The Coordination Tax: Why Adding More People Makes Your Company Slower

You hired to go faster. Two new people in operations, one more in finance, a couple in the project teams. Six months later the work is not moving faster - it is moving slower, and everyone is busier than ever. The calendar is a wall of meetings. Decisions that used to take a morning now take a week. And nobody can quite say who owns the thing that slipped.

This is not a people problem, and it is not a motivation problem. It is a maths problem. Every person you add creates new links that have to be kept in sync, and the number of those links grows faster than the number of people. Past a certain point, each hire buys you a little more output and a lot more overhead. Economists and engineers have a name for the overhead. Call it the coordination tax.

This piece is for the founder, COO, or team lead who feels the drag and is about to reach for the obvious lever - hire more people - without realising that lever is part of the problem. We will show where the tax comes from, why it compounds, why the usual fixes fail, and what actually breaks the pattern: leverage, not headcount.

What the Coordination Tax Actually Is

Every company runs on two kinds of work. There is the real work - the thing the customer pays for, the part that creates value. And there is the work of keeping the people who do the real work aligned with each other. That second category is the coordination tax, and most companies never put a number on it even though it consumes a large and growing share of the day.

• Coordination is not the value - The customer does not pay for your Tuesday alignment meeting. They pay for the product that ships. Coordination is the overhead required to make the product ship when more than one person is involved.
• It hides inside useful-looking activity - Nobody books a calendar slot called “waste”. The tax lives in status updates, handoff emails, cross-team reviews, and re-explaining context. Each one feels reasonable. The sum is a full day a week.
• It is a tax, not a purchase - You do not choose to pay it item by item. It is levied automatically on every unit of real work the moment that work has to pass between people.
• It is invisible on the org chart - The org chart shows boxes and reporting lines. It does not show the hundreds of informal links along which requests, questions, and approvals actually travel.
• It compounds silently - Because no single meeting or message looks like the culprit, the tax can double while every individual decision that created it looked sensible.

The German management literature has a precise term for one part of this - Abstimmungskosten, the cost of aligning and integrating the activities of multiple actors¹⁷. The broader concept, Koordinationskosten or coordination cost, is one of the oldest ideas in the economics of the firm. What is new is how large it has grown in modern knowledge work, and how badly it scales as companies add people.

Brooks was writing about software in 1975, but the mechanism is universal. He observed that a new person needs information before they can contribute, and that information comes from the people already busy. The newcomer asks a blizzard of questions, which slows the very people who were the bottleneck. The same dynamic plays out in finance, operations, service, and sales the moment a team gets big enough to need constant alignment.

Why the Tax Grows Faster Than Your Headcount

The reason adding people backfires is not psychology. It is combinatorics. When you add a person, you do not add one relationship - you add a relationship with everyone already there. The number of possible one-to-one links in a team of n people is n(n-1)/2, which grows with the square of the team, while headcount grows in a straight line³.

Read the last column carefully. Going from 5 to 100 people multiplies headcount by 20, but multiplies the coordination links per person by more than 12. You are not just bigger. Every single person now sits at the centre of far more connections that need maintaining. That is why the tax rises as a share of each hire the larger you get.

The three forces stacked on top of the maths

1. Ramp time - A new hire is a net consumer of attention for weeks or months. They pull answers, context, and review time from the people already at capacity before they return any.
2. Partitioning limits - Some work simply cannot be split. Nine women cannot make a baby in one month, as the old line goes. When work is sequential or deeply interdependent, extra hands mostly add communication, not throughput.
3. Diffusion of ownership - The more people who could own something, the more likely nobody clearly does. Bystander effects that are harmless in a crowd become expensive in a company: the task waits while everyone assumes someone else has it.

The most uncomfortable implication is about scale. As companies grow, complexity tends to grow faster than the value that complexity produces. That is the mechanism behind a pattern many operators feel but rarely name: profit per employee often falls as headcount rises, because each new person adds coordination cost faster than they add output⁵.

Where the Coordination Tax Actually Shows Up

The tax is abstract until you see the line items. It arrives disguised as normal work, which is exactly why leaders rarely target it. Here is where the money goes, with the evidence for each.

1. Meetings that multiply

• Ad hoc everything - Microsoft found 57 percent of meetings are now ad hoc calls with no calendar invite, and 1 in 10 scheduled meetings is booked at the last minute⁷. Alignment that used to happen in passing now becomes a scheduled interruption.
• Meetings breed meetings - The larger the group, the more pre-meetings, syncs, and scrum-of-scrums appear just to prepare for the real meeting. A chessboard calendar of coordination is the visible symptom of the tax³.
• Little gets resolved - A study for the Hans-Boeckler-Stiftung found many organisations treat meetings as the cure for every coordination question, yet only a fraction of participants can genuinely focus on the content¹⁵.

