The hidden cost of admin work

It is Friday at six. The founder of an eight-person firm closes her laptop and tries to remember what she actually did this week.

The calendar is full. Every block has a name on it. There were calls, the right ones. There were emails answered, the right ones. There was a vendor renewal she pushed through, an onboarding kit she rebuilt for the new client, two reports she pulled together herself because the analyst was out, a contract she had to read twice because it was the second version. The week happened.

Nothing shipped forward.

She knows what shipping forward feels like. It feels like a shape getting clearer. It feels like the company being slightly more itself at the end of the week than at the start. This week, the company is in the same shape. The work she did kept it there. She is not sure that should count as work.

The line item that isn’t there

Admin work is the work no one writes down.

It does not appear on the income statement. It is not a vendor and it is not a deliverable. There is no row labeled “coordination” or “follow-up” or “the third email it took to get an answer the first email should have produced.” The general ledger does not see it because the general ledger sees only what changed hands.

This is the first cost. Salary, fully loaded, paid in exchange for hours. Some fraction of those hours produces what the company sells. The other fraction produces the conditions under which the first fraction can happen. A services firm pays for the second fraction the same way it pays for the first. The difference is that only the first one is named.

The second cost is harder to see and larger. It is the work that did not happen because the time was already spent. The proposal that did not get rewritten, the new offering that did not get scoped, the recruitment that took two extra weeks because nobody had three uninterrupted hours in the same week. Opportunity cost is real money. It just does not have a vendor invoice attached, which is why it is almost never managed as a cost at all.

Admin work is paid for twice. It is itemized neither time.

How it compounds

Admin does not scale linearly with the business. It scales with the number of relationships the business has to maintain.

Add a client and you do not add one extra task. You add a recurring stream of small tasks: status updates, invoice reminders, project setup, the inevitable mid-engagement clarification thread. Add a tool and you do not add productivity. You add a notification source, an integration to maintain, a place where information can sit and be overlooked. Add a process and you add the meta-process of remembering when to use it.

Each item is small. None of them is worth fixing on its own.

Together they explain why an eight-person firm starts looking for a ninth, not because the work has grown by twelve percent, but because the coordination overhead has grown faster than the work itself. The ninth hire often exists to keep the first eight aligned. The tenth exists because the ninth needs help. By twelve people you have a head of operations, which is a polite way of saying the company has accumulated enough invisible work to require a department.

Software companies hit this wall too, but they can sometimes hide behind growth. Services firms cannot. The cost is paid out of margin every quarter, in cash, by people who could have been billing.

The focus tax

Even when admin fits into the cracks of a working day, it costs more than the time it takes.

The reason is attention residue. A 2009 paper by Sophie Leroy, Why Is It So Hard To Do My Work?, showed that part of your attention stays attached to the previous task after you switch to a new one, and that the residue measurably degrades performance on the second task. The cost is not the minute you spent answering a one-line email. It is what happens to the next forty minutes of work you tried to do afterwards.

Gloria Mark’s research at UC Irvine, The Cost of Interrupted Work, found that interrupted workers take an average of around twenty-three minutes to return to the original task. That number gets quoted often, but the underlying finding is more useful: people do not recover for free, and the longer they have been interrupted, the worse the recovery is.

Stack a working day full of small interruptions and the math gets brutal before you have even added up the minutes themselves. Every interruption is a small re-entry tax on the next attempt at concentrated thought. The hour the calendar says you spent on admin took two from the work you were trying to do. Sometimes more.

This is why “I worked hard, I just have nothing to show for it” is not a contradiction. It is a description.

Why software has not fixed it

The intuitive answer to admin overhead is to buy a tool that handles it. The last fifteen years of SaaS were sold on this premise. The aggregate result was more admin, not less.

Tools were added to the inbox, not removed from it. Each new product brought a login, a notification stream, a settings panel, an integration that almost worked, and a billing line. The work did not move. It duplicated. A team that used to track tasks in one place now tracks them in three, partially, with the canonical version in nobody’s head.

The incentive structure is the problem. SaaS is sold per seat, and the long-run goal of every per-seat product is to be opened by more people more often. That is the opposite of what good administrative software should aim for, which is to disappear. A product that runs quietly in the background and that nobody logs into is, from the vendor’s point of view, a renewal risk. So the vendor adds dashboards. So the team adds another window.

Most knowledge-work software is not built to remove work. It is built to be present.

What actually fixes it

The fix is not another tool. The fix is to stop doing the work.

Most admin tasks are not problems of capability. They are problems of routing. A human is acting as the connector between two systems that could talk to each other directly, or between a piece of data and the person who needs it, or between a request and the standard reply that always answers it. The hours go to the routing, not to any judgment that required a person.

The work disappears when the routing disappears. That requires looking at one workflow at a time, finding the specific point where the human is acting as a switchboard, and replacing the switchboard with something that just happens. It is narrow work. It is unglamorous. It does not produce a screenshot anyone wants to share.

It does, however, compound in your favor. An hour saved this week is also saved next week, and the week after that, and every week the system runs. Most of what you pay for in your business compounds against you. The systems that quietly remove admin are one of the few things that compound for you.

The cost worth managing

The hardest cost to manage is the one nobody invoices for. It shows up in the work that did not get done, in the hire that was not strictly necessary, in the Friday evening when a thoughtful person looks back on a full week and cannot find what they actually built.

It is worth managing anyway.