Why Work About Work Grows Faster Than Real Work as Companies Scale

By Victor Hoang · Updated Oct 10, 2026

Every founder remembers the early days as fast. Nobody held a status meeting because everybody knew the status. Nobody needed an approval because the person who'd approve it was sitting four feet away. Then the company tripled, and somehow shipping the same kind of work started to take three times as long with three times the people.

That's not a morale problem or a hiring problem. It's arithmetic. The work that exists to keep people informed, aligned and approved grows faster than the number of people doing it. Our view is that scaling companies can't treat this as a one-time cleanup. They have to prune process on purpose every year, the same way they review budgets, or watch a growing share of capacity go to managing the work instead of doing it.

This article explains why the growth is faster than linear, what the evidence says about how much it costs, and gives you a decision table and a yearly routine for cutting it back. If you want a way to measure the damage first, the Coordination Tax framework is the right companion. This piece is about why the number keeps climbing and what to do about it every year.

Key Facts

• Microsoft's 2023 Work Trend Index, built on Microsoft 365 activity data for a 28-day period ending March 2023, found the average employee spends 57% of their time communicating in meetings, email and chat, and 43% creating. It also reported 192% more Teams meetings and calls per week since February 2020.
• Asana's Anatomy of Work Index, a vendor survey of more than 10,000 knowledge workers, reports that 60% of time at work goes to work about work rather than skilled work.
• Writing in Harvard Business Review in 2016, Rob Cross, Reb Rebele and Adam Grant found that time spent in collaborative activities had ballooned by 50% or more over the previous two decades.
• Bain & Company examined the calendars of 17 large corporations and found 15% of collective time went to meetings of three or more people, a share that had risen every year since 2008. Bain estimated most companies could free at least 20% of collective hours.
• In Fred Brooks's classic argument, if every part of a task must be coordinated with every other part, effort grows as n(n-1)/2, so three workers need three times the pairwise communication of two and four need six times.
• Gary Hamel and Michele Zanini estimated in 2016 that excess bureaucracy costs the US economy more than $3 trillion, or about 17% of GDP. That's the authors' own estimate, not a measured statistic.

The Arithmetic Behind the Slowdown

Start with the number of relationships. Two people have one communication path. Three have three. Ten have 45. The formula is n(n-1)/2, and Brooks used it in The Mythical Man-Month to show why adding people to a project doesn't add capacity in proportion. In his chapter on the man-month, Brooks wrote that intercommunication is the worse of two added burdens, the other being training new people, which grows only linearly.

Here's what the formula does to a growing company:

Go from 10 people to 100 and headcount is up 10x. The possible paths are up 110x.

Now the honest caveat. Nobody in a 200-person company talks to all 199 colleagues, and Brooks's formula assumes every part must be coordinated with every other part. Real organizations cut the paths with teams, managers and documentation. So treat the formula as an upper bound, and the lesson as directional: every added person adds more coordination surface than the one before. What actually happens is that organizations respond by adding structure to contain the paths. And each piece of structure is itself work about work.

That's the part leaders miss. The status report, the approval step, the weekly cross-team sync and the extra management layer are all rational responses to a real problem. Each one is defensible alone. Together they become the load.

Four Ways Coordination Compounds

The path count is only the start. Four mechanisms stack on top of it.

Every fix adds a standing obligation. A missed handoff leads to a new checklist. A surprised executive leads to a new weekly report. A budget overrun leads to a new approval. Nothing in that sequence includes a step where something gets removed. Bain's team described the same pattern in its work on organizational time, calling it initiative creep and noting that seemingly sensible projects are added incrementally and few are ever formally terminated. At Newmont Mining, 87 initiatives were running when a new CEO arrived in 2013, each claiming executive attention.

Meetings multiply down the org chart. Bain traced one company's weekly senior-team review. The meeting itself consumed 7,000 person-hours a year. Pre-meetings between each executive and their unit added 20,000. Team meetings feeding those added 63,000. Preparatory meetings added 210,000. The total reached 300,000 hours a year, before counting any data collection. One recurring meeting at the top created roughly 43 hours of work below it for every hour in the room.

Approvals chain. Each new layer gets a veto because the person at that layer is accountable for what happens below them. The cost isn't only the approver's time. It's the waiting, the chasing and the context-switching while a request sits. Asana's survey lists chasing for approvals among the main reasons people stay late, next to unexpected meetings and unclear goals. That's one vendor's survey, but the mechanism is easy to recognize in your own calendar.

