Perspectives on complexity

How to bridge the complexity gap

My colleagues and I have been studying the complexity gap for about 15 years now.

It all started when we were hired in 2002 by a U. S. federal agency to conduct research on what it takes for leaders to make good decisions under VUCA (volatile, uncertain, complex, and ambiguous) conditions.

I’m using the term leader here, fully aware that some people might use manager instead.

We began the project with a study of leader decision making, ultimately interviewing several hundred employees representing every level in a large federal agency’s employee hierarchy. The only group we were unable to interview were leaders at the very top.

Thinking complexity

The interviews were wrapped around a set of wicked workplace problems — complex problems with VUCA features and no “correct” solutions. We studied these interviews, scoring them for their complexity level with the Lectical Assessment System, while thoroughly documenting the ideas and skills demonstrated in responses (so we could construct descriptions of how ideas and skills changed from level to level).

Role complexity

As we completed the first phase of the project, which focused on thinking complexity, we began developing an approach to determining role complexity, again using the Lectical Assessment System. This was an interesting challenge. We settled on an approach that started with a basic Lectical Analysis of easily observable increases in the number and complexity of stakeholder interests associated with decision making at each higher level in the agency’s hierarchy. Then we zoomed in, examining the complexity associated with the work of different departments within the organization, and finally, the complexity associated with work in specific roles.

Comparison of the complexity level of jobs with the complexity of leader/manager’s thinking

The figure on the left shows the results of the thinking and role complexity analyses. The basic story here is that the complexity of roles increases in a linear fashion as we move up the hierarchy, but the average complexity of leaders’ thinking does not.

However, that’s not the whole story. From semi-skilled roles to entry level roles, role complexity and thinking complexity are pretty well aligned. Then there’s a flattening out of the curve from the mid- to upper-levels, and a return to growth in higher work levels that’s parallel to, but well below, the role complexity curve.

Our client wasn’t surprised by this pattern. She reported that the agency routinely hired for senior roles by going outside the agency — because existing employees were not developing the skills required in higher roles. We weren’t surprised either. The agency had a command and control culture. Nothing stifles development like command and control, because there is generally no role for critical reflection — essential for development — at lower levels in a command and control hierarchy.

Additional evidence for the complexity gap

We have continued to observe similar patterns in the relation between role complexity and thinking complexity, with a few notable differences. For example, as you can see in the “average Lectical Score” results in the chart below (sample size > 2000), we have found higher mean levels of performance in lower-level roles than observed in our original government sample. In contrast, we have found lower mean levels of performance for more senior roles. On average, leaders in executive level roles in the government agency performed at level 12. In our larger sample, this average was just under 1160.

The chart below shows the results for several thousand leaders working in numerous industries. The data are not, for the most part, from carefully designed experiments. This is in every way a “convenience” sample. Skepticism is appropriate, but given the size and diversity of the sample, general trends are likely to be relatively accurate.

An advantage of 40 points may not sound like much, but a score of 1200 represents a very different kind of thinking than a score of 1160, as illustrated below.

Take another look at the Lectical Score vs. management level chart, above. In addition to showing role complexity alongside average Lectical Scores, it includes benchmarks. These benchmarks can be thought of as “practical objectives,” and they represent a major shift in our treatment of the relation between the complexity of leaders’ thinking and the actual complexity of roles. Benchmark estimates incorporate two factors. First, the actual distribution of complexity levels in the population, and second, evidence that highly skilled leaders use VUCA skills to bridge the gap between the complexity of issues confronted in their roles and the complexity level of their thinking.

The image on the left shows the relationship between benchmarks and role complexity. As you can see, by CEO level 1, the gap between the complexity level of benchmarks and the complexity level of roles is more than a full Lectical Level. Because the distance between benchmarks and role complexity increases with every level in an organizational hierarchy, leaders’ mastery of VUCA skills must also increase if they are to bridge the gap.

VUCA skills to the rescue

VUCA skills are (1) required for making good decisions in volatile, uncertain, complex, or ambiguous contexts. They also (2) support leader development, (3) improve human interactions, and (4) help bridge the gap between the complexity level of leaders’ thinking and the complexity of issues faced in their roles. The VUCA skills we measure are…

  • perspective coordination — determining which perspectives matter, seeking out a diversity of relevant perspectives, and bringing them together in a way that allows for the emergence of effective solutions.
  • decision-making under complexity — employing a range of decision-making tools and skills to design effective decision-making processes for complex situations.
  • contextual thinking— being predisposed to think contextually, being able to identify the contexts that are most likely to matter in a given situation and determine how these contexts relate to a particular situation.
  • collaboration — understanding the value of collaboration, being equipped with the tools and skills required for collaboration, and being able to determine the level of collaboration that’s appropriate for a particular decision-making context.

The most highly skilled leaders we’ve encountered not only have scores that meet benchmark estimates for their roles — they are VUCA virtuosos. We’ve set VUCA skills benchmarks based on the levels of performance achieved by these leaders.

Unfortunately, as you can see in the graph below, the overall averages for VUCA skills are not as high as they need to be for optimal decision making.

On average, leaders at the executive level do a bit better than those at lower levels, but the disparity between VUCA benchmarks and leaders’ average scores increases steadily as role complexity increases. This is a problem.

The good news is that this problem can be remedied. Here is the recipe:

  1. Measure the level of leaders’ VUCA skills to identify opportunities for growth.
  2. Help leaders leverage everyday workplace experience to build skill continuously in identified growth areas.
  3. Implement iterative learning, decision-making, planning, and design processes throughout the organization.
  4. Reward small-scale, low-risk collaborative experimentation.

By the way, this is also part of our recipe for engagement, satisfaction, productivity, and organizational health.