2. Interruptions that shred focus

• Every two minutes - Microsoft found knowledge workers are interrupted by a meeting, email, or notification roughly every two minutes during core hours - about 275 interruptions a day⁷.
• The recovery cost - Research popularised through the Anatomy of Work index puts the time to fully regain focus after a significant interruption at around 23 minutes¹⁰. Interruptions do not just cost their own length; they cost the runway to get back.
• The infinite workday - When focused work cannot happen during the day, it migrates to the edges. Microsoft found after-hours meetings up 16 percent year over year and a meaningful share of workers back in the inbox at 10 pm⁷.

3. Handoffs and tool-switching

• Constant toggling - Harvard Business Research found employees toggle between applications more than 1,200 times a day - roughly once every 24 seconds of active work⁹.
• App sprawl - The average worker moves between about nine applications to get their job done, and every switch carries a re-orientation cost¹⁰.
• Message volume - The average worker receives 117 emails and 153 Teams messages per weekday⁷. Much of it is coordination: where is this, who approved that, can you confirm.

4. The “who owns this?” tax

• Diffuse responsibility - The German operations literature describes how unclear responsibilities and missing decision rules force permanent feedback loops through meetings and status mails, turning knowledge work into coordination work¹⁶.
• Escalation as a habit - When ownership is unclear, the default is to escalate. The decision climbs the hierarchy, consuming senior time and adding days of delay for a call someone closer could have made.
• Rework from misalignment - Two people build against different assumptions because the coordination link between them was weak, and the work has to be redone. Rework is the coordination tax paid twice.

None of these line items looks like waste on its own. That is the trap. The coordination tax is death by a thousand reasonable meetings, and you cannot cut it by asking people to have fewer good intentions.

The Second Tax: Coordination Knowledge Lives in People’s Heads

There is a deeper reason coordination is so expensive: most of it exists to move knowledge from one person’s head into another’s. Who signs off on this. Which system holds the master record. What we decided last quarter and why. When that knowledge is undocumented, every alignment is a live retrieval from a human being, and the tax is paid again every single time.

• Knowledge as tacit - The map of how work really flows - the exceptions, the workarounds, the unwritten rules - usually lives in a few experienced people, not in any system.
• Coordination as retrieval - A large share of meetings and messages are just queries against those human databases: “quick question, who handles X?” Each query interrupts an expert.
• The tax spikes when people leave - When an experienced person resigns or retires, the coordination knowledge leaves with them. The remaining team pays a re-learning tax on top of the ordinary one.
• Onboarding is the tax made visible - Every new hire has to reconstruct the coordination map from scratch by asking the busy people, which is precisely the Brooks dynamic in slow motion.
• Seasonal work is the worst case - Processes that run once a year - the annual close, an audit, a renewal cycle - are the hardest to remember and the most dependent on one person’s memory.

This is the case for building a durable, shared memory of how the company actually operates - a Company Brain. If the coordination knowledge lives somewhere that both people and software can read, the routine questions stop routing through a human, and the answers survive when the human leaves. We will come back to what that looks like in practice.

Why the Usual Fixes Fail

When the drag becomes undeniable, leaders reach for a familiar set of moves. Most of them treat the symptom and leave the mechanism intact - or make it worse. Here is why the four most common responses disappoint.

Fix 1: Hire more people

• The obvious move that backfires - The team is overloaded, so add capacity. But the new person draws information and review time from the people who were already the bottleneck, exactly as Brooks described¹.
• New links, not just new hands - Each hire adds coordination links faster than output, so the effective tax rate on the whole team ticks up³.
• The market is against you - Hiring is slow and expensive, and skilled roles stay open for months. The capacity you need rarely arrives when you need it.

Fix 2: Reorganise

• Redrawing boxes - A reorg changes who reports to whom, but the informal links along which work actually travels often survive untouched.
• Coordination moves, it does not vanish - Splitting a team into two can cut internal links while adding a new, thick coordination seam between the two halves.
• The disruption tax - Every reorg resets relationships and context, temporarily raising the very coordination cost it was meant to lower.

Fix 3: Flatten the hierarchy

• The 2025 trend - Companies are stripping out middle management: 41 percent cut management layers in 2025, and average spans of control have widened to about 12 people per manager, up nearly 50 percent since 2013¹³.
• Overload moves up - Wider spans push coordination onto fewer managers. Gartner found 75 percent of HR leaders believe their managers are overwhelmed, and middle managers report the highest burnout of any group¹³.
• The work is still there - Removing a layer without removing the underlying coordination just redistributes the load onto people who already had no slack.

Fix 4: Buy more collaboration tools

• More surface, more stalls - Each new tool is another inbox, another channel, another place a request can wait. Workers already toggle apps 1,200 times a day⁹.
• Notifications are interruptions - A tool that pings is a tool that fragments focus. Adding channels tends to raise the interruption count, not lower it.
• Tools coordinate people; they do not do the coordinating - Software that shows you the handoff still leaves a human to perform it. The manual work is untouched.

The Real Escape: Leverage, Not Headcount

There are only two ways to get more output from an organisation. Add more people, which adds coordination cost faster than it adds capacity. Or add leverage, which adds capacity without adding a new person who must sit in every meeting. The whole argument of this piece comes down to that choice.

