Finding Root Causes in a System

In light of the principles of systems thinking we have described, let’s return briefly to ProductCo, the manufacturing company whose story opened this chapter. In cases like this one it is impossible to solve the problem by treating the symptoms. Instead, it is necessary to identify and attack the root causes of the symptoms. But how can they be found? Causation in a complex system has many pathways; those pathways are not always obvious, but they can be identified. In ProductCo’s case we approached causation through a mixed strategy of qualitative data and then a deep dive into quantitative analysis.

Gathering Data

Data gathering through interviews with executives and on-line surveys of managers was an important part of the analytic process. The insights gained were wide-ranging and essential. For example, ProductCo’s implicit strategy for developing leadership talent emerged from these “soft” methods of data gathering. That strategy was predicated on the development of generalist capabilities through the movement of people through many jobs within their functions and occasionally through jobs outside their functions. That strategy was abetted by the perceived importance of filling leadership positions from within and the respect given to those who played the game.

The soft data were supplemented by “hard” quantitative data obtained by digging into the organization’s human resource information system (HRIS). The combination of soft and hard data revealed both what people said about the system and how it worked in reality.

The quantitative approach used in this case was an Internal Labor Market (ILM) analysis, a bedrock tool of human capital management that is explained later in this book. From it we learned a wealth of facts about how the system actually operated, including the following:

• The rates of movement through jobs

• The rates of promotion

• The financial gains that employees experienced through lateral moves and promotions

• The (low) number of outsiders hired into leadership positions

• The job performance of those outsiders

These descriptive measures were essential for understanding the system, but they were not sufficient to establish causal connections.

Taking a Deep Dive

What we call a deep dive involves statistical modeling of how events unfold over time. This final step brings one face to face with causal relationships. It tells an analyst which factors contribute the most to the outcomes of interest, the magnitude of their impact on those outcomes, and their ultimate consequences.

The virtues of statistical modeling as a foundation for management action are communicated in more detail in the following chapters. For now suffice it to say that modeling reveals and explains interrelationships among organizational practices, conditions, and consequences. Modeling also helps reveal the unintended consequences of actions and provides a way to recognize complementarities, both positive and negative. In short, statistical modeling of causes and consequences provides a disciplined foundation for both understanding and action.

One of the things revealed by our deep dive at ProductCo was the price tag of its manager mobility strategy: tens of million of dollars each year. Knowing that monetary cost helped the leadership answer two important questions: Does this policy make sense? and Are we getting value for our investment?

Once we had uncovered and determined the impact of the facts, we were in a better position to make recommendations and the company had a solid foundation on which to make decisions about human capital. The data not only supported management’s suspicions but also made critical connections between processes and results.

In the end we were able to provide the company’s leadership with a concrete plan for redesigning its program for promising managers. Here are the high points:

Changes at ProductCo

Current Situation	Recommendation
Large rewards for promotion	Strengthen rewards for performance in place Enhance reward and recognition for those who develop technical depth and excellence
Insufficient time in position	Clarify the required minimum time in position for different roles and functions Establish new attitudes, behaviors around time in position
Bias toward developing generalists	Restore a balance of generalist and specialist capabilities through managed career paths Focus on business-critical competencies
Quality and performance problems	Monitor implementation of changes and their business impact