Collaboration demand keeps rising. Cross, Rebele and Grant's finding that collaborative time ballooned by 50% or more over two decades describes the same dynamic from the individual's side: every new colleague, team and process is one more source of requests. If you want the leadership angle on that, collaboration overload as a leadership failure covers it directly.

What the Evidence Says It Costs

Several independent sources land in the same range, though they measure different things, so don't add them together.

Microsoft's activity data says 57% of employee time is communication. That's measured from telemetry across Microsoft 365, so it covers people who use Microsoft's tools, and it counts communication as a category, not all of it as waste. Plenty of communication is the job.

Asana's 60% comes from self-reported survey data from a vendor that sells work management software. It's useful as a signal of how workers perceive their days and weaker as a precise number.

Bain's calendar analysis is the most useful for scaling questions because it's about large organizations and it tracks a trend, not a snapshot: meeting share up every year since 2008. And the 20% recoverable figure is a conclusion drawn from their data, not a guarantee.

Taken together, the safe claim is this: in large knowledge-work organizations, somewhere around half of working time is communication and coordination of some kind, that share has been rising, and a meaningful fraction of it, Bain says at least a fifth of collective hours, is recoverable. For the human side of what happens when you leave it alone, see why your best people leave meeting-heavy teams.

A Hypothetical: The 60-Person Company

Here's an invented but typical example. A company grows from 20 to 60 people in two years. Along the way it adds a weekly all-hands status meeting, a Monday leadership sync, a Wednesday cross-functional sync, a weekly project report for each of six projects, a two-step approval for any spend over a small threshold, and a monthly review deck. Each was added after a specific problem.

None of those items is large. But suppose each person touches an average of six hours a week across them, counting preparation and follow-up. That's 360 hours a week across 60 people, the equivalent of nine full-time employees whose entire output is the machinery. Nobody chose that. It accumulated one reasonable decision at a time.

The point isn't the specific numbers. It's that the cost is invisible when each item is judged alone and obvious when you add them up.

A Decision Table for What to Keep

Not all coordination is waste. Brooks's argument doesn't say avoid communication, it says communication has a cost that grows, so spend it where it buys something. Use this table to sort any recurring status update, approval, meeting or report.

The approval rows matter most. An approval that is nearly always granted is a tax with no return. Replace it with a limit ("anyone can spend up to X without sign-off") and a periodic audit. A delegation of authority matrix is the usual tool for setting those limits, and decision-making velocity explains why the speed of those decisions matters competitively.

The Yearly Pruning Routine

Good intentions don't prune. A routine does. Here's a version that fits in one quarter.

1. Build the inventory. List every recurring meeting, report, standing approval and status ritual. Include owner, attendees, frequency and why it was created. Most companies find the list is longer than anyone guessed.
2. Put a price on it. For each item, estimate person-hours per year including preparation and follow-up. Bain's example shows why you have to follow the ripple: the room time was 7,000 hours and the true cost was 300,000.
3. Run the decision table. Sort every item into keep, convert to async, merge, or delete.
4. Set a cut target. Bain's estimate that at least 20% of collective hours is recoverable is a reasonable ceiling to aim at, not a promise. Start with a 15% reduction in recurring meeting hours.
5. Give everything an expiry date. Any new recurring meeting, report or approval gets a sunset date, say six months out. It survives only if its owner can name a decision it served.
6. Cap the number of live rituals. Treat organizational time as fixed, which is how Bain describes the approach at Ford. If something new is added, something else goes.
7. Repeat next year. Calendar the review the same way you calendar budget planning.

Here's a quick triage of which categories usually give the fastest return:

For the meeting-specific rules, 11 rules for cutting pointless meetings is a practical starting point. And async-first operating models cover what replaces them. If your inventory turns up a queue where work simply waits, finding bottlenecks in business processes will help you separate a coordination problem from a capacity one.

What to Do Next

Don't launch a transformation. Do these four things in the next 30 days.

1. Count the rituals. Spend one afternoon listing every recurring meeting, report and approval. Just the count and the owner. The number alone usually starts the conversation.
2. Pick one approval and kill it. Choose the one with the highest approval rate and replace it with a limit and a quarterly audit. You'll learn quickly whether anything breaks.
3. Attach an expiry date to the next new ritual. Make it the rule that nothing recurring gets created without a sunset date and a named decision it supports.
4. Put pruning on the annual calendar. Same level of attention as budget planning, with an executive owner. Without an owner, the inventory grows back within a year.

The goal isn't zero coordination. Coordination is what lets 200 people act like one company. The goal is to keep the share that goes to coordination from creeping up every year simply because nobody's job is to take things away. Make that someone's job.