Open-office noise and team communication: why moving desks further apart does almost nothing

Fix the workstation before you rearrange the floor

The lever with the best return is the panel between people, not the gap between them.

Panel height and panel material determine whether a seated person is acoustically separated from the one opposite, and both are properties of the furniture rather than the building. A guide to choosing office workstations will list panel height as a headline variable, alongside footprint and storage, precisely because it is the decision that changes the acoustic result rather than a finish option. The same panel that separates two seated people is also the one standing between a colleague's voice and your microphone.

Rearranging desks costs a weekend of disruption and buys you very little, for a reason that is worth understanding.

The distance maths, which is worse than people assume

Spacing feels like the obvious fix and the physics does not support it.

The GSA's guidance on acoustic comfort in offices, published as Sound Matters, states plainly that "doubling the distance will only drop the decibel level a receiver hears from the sender by 5 points". To reach genuine seated privacy through distance alone, workers would need to sit at roughly 12 feet on centres, which the document describes as unrealistic given space constraints.

Read that against a typical desk pitch of 5 or 6 feet and the conclusion is immediate. You cannot buy your way out of this with floor area. Even a generous refit gets you one doubling, and one doubling is 5 dB, which is perceptible and nowhere near sufficient.

What actually moves the number

Four levers, in rough order of effect per dollar.

The measure behind all of this is the Privacy Index. A PI of 80 is what the guidance calls normal privacy, at which roughly 20% of words can be made out, which is too little content to break concentration. Most commercial buildings sit well below that, which is why eight people can follow every word of a ninth person's client call whether they want to or not.

The counterintuitive item is sound masking. Adding noise to a noisy room sounds absurd and is the single most effective intervention available, because intelligibility rather than volume is what pulls attention away from work. You can ignore a hum. You cannot ignore a sentence.

The video call problem nobody specifies for

Acoustics used to be about people overhearing each other. It is now mostly about what the microphone picks up.

A person on a call in an open bay is broadcasting the room to everyone on the other end, and hearing the room instead of the meeting. Panel height matters here in a specific way: a panel high enough to block a seated sightline also blocks the direct path to a neighbouring microphone, which is the difference between a colleague being background and a colleague being a participant.

The cost lands on both ends of the call. People in the room raise their voices to be heard over themselves, which makes the bay louder for everyone not on the call. Remote participants spend attention separating one voice from the wash behind it, catch about eight words in ten, and ask for repeats. The meeting runs fifteen minutes long and half the team leaves it less certain than they should be. This is the point at which open-plan acoustics stop being a facilities complaint and start being a tax on every hour the team spends together.

This is where the acoustic decision and the collaboration stack meet. Tooling absorbs some of the problem through noise suppression, and the broader pattern of scaling teams without scaling the friction is covered in this piece on digital tools that help businesses scale. Software can cancel steady noise well. It handles a nearby intelligible conversation far less well, because that is signal rather than noise as far as the algorithm is concerned.

Which is why the fix for calls is rarely acoustic alone. A headset with a tight pickup pattern, two or three enclosed booths within walking distance of the bay, and a plain norm that camera-on calls happen inside them will do more for meeting quality than any panel specification on its own.

What noise does to the way a team actually communicates

The second-order effect is the one nobody puts on the refit budget.

In a loud bay, the cheap question stops being cheap. Asking something across the divider now costs an interruption you can hear coming, so people stop asking out loud. Some of that migrates into a chat thread, which is usually an upgrade: it is written down, it is searchable later, and the other person answers when they surface rather than when you happen to arrive at their desk. Some of it migrates nowhere at all, and that is the expensive part. The half-formed question that would have taken nine seconds becomes a quiet assumption, and the quiet assumption surfaces a week later as rework.

Noise also changes who is reachable. Headphones are a do-not-disturb sign that nobody has to justify, and a bay where everyone is defending their own attention is a bay where nobody is easy to reach. Teams tend to read that as a culture problem and treat it with more standing meetings, which is the one intervention guaranteed to make the room louder.

The workable response is to make the quiet default deliberate instead of accidental. Decisions and requests go somewhere they persist and can be read later; real-time interruption is reserved for the things that genuinely cannot wait; and there is enough enclosed space that a call or a focused hour does not have to be negotiated with whoever is sitting nearby. Physical treatment lowers the noise the room produces. Communication norms lower the noise the team produces, and no amount of ceiling tile substitutes for that.

A sequence that works within a normal budget

The order below is deliberate, cheapest and most effective first.

1. Measure before spending. A phone sound level meter is not laboratory grade, and it will tell you whether you are at 50 dBA or 65 dBA, which changes the answer.
2. Treat the ceiling. If the tiles are hard-faced or, worse, absent in an exposed-services fit-out, this is where most of the reverberation lives.
3. Add masking. It is a fraction of the cost of a refit and it works from the day it is commissioned.
4. Then look at panels. Height to block seated sightlines, absorptive facing rather than laminate.
5. Fix the communication defaults in parallel. Agree what belongs in a written channel, what genuinely needs a call, and where calls happen. It costs nothing, and it is the only item on this list that reduces the noise being generated rather than absorbing it after the fact.
6. Rearrange last, if at all. It is the most disruptive and least effective of the six.

Where the honest limit is

Open plan has a floor below which acoustics cannot take it, and pretending otherwise wastes money.

Sustained concentration work and frequent confidential calls are not open-plan activities, and no ceiling treatment converts them into one. The response there is enclosed space in the mix, whether that is booths, focus rooms or genuinely private offices for the roles that need them.

What acoustic treatment does buy is a room where ordinary collaborative work stops being interrupted every few minutes, and a team that is not quietly rerouting every conversation around the noise. That is a real gain and it is achievable. It is just not the same thing as silence, and offices that promise silence to their staff after a refit tend to be judged against the promise rather than the improvement.