Add people, and each one pulls less

Series「People Don't Follow the Org Chart」 · Part 8 of 13 · Part 2 · How structure shapes behavior
Prologue：The people who aren't on the org chart

A project moves fast with three people and slows down after it grows to eight. The manager's first thought is "communication cost," so more meetings are added to fix the communication cost, and it slows down further. This was measured over a century ago.

The rope experiment

In the 1880s the French agricultural engineer Max Ringelmann ran a simple experiment: have people pull a rope and measure the force. One person averaged 63 kilograms. Three people together pulled not 189 but 160, or 53 each. Eight people pulled 248, or 31 each, less than half the individual figure. The more people, the less each one pulls.

The result was buried for almost a century and rediscovered in the 1970s. A 1974 experiment by Ingham and colleagues took it apart: subjects were told they were pulling with others, while the others only pretended to pull. The subjects' force still dropped. So the loss isn't only coordination (people pulling out of sync); part of it is motivation. Knowing others are there, you ease off a little. The latter became known as social loafing.

The dispute is over proportions: how much is coordination and how much motivation varies widely by task. Research has also found that when individual contributions are identifiable, motivation loss nearly disappears. That last finding is the most useful conclusion in the whole story.

Tug of war inside organizations

Most work in organizations is tug of war: several people pulling toward one goal, with no way to see how hard each rope is being pulled. A sales team's quarterly target, a project's delivery date, a department's KPI are all one shared rope.

The Ringelmann effect shows up like this: the goal is collective, so everyone rationally eases off a little; contributions are invisible, so those who pull harder aren't recognized and, after a few quarters, ease off too; and the more people, the worse both get. Adding people to fix a schedule hands the rope to more hands, each pulling more loosely.

Japanese organizations have an amplifier: the tradition of collective responsibility. "Everyone is responsible together" is a virtue in Japan, but in the face of the Ringelmann effect it is precisely the condition that makes each person's effort invisible.

The design decision: responsibility modules, one owner each

Since motivation loss nearly disappears when contributions are identifiable, the direction is clear: cut the rope apart.

My method is responsibility modules. Divide a team's work into blocks, each with a defined output and boundary, each with exactly one owner. Not a "lead owner," not "jointly responsible": one name. Others can contribute, but when the block fails, who to ask, and when it succeeds, whose result it is, has one answer.

Three concrete rules. One: the number of modules roughly matches the number of people, so everyone is sole owner of at least one and has a rope of their own. Two: write down module boundaries and decide in advance who owns what falls on the boundary; otherwise the gaps between modules become new zones of collective responsibility. Three: tie evaluation and bonus to module outputs, not to the team's total goal. The team goal sets direction; it doesn't distribute money.

This rule is the basis for the middle track of my three-track pay design, where the grade sets base pay, the role sets responsibility pay, and results set the bonus: the role component pays for sole ownership of one responsibility module. Part of the problem in A 50% raise and he still left came from responsibility nobody could see.

Back to the diagnosis

In an organizational diagnosis I ask one simple question: which of this team's outputs can be pointed to a single name? The part that can't is where Ringelmann is at work. It is usually far larger than the manager thinks.

Your team's quarterly target: how many modules with a single owner can it be split into?

In this series
Previous：Your bilingual staff are not translators
Next：Four kinds of relationships, and why pay design goes wrong