Most of the workday is not the work

The 60 percent nobody measures

Every owner has felt it. The team is busy all day. Email pings, calendars are full, calls are constant, and somehow the actual work moves slowly. The boring reason is that most of the day is the work around the work, not the work itself.

That sounds like an opinion. It is measurable. Asana has been counting it for years across tens of thousands of knowledge workers in its Anatomy of Work Index. The headline number from the most recent release lands around 60 percent.

Forty percent of the day is the job

Forty percent of a knowledge worker's day is the job they were hired to do. The other 60 percent is coordinating to get the job done. Status meetings. Asking where a file is. Re-explaining a decision someone missed on Tuesday. Switching tools to move one task forward. Looking for the right person to approve something only one person can approve.

The number has been stable, more or less, across surveys and through the rise of every tool that was supposed to fix it. Slack did not fix it. Project management software did not fix it. Generative AI has not fixed it. Mostly, the new tools just give coordination a new place to happen.

What is striking is how few mid-market operations have ever calculated their own version of this number. Most owners would guess their team spends 70 or 80 percent of the day on real work. The honest measurement, when it gets done, is closer to 30 or 40. Owners are not lying to themselves on purpose. They are looking at a number nobody on the team has any incentive to surface.

What coordination work actually looks like

A coordination day is hard to picture in the abstract, so picture it on a Wednesday at 10:42 a.m. An operations lead is in a recurring status sync on Zoom, half-listening, while writing a Slack reply to a sales engineer asking which version of a quote is current. There are three versions in three places. The right answer takes opening two more tools to check. By the time she answers, she has missed what was said in the meeting, so she pings a colleague at noon to recap the part she missed.

Nothing in that scene is wasted in the sense of being lazy. Every individual act is reasonable. The cumulative result is half a day spent in service of a single quote that, end to end, should have taken twenty minutes.

Now repeat the scene with five people, six departments, and twenty quotes a week. That is the shape of coordination work in most mid-market companies. It is small, defensible decisions piled up until the day is gone. It is rarely any one person's fault. It is rarely solvable by adding more meetings, which is the most common attempted fix and, predictably, makes it worse.

Nobody owns it, because nobody's pay depends on it

A line item gets attention. An expense category gets attention. A meeting nobody likes attending gets attention, eventually, when somebody decides the headcount cost has crossed a line.

Coordination overhead has none of those affordances. It does not appear on a P&L. It is spread in thin slices across everyone's day, so no single person sees the whole cost. Nobody is graded on reducing it. Nobody gets a bonus for taking thirty minutes out of a process. The owner sees only the symptom (people are stretched, the team needs to hire) and treats the symptom by hiring.

The loop is self-reinforcing. A missing system creates a workaround. The workaround needs coordinating. The coordinating needs a meeting. The meeting needs a follow-up. The follow-up is what people are doing instead of the work.

Pointing at it is also awkward in the way office politics are awkward. Coordination overhead has no owner because nobody wants the job of owning it. Whoever shrinks the loop has, by implication, said the loop should not have been there, which is rarely a popular thing to say to whoever put it there in the first place.

The duplicate-work tax

One slice of the 60 percent is pure repetition. A task gets done, and then it gets done again by someone who did not know it had been done the first time. Asana put a number on this too.

Five weeks per person, per year

209 hours is roughly five working weeks. Per worker. The cause is rarely effort or competence. The cause is that the second person did not have a way to know about the first. The information existed; it was just not somewhere the second person would have looked.

Multiply 209 hours by a loaded cost per knowledge worker, multiply by headcount, and the duplicate-work tax for a 60-person operation lands in the seven figures, conservatively. Owners who see that number for the first time usually do not believe it. The number is right.

What makes it harder to attack is that duplicate work feels like productive work to the people doing it. Both versions of the task look like effort. Both versions get done well. Neither person knows they are paying a tax until someone outside the day-to-day adds the columns up.

The context-switch tax

The other half of the overhead is harder to measure but probably larger. It is the cognitive cost of bouncing between tools, channels, and conversations.

A person who spent the morning writing a proposal does not lose only the minutes spent on email. They also lose the ramp-up time it takes to get back into the proposal. That ramp-up is invisible. It happens privately. It is rarely counted, even by the person paying it.

Workers in a typical mid-market operation switch between something like eight to ten applications in a normal day, often more on Mondays. Every switch carries a cost the person paying it would describe as small. The day is small switches.

A useful test: ask anyone on the team how long it took them to do one thing yesterday. They will give you a real, honest answer that turns out to be wrong by a factor of two. The hands-on time was thirty minutes. The end-to-end time was three hours. They were busy the whole time, but not on that.

It compounds quietly

A slow process does not announce itself. It feels like the business is simply busy. Every wait pushes back the next step, every interruption pushes back the wait, and the delays stack. The result is longer delivery times, more work in progress, and capacity tied up in jobs that are not moving.

This is the part owners eventually notice, usually during a quarter that should have been good and was not. Revenue is fine, the team is overloaded, the margin is off. The diagnosis is rarely 'we have coordinated ourselves into a slow operation.' Usually it is something more flattering: demand is up, the team needs more people, the systems need replacing. All of those can be true. So is the coordination problem.

The simplest test is the one almost nobody runs. Take a single workflow. Time the active work. Time the elapsed total. Compare. If the active time is under twenty percent of the elapsed total, coordination is most of what your business is paying for.

What an audit does with it

This is the kind of leak an operational audit is built to surface. It is not urgent, so it never gets attention. It is not dramatic, so it never gets measured. It quietly costs more than most of the visible problems combined.

The audit maps how work actually flows. It interviews the people who do the work, not only the people who manage them. It counts the meetings, the handoffs, the duplicate paths, and the loops. It assigns a defensible dollar figure to each one, conservatively, using the team's own loaded cost numbers.

When that is done, the team can stop arguing about whether the coordination problem is real. They can start arguing about which slice to fix first, which is the more useful argument.