Most reorganizations begin with the current organization chart. Boxes are moved, merged and renamed, and a few layers are removed. The result looks different on paper but tends to keep the habits, duplications and unclear accountabilities of the structure it replaced.
Zero-based organization design takes a different starting point. Instead of asking how to adjust the current structure, it asks what work the organization has to do to deliver its strategy, and what the simplest structure to do that work would be. Only then does it compare the answer with what exists today.
The phrase can sound like a euphemism for deep cuts. It does not have to be. In our experience the main benefit is clarity, and lower cost tends to follow from it.
Start with the work
The first step is a plain description of the work: the outcomes the organization must deliver and the activities that produce them. For a retail bank this might include acquiring customers, approving loans, servicing accounts and managing risk. Each is broken down into activities, and each activity is tested with three questions.
Does this activity contribute to an outcome a customer, regulator or shareholder cares about?
Is it being done in more than one place?
Could it be done with less effort, or not at all?
In most organizations, a fifth to a quarter of activity fails at least one of these tests. Reporting is produced that nobody reads. Approvals are repeated at three levels. Two teams in different divisions do the same analysis for different managers.
Design roles around decisions
Once the work is clear, the next step is to decide where each important decision should sit. This is where most of the value lies. Organizations with too many layers usually have them because decisions have drifted upward; each layer exists partly to review the one below.
We map the fifty or so decisions that drive performance, assign each to the lowest level that has the information and accountability to make it, and design roles around those assignments. Spans of control follow. A manager whose team makes routine decisions without escalation can lead eight to twelve people. One who must approve everything can lead four.
A layer that exists only to check the layer below is not management. It is insurance, and an expensive form of it.
Fill the structure from the top down
A common mistake is to design the entire organization before appointing anyone. That takes months, during which uncertainty spreads and good people begin to leave.
A better approach is to design the top two or three layers, appoint the leaders for them, and then ask each leader to design their own area within agreed principles: spans, layers, cost envelope and decision rights. Each leader has a fixed period, usually six to eight weeks, to complete the work. This spreads ownership, keeps the program moving and puts design decisions with the people who will live with them.
What it delivers
In a recent redesign at a large insurer, the zero-based approach reduced management layers from nine to six between the chief executive and the front line. Average spans rose from 4.2 to 7.1. Total cost fell by 14 percent, but the more important change was in decision speed: the time taken to approve a new product fell from seven months to under three.
None of this is achieved by the design alone. It depends on leaders who are willing to give up decisions, employees who are trusted to make them, and a few months of steady attention to how the new roles actually work. But the starting point matters. Begin with the current chart and you will mostly get a version of it. Begin with the work and you have a chance of getting something better.
