Optimizing Each Part of A Firm Doesn't Optimize The Whole Firm
Optimizing Each Part of A Firm Doesn't Optimize The Whole Firm
Most people found the General Motors ignition switch scandal appalling. Not only did the defect result in over 100 deaths, but it
turned out that fixing the problem would have cost less than $1per car. For many, it was a horrible indictment of corporate greed.
Profits, it seems, were valued more than human lives.
Yet look a little closer and it becomes clear that the real problem wasnt callousness, but mismanagement. The defect in the ignition
system was, in fact, relatively minor. The real problem was that it caused airbags not to deploy. Each subsystem was performing to
standard, but the interaction between them resulted in disaster.
Unfortunately, most organizations today fall into the same trap: they look at isolated metrics, but fail to see the whole system. They
optimize each part of the business separately, and fail to consider how they interact. When we see an operation as a set of isolated
metrics to optimize, we can lose our sense of context and decrease overall performance an efficiency paradox.
In The Good Jobs Strategy, MIT professor Zeynep Ton makes exactly this point through her study of retailers. The conventional
wisdom says to maximize profits through low wages, optimized scheduling, and extensive inventory management systems. Yet her
research finds that these practices often serve to reduce overall efficiency and profitability.
One example she gives is how bad weather can wreak havoc on an optimized system. A rainy day reduces store traffic, resulting
in lower sales. Lower sales translate into reduced staffing forecasts. Stores are then understaffed when the sun comes back out,
leading to poor customer service and inventory tracking problems. The result is even worse sales, even more reduced staffing, and a
vicious circle of lost revenue.
By contrast, some high-performing retailers such as Trader Joes, Costco, and Quick Trip, deliberately increase their labor
investment through better wages, increased training, and over-staffing. While these extra costs might not look good on a
spreadsheet, it allows them to handle a lot more complexity and generate new ideas, which increases performance.
In his book Team of Teams, General Stanley McChrystal describes another aspect of the over-optimization problem. Although his
soldiers were winning every battle, somehow they were losing the war. Whats more, every time they began to gain the upper hand
by shifting tactics, the enemy would adapt. It was beginning to seem like they were engaged in a never-ending game of whack-amole.
McChrystal realized that although his squads of highly trained commandos and intelligence analysts were performing their
individual tasks with world class alacrity, they were failing to, as he put it, see the whole system. For example, teams of
commandos would go on a raid and capture valuable intelligence, but then bags of documents and hard drives would sit in a closet
for weeks before anyone got a chance to look at it. Other times, an analyst would make an important breakthrough, but was unable
to get that intelligence to the ground units that could make best use of it.
McChrystal took the unusual step of decreasing the emphasis on efficiency and focused his sights on agility and interoperability. By
beefing up the roles of liaison officers and embedding specialists in each others units, he slowed each unit down slightly, but
overall operational efficiency increased by a factor of seventeen.
Managers often fall into the trap of thinking that by improving each part of their enterprise, they will improve the whole. However,
the opposite is often true.
Success and failure are rarely determined by performance against a plan, but rather how you adapt for events that cannot be
foreseen. If, as Zeynep Ton described, an errant thunderstorm or blizzard can throw your system off, its not much of a system.
The problem is that the world is far too complex to be reduced to excel sheets, organization charts, and diagrams. In the final
analysis, nobody cares what your internal metrics are. Whats really important is not the nodes, but the network. Thats what
McChrystal means when he speaks of seeing the system.
If everyone is trained and compensated to focus on only their part of the task, the shared mission is lost. Thats not a path to
greater efficiency or to profitability, but to oblivion.
Greg Satell is a U.S.-based business consultant. You can find his blog at Digital Tonto and follow him on Twitter @DigitalTonto.
Comments
Leave a Comment
POST
8 COMMENTS
REPLY
POSTING GUIDELINES
We hope the conversations that take place on HBR.org will be energetic, constructive, and thought-provoking. To comment, readers must sign in or register. And to ensure the quality of the discussion,
our moderating team will review all comments and may edit them for clarity, length, and relevance. Comments that are overly promotional, mean-spirited, or off-topic may be deleted per the
moderators' judgment. All postings become the property of Harvard Business Publishing.