Deep work at scale

The uncomfortable baseline

Our focus scoring looks for sustained, low-switch engagement with demanding work. Across all teams on the platform, the median knowledge worker manages one hour and 43 minutes of it per day. Not eight. Not five. Under two.

When we show executives their own organisation’s number, the first reaction is always denial, and the second is to ask which individuals are dragging the average down. That is the wrong question. The distribution is remarkably tight; almost everyone is close to the median, because the constraint is environmental.

You cannot hire your way out of a calendar problem.

The number also moves with company size in a way that should worry anyone scaling fast. Teams under 50 people average around two and a quarter hours of daily deep work. Past 500 people, the average drops below ninety minutes. Coordination is a purchase, and focus is the currency it gets paid in.

Why personal systems fail in public

Every organisation has that one engineer with noise-cancelling headphones, a blocked calendar, and a two-day response time. Their focus hours are triple the median. They are also quietly resented, because their protection is funded by colleagues who absorb the interrupts they deflect.

This is the core reason individual focus advice does not scale. Attention in a team is a shared pool. One person defending their slice, without an agreed system, just moves the drain to someone with less standing to refuse.

The teams that succeed make protection collective, so nobody has to be the difficult one.

There is a fairness dimension here that leaders underestimate. In our category data, deflected interrupts land disproportionately on newer employees and on women, the people least positioned to decline them. A collective policy is not just more effective than individual heroics; it is considerably more equitable.

The 200-person experiment

A software firm in Bengaluru, around 200 engineers, ran the cleanest test we have observed. Half the teams adopted synchronised quiet hours, 9:30 to 11:30 daily, with chat expectations formally relaxed. The other half continued as usual for a quarter.

The quiet-hours cohort added 38 minutes of measured deep work per person per day by week six. More interesting was what did not happen: incident response times were unchanged, because genuinely urgent things always found a phone. The always-on habit had been protecting against an emergency that occurred roughly never.

The control teams demanded the same policy before the quarter ended. Some experiments conclude themselves.

Synchronise, don’t just schedule

The detail that made Bengaluru work was synchronisation. Everyone dark at the same time means nobody is waiting on anybody. Staggered individual focus blocks, the more intuitive design, performed far worse in our data, because a blocked calendar does not stop the person who needs you from needing you.

It only works team by team, though. Whole-company quiet hours collide with time zones and with roles that must stay responsive. The unit of synchronisation is the group that actually interrupts each other daily, which is usually five to twelve people.

Find that unit, and two shared hours beat any amount of personal discipline.

The window’s placement matters less than its consistency. Teams that anchored quiet hours to their measured focus peak did slightly better, but teams that simply kept the same two hours every day, wherever they fell, retained the habit far longer than teams that let the block float around the calendar week by week.

What managers do during quiet hours matters most

In deployments where quiet hours decayed, we traced the decay to leaders almost every time. A director who sends 10am messages ‘just to read later’ has ended the policy, whatever the policy says. Seniority converts every message into an implicit deadline.

The durable version we see has managers doing their own deep work in the window, visibly. One VP we work with writes her board updates then and says so in her status line. Her organisation’s quiet-hour compliance is the highest we measure.

Culture is downstream of what the most senior person does at 9:45 on a Tuesday.

We now recommend making leadership behaviour part of the measurement itself. In accounts where executives can see their own interrupt footprint, the volume of messages senior people send into protected windows falls by roughly half within a month. Nothing was mandated. The mirror was enough.

Measuring without turning focus into a leaderboard

The obvious failure mode is ranking people by focus hours. We refuse to build that view, and customers occasionally ask. A leaderboard converts deep work into performance theatre within weeks: long, empty sessions in an IDE, chat left deliberately unanswered, the metric fed and the purpose starved.

What we surface instead is the team trend and its relationship to output signals the customer already trusts, cycle time, tickets resolved, features shipped. Focus is an input. Inputs are for diagnosis.

When a team’s focus rises and its output does not, that is a finding too. Usually it means the recovered hours are going into work that should not exist.

Protecting the gains

Focus improvements erode. New hires arrive without the norms, a crunch suspends quiet hours ‘temporarily’, a reorganisation splits a synchronised team. In our retention data, roughly half of teams that gained an hour of daily deep work had given a third of it back within two quarters.

The half that kept the gains shared one habit: the focus number appeared in a recurring leadership review, next to revenue and attrition, and someone owned it.

What gets reviewed gets protected. Everything else was always going to be borrowed.

Onboarding deserves particular attention. We suggest putting quiet hours in the first-week induction, explained as plainly as the expense policy. A norm stated on day three becomes furniture; a norm discovered by accident in month four becomes optional. New joiners will protect whatever they were told the team protects.

Writes for The Signal about culture and the future of measurable, humane work — drawing on anonymised patterns from the teams and focus hours analysed on Momentum.