• Headcount is linear capacity, super-linear cost - Every hire is one more node in the coordination graph, and the graph grows with the square of the nodes³.
• Leverage is capacity without a new node - A standard playbook, an automated workflow, or an AI employee absorbs work without needing to be aligned, updated, or invited.
• The best leverage removes the coordination, not just the task - Automating a form still leaves the handoff. Absorbing the whole routine loop - read the request, check the systems, take the action, escalate only the exception - removes the coordination itself.

What an AI employee actually absorbs

An AI employee is not a chatbot in a corner. It is a worker wired into the systems your team already uses - email, Teams, SharePoint, CRM, ERP - that takes over concrete routine work end to end. Applied to coordination, that means the handoffs stop bouncing between people.

• Status chasing - Instead of a person pinging three colleagues for an update, the AI employee reads the systems of record and answers the question directly.
• Routine approvals - It checks a request against the rules, approves what clearly qualifies, and routes only the genuine exceptions to a human.
• Data entry and reconciliation - It moves data between systems so people stop toggling apps 1,200 times a day to do it by hand⁹.
• Answering the routine question - It reads the Company Brain and answers “who owns this, what did we decide, where is the master record” without interrupting an expert.
• Follow-through - It drafts the email, files the document, updates the ticket, and closes the loop, so a decision does not wait on someone finding time to execute it.

Why this beats every fix in the last section

How Superkind Fits

Superkind builds custom AI employees that plug into the systems your team already uses and take over concrete routine work - the exact work that generates most of the coordination tax. The approach is process-first: we start from how your company actually operates, not from a generic template you have to bend to fit.

• Wired into your stack - AI employees connect to email, Teams, SharePoint, CRM, and ERP as one layer over everything you already run. No rip-and-replace, nothing new for the team to learn.
• Built on a Company Brain - Each AI employee reads a structured memory of your processes, decisions, and systems, so it answers routine questions and takes routine actions with real company context - not internet generalities.
• Absorbs the handoffs - It chases the status, checks the request, updates the record, and routes only the genuine exception to a person, so coordination stops bouncing between inboxes.
• Live in weeks - The first AI employee usually goes live within about two weeks. No six-month rollout before anything changes.
• Knowledge that survives turnover - Because the coordination map lives in the Company Brain, it stays when people leave and shortens onboarding for the people who arrive.
• Grows in performance, not headcount - Your team carries more real work without new hires, because the routine load is absorbed rather than passed to another person.
• Improves daily - The AI employees get better because your team works with them and corrects them, so accuracy rises with use.
• Use case by use case - We start with one high-tax process, prove the ROI, then expand, so the investment tracks measurable results rather than a big-bang promise.

A Practical Playbook to Cut the Coordination Tax

You do not fix the coordination tax with a slogan about fewer meetings. You fix it by finding where the tax is highest and absorbing that specific coordination work. Here is a sequence any operator can run.

1. Measure the tax - Track three numbers for a fortnight: the share of the week spent in meetings and messages versus focused work, the number of handoffs in your top five processes, and the average wait time for a routine approval or answer. This is your baseline.
2. Map the handoffs - Pick the process that scored worst and draw every point where work passes between people or systems. Each arrow is a link that costs coordination. Count them.
3. Find the routine loops - Separate the genuine judgment calls from the routine loops - the status chases, the standard approvals, the data moves. The routine loops are the tax you can absorb.
4. Capture the knowledge - Write down how the process actually works, including the exceptions and the unwritten rules. This becomes the seed of a Company Brain and stops the knowledge living only in one head.
5. Absorb one loop with leverage - Put an AI employee on a single high-tax loop end to end: read the request, check the systems, take the routine action, escalate the exception. Do not boil the ocean.
6. Route only exceptions to people - Redesign the process so a human is involved only when judgment is genuinely required. Every routine case that no longer needs a person is coordination tax removed.
7. Measure again and expand - Re-run the three baseline numbers. When wait times and meeting load drop, move to the next-worst process. The gains compound as the Company Brain grows.

Decision Framework: Are You Paying Too Much Coordination Tax?

Not every company should reach for leverage today, and not every drag is a coordination problem. Use these signals to decide where you stand and what to do next.

For a broader look at the failure modes this connects to, our pieces on the interruption tax, the founder bottleneck, and seasonal knowledge that nobody remembers how to run each pull on one thread of the same problem.

Frequently Asked Questions

Related Articles

• The Interruption Tax: What “Quick Questions” Really Cost Your Experts
• The Founder Bottleneck: When Everything Runs Through One Person
• Seasonal Knowledge: The Once-a-Year Processes Nobody Remembers How to Run
• AI Agents for the Mittelstand: Deploy AI Without Losing What Makes You Great

Henri Jung

Co-founder of Superkind, where he helps SMEs and enterprises deploy custom AI employees that actually fit how their teams work. Henri is passionate about closing the gap between what AI can do and the value it creates in real companies. He has watched too many good companies try to grow their way out of drag by hiring, only to buy more meetings and slower decisions. He believes the way forward is leverage over headcount - absorbing the routine coordination so people can do the work only people can do.