Count the layers between the person who sees a need and the person allowed to meet it. In many of our organizations the honest number is six or seven. A field officer reports to a coordinator, who reports to a manager, who reports to a head of programs, who reports to a country director, who reports to a regional desk, who reports to a head office. By the time a simple decision climbs that ladder and a yes climbs back down, the moment it was meant for has often passed. We did not design this depth on purpose. We grew it, one reasonable post at a time, and now we pay for it in a currency the chart never shows.
This is not a complaint about managers. The people in those roles are mostly capable and mostly stretched. The problem is structural, it is shared across the sector, and because it is structural it is ours to redesign.
Why the chart keeps getting taller
Each new layer is added for a defensible reason. A span of control grows too wide, so we insert a coordinator. A risk gets missed, so we add an oversight role above the one that missed it. A senior person needs a title to match their tenure, so we create a tier to hold them. None of these choices is wrong on its own. Together they produce an organization where the average decision touches more hands than it needs, and where each hand can slow a thing down but few can speed it up. We optimized every box for control and never measured what the stack of boxes cost in time.
There is a second cost, quieter than delay. Every layer is a place where information is summarized before it moves on. A field reality arrives at the top sanded smooth, its urgency rounded off, its texture gone. The people with the most context end up the furthest from the decision, and the people with the most authority end up the furthest from the facts. Depth does not just slow us down. It dulls what we know.
What fewer layers actually buys
Flattening is not a slogan about removing managers, and we should say that plainly. It is a discipline of asking, for each layer, what decision lives here that could not live one level down. A layer that only forwards work is not managing. It is queuing.
Three moves make this real.
Push the decision to the lowest level that holds the context. For each recurring choice, name the most junior role that has the information to make it well, and place the authority there in writing. The test is simple. If a layer adds a signature but no judgment, the signature is delay wearing the costume of control.
Widen spans where the work is stable, narrow them where it is hard. A manager overseeing a settled, repeatable activity can hold many reports. A manager guiding new, high-risk work needs few. We tend to apply one ratio everywhere, which over-supervises the routine and under-supports the difficult. Match the span to the work, not to the org-chart symmetry.
Count the layers a decision crosses, and treat that count as a number worth managing. Pick a handful of common decisions and trace how many levels each one climbs. Most teams have never measured this, and the measuring alone tends to expose two or three tiers that exist out of habit rather than need.
A layer that only forwards work is not managing. It is queuing.
None of this asks us to be reckless with oversight. It asks us to stop confusing the number of people who can say no with the quality of the decision. Real accountability is a clear line from a choice to a named owner, not a long corridor of people who each touched the file and none of whom feel responsible for the time it took.
The measure of a well-built structure is not how many levels it has. It is how short the honest distance is between a need being seen and a need being met. That distance is something we can shorten, starting with the next layer we are tempted to add.