← Blog

Focus, explained

The open-plan office and your attention

The open-plan office is the rare productivity intervention that was deployed at enormous scale before anyone measured it, and then kept after the measurements came in. The research is genuinely interesting — not because it says "open plan is bad", which is too blunt to be useful, but because it identifies quite precisely which part of the environment costs you attention, and that part is not the one most people try to fix.

By ComunAI · 10 August 2026 · ~6 min read

The claim the layout was built on

Open plan was justified by collaboration: remove the walls and people will talk to each other more. That is a testable claim, and Ethan Bernstein and Stephen Turban tested it in the most direct way available — by fitting employees at two organisations with sensors that recorded face-to-face interaction, before and after a move from cubicles to open plan.¹

Face-to-face interaction did not rise. It fell substantially, while email and instant messaging rose. Evidence: Moderate The authors' reading is that removing physical boundaries prompts people to construct social ones instead — headphones on, eyes down, conversation moved to a channel where it can be ignored.

The cost is speech, specifically

Ask people what is wrong with an open office and they say "noise". The research is narrower than that, and the narrowness is the useful part.

Simon Banbury and Dianne Berry surveyed office workers about which sounds disrupted them and followed it with laboratory work on the same question. The disruptive ingredient was speech — particularly speech you can make out — rather than volume.² Evidence: Moderate This lines up with a much older and very well-replicated laboratory effect: background speech impairs verbal short-term memory tasks even when the speech is irrelevant, unattended, and in a language you do not speak. We went through that irrelevant speech effect in more detail when we looked at the evidence on focus audio. Evidence: Strong

So the relevant variable is not decibels. It is intelligibility. A café at the same sound level as a bank of desks is less disruptive because you cannot resolve any single conversation out of it — which is why "coffee shop noise" and open-office noise, superficially similar, are not the same stimulus at all.

You do not get used to it as much as you assume

The standard consolation is habituation: give it a few weeks and you stop noticing. You do largely stop noticing. Whether you stop being impaired is a different question, and Banbury and Berry looked at that too, finding that adaptation to speech and office noise was limited and did not straightforwardly eliminate the performance cost.³ Evidence: Moderate

Gary Evans and Dana Johnson ran a related study in which clerical workers were exposed to low-intensity open-office noise, and found effects on physiological stress markers and on motivation — measured behaviourally, after the exposure ended, rather than by asking people how they felt.⁴ Evidence: Moderate Participants did not report being bothered. The measurements disagreed with them.

This is the same dissociation we ran into with sleep: the subjective sense of "I'm fine" is a poor instrument, and it degrades in the same direction as the thing it is supposed to be measuring. It is exactly the pattern in the sleep restriction literature, and it should make you distrust your own verdict on your office.

What people actually trade away

Jungsoo Kim and Richard de Dear analysed a large body of post-occupancy survey data across office types and framed the result as a trade: open plan buys easier interaction and pays for it in sound privacy. Their conclusion was that for most occupants the penalties outweighed the benefit.⁵ Evidence: Moderate

And the ambient interruption rate matters as much as the sound. Gloria Mark's observational work found information work already broken into short fragments with frequent shifts,⁶ and her later study found that interrupted work got finished faster but at a cost in stress and effort.⁷ Evidence: Moderate An open floor removes the last physical friction from someone deciding to interrupt you. It does not create the problem; it lowers the price of causing it.

What actually helps — and how thin the evidence for the popular fixes is

Here is where we have to be careful, because this is the part every article gets confident about and the part where the research is weakest.

Annu Haapakangas and colleagues studied two office relocations where quiet rooms and quiet zones were provided alongside the open floor, and reported improvements in perceived conditions where those spaces existed and were used.⁸ Evidence: Early That is the most direct support available for the obvious intervention — somewhere else to go — and it is field research with the usual limitations, not a controlled trial.

On headphones: we know of no good controlled evidence that wearing headphones in an open office protects task performance. Evidence: None There is a plausible mechanism — masking speech so it cannot be resolved — and there is a competing one, since music with lyrics is itself intelligible speech. We are not going to claim a result that has not been produced.

What we can and can't say

The calibrated bottom line

If you can move for an hour, move — that is the intervention with the most direct support, and it works by removing intelligible speech rather than by making you more disciplined. If you cannot move, the position the evidence supports is that you are managing a cost rather than eliminating one, and the remaining lever is the interruptions you can control.

That distinction matters more than it sounds. The open floor sets a floor on how much ambient disruption you're exposed to, and no app changes that. What an app can change is the second stream — the notifications, the tab you open at the first dull moment, the switch you make yourself. In an environment that is already noisy, the self-inflicted interruptions are the ones with any slack in them, and there is a decent case that they are worth batching out of the way precisely because the external ones cannot be.