One of my first teaching experiences involved an evening MBA class. At least half of the students were older than I was and all had more real-world experience. We were discussing organization design and the choice to centralize or decentralize. It’s a choice with profound implications for the structure of incentives, communication patterns, and decision rights.
We reviewed how decentralized designs promote innovation, motivation, and autonomous action, while centralized designs support control, efficiency, and coordination. My central and rather predictable message focused on the need to “design for fit” and to match organizational structure to strategy.
We then launched into our case for the day. It illustrated Hewlett-Packard’s decision to centralize after decades of decentralization. On the surface, this message and case illustration seemed completely sensible.
After the case discussion ended, a student who worked at McDonnell Douglas (now Boeing) raised his hand and said, “I disagree with everything you have said. I have watched our company move back and forth between centralization and decentralization repeatedly, and it has nothing to do with what you just described.” A classmate chimed in, “I agree completely with [the first student]. Our company has also bounced back and forth between centralization and decentralization. I am quite sure management has no idea what they are doing.”
This sobering start to my teaching career presented a puzzle that remained with me for years. If fit with strategy or the environment stands as the central object of organization design, why do firms change their design so often? Are strategies or environments really changing at this pace? Or, are managers just flailing about to elevate performance?
Perhaps we’re asking the wrong question. For any corporation, sustaining value creation almost always requires attending to a rather sweeping agenda of objectives—accelerate innovation, streamline supply chains, elevate customer satisfaction, reduce production cost, reduce production defects, and increase capacity utilization. The specific components of your own list may vary, but the length frequently does not.
Organizations often go to great lengths to compose metrics, structures, and incentives that target all these directions. But the reality is that aggressively driving behavior in all these directions at once is a fool’s errand. Individuals are greatly limited in their capacity to multi-task and organizations are not much different.
Yes, you may allocate different agendas to differing teams, but often these objectives transcend any given team. Moreover, organizations are massively inert, and a nudge in this direction or the other is simply ineffective in generating any form of organizational momentum. Instead, achieving the broad set of outcomes you envision almost always demands ongoing directional shifts.
Often these entail centralizing or decentralizing. Note that in this view, changes in design or emphasis are not driven by dynamics in the environment or strategy that shift the target, but rather by the complexity of the target and the limits of design in optimally shaping both individual and organizational attention.
In hindsight, it’s ironic that the case I used in that early class was Hewlett-Packard. True to the students’ implied prediction, at precisely the time we were discussing HP’s brilliant shift to centralize, they were actually in the process of shifting back, now six years later. Fast forward another five years and HP was once again centralizing, only to three years later decentralize.
Then, in 2000, Carly Fiorina steps in and dramatically centralizes HP, only to then refuse to decentralize and as a consequence be pushed out the door by the board. Mark Hurd then steps in 2005 and predictably decentralizes. Thus in the 20 years that followed that classroom discussion, HP made an additional five directional shifts. Over these years, HP composed tremendous shareholder value and emerged by 2010 as the world’s largest IT company. They achieved this position not by statically matching strategy to organization design, but through skillful dynamic design.
I often liken designing organizations as akin to sailing into the wind. Rather than aiming directly toward the target, the quickest path to progress is setting a course 30 degrees off wind, only to then tack 30 degrees off wind, setting a course in the other direction. Through this process of tacking or dynamic direction setting, organizations make faster progress to their envisioned destination versus composing some clever, complex design at the outset that attempts to push in all desired directions. As you’ll discover, effective strategic leaders are skilled dynamic designers.
This post is an adapted excerpt from my book, Beyond Competitive Advantage. You can download a free chapter below.