either blame individuals for poor performance or put in place patch after patch to try to solve problems rather than deal with an integrative approach to problem solving. In particular, who you hire, who you fire, and who you promote has performance impact on your organization measured in years, making it difficult to get better because the feedback cycle takes so long.
• Most long term solutions and systems approach to problem solving are counter-intuitive and difficult to sell to short-term oriented business cultures. 

The tool that the book uses to illustrate this is the Beer Game, developed at MIT's Sloan school of management. The structure of the game ensures that very similar outcome happens despite having  explained the rules to very smart people and having very smart people play them. The game illustrates that given a poorly structure system and organization in place, it doesn't matter who's playing the game---it's very difficult to do a good job. In fact, the wider economy exhibits this behavior in the form of bubbles of various forms.

You can see many examples of this kind of organizational pitfalls in many Silicon Valley tech companies:

• Conventional wisdom is that whenever you promote a great engineer into being a manager you lose a great engineer and get a lousy manager. Companies frequently therefore hire managers from the outside. In the short term this solves the above problem. In the long term, however, outside managers frequently dilute the culture, and more perniciously, by not having a culture of promoting from within, in the long term you get demoralized employees and end up with retention problems.
• I can think of one case where a great engineer was promoted into a lead. This person only worked at night and never met her team-mates. She was, however, very productive (since she never actually spent any time on leadership). This led to a promotion since leads weren't evaluated on leadership. Other leads in the department took note, however, and soon leadership was devalued as everyone followed her lead. The company would end up with a culture where leads would grab all the great work for themselves, since promotions depended not on ensuring that your team was successful, but on individual performance.

The book not only illustrates the approach (called "Systems Thinking"), but in one of the appendices has a complete catalog of corporate dysfunction patterns, with diagrams of feedback cycles, diagnoses, and solutions. So if you're in a hurry, just read "The Beer Game" chapter, the section on "Systems Thinking", and then flip straight over to Appendix 2. Everything else can be treated as fluff.

Recommended.