Fifty meters apart, conversations stop

Series「People Don't Follow the Org Chart」 · Part 4 of 13 · Part 2 · How structure shapes behavior
Prologue：The people who aren't on the org chart

A Japan entity moves to a new office. Sales on the third floor, engineering on the fifth, an elevator in between. The org chart doesn't change. Reporting lines don't change. Three months later sales complains that engineering doesn't respond, and engineering complains that sales makes promises. Each side thinks the other changed. Only the floor did.

A curve from 1977

In the 1970s Thomas Allen at MIT studied technical communication among engineers and measured how often people talked against how far apart they sat. The result is a steep downward curve: within 10 meters the probability of weekly contact is high; by around 30 meters it has fallen by more than half; past about 50 meters the curve flattens, and contact is not significantly different from being in another building or another city.

The curve has been confirmed many times since. What surprised people is that phone, email and chat did not flatten it. Allen's own 2007 follow-up found that people who sat close not only talked face to face more, they also phoned and emailed each other more. Proximity doesn't produce one channel; it produces the willingness to communicate at all.

The dispute is about remote work. Some studies find that fully remote teams can compensate with deliberately designed communication rhythms. Others find that after going remote, weak ties across teams fall sharply while strong ties within teams get stronger. My reading: the curve still holds, but the unit of "50 meters" has shifted from physical distance to the cost of deliberately initiating a conversation.

The seating plan is an org chart

Most companies treat office layout as facilities work, done by department. Seating by department means the teams that already talk plenty are moved closer, while the cross-department pairs that talk too little are pushed beyond 50 meters. The collaboration lines drawn on the chart are cut by the seating plan.

Japanese offices add a feature: island seating by department, with the department head at the end of the island. This layout draws the boundary of the place very clearly. Vertical communication within the island is smooth; horizontal communication between islands is nearly zero.

The design decision: decide who sits within 10 meters first

Treat the seating plan as part of organizational design, drawn at the same time as the org chart.

Step one: list the two or three pairs of roles that will need the most frequent communication over the next six months but are not in the same box on the chart. Enterprise sales and solutions engineers, say, or support and product. Step two: put those pairs within 10 meters, even if it means splitting a department's island. Step three: don't worry about communication inside a department. Reporting lines and weekly meetings carry it; a few extra meters won't break it.

For hybrid teams, the same principle applies to office days: the pairs that need frequent contact must overlap. Letting each department set its own office days puts every cross-department pair beyond 50 meters.

Choosing an office works the same way. A two-floor office has one more 50-meter wall than a single open floor, and the rent saved rarely covers what that wall costs.

Back to the diagnosis

The context layer of my organizational diagnosis asks about the combination of person and environment. Physical distance is the most overlooked and most changeable item in it. In a diagnosis I take a seating plan, mark the people who appear on the three decision chains, and look at how far apart they sit. The next post is the other side of the same thing: communication structure gets copied into the product.

The two teams in your company that most need to work together: how many meters apart are they right now?

In this series
Previous：The favor ledger is more honest than the org chart
Next：Conway's law: draw the team before the product