MELLORCATalk to Mellorca
Patterns

Manual Work & Administrative Load

Why Small Tasks Become Expensive at Scale

A two-minute task looks too small to fix, but its cost is set by how often it runs, who does it and what it interrupts. Counting the multiplier shows which small tasks have quietly become large ones.

Mellorca Patterns2026-10-04

Observable condition

Ask anyone in the business about their worst task and they will rarely name the small ones. Saving an emailed purchase order into the right folder and renaming it. Checking that a payment has reflected before releasing an order. Copying a delivery address from an email into the dispatch system. Sending a "received, thank you" reply. Updating the job status in a spreadsheet after it has already been updated in the job system. Each takes a minute or two. None feels worth a meeting, let alone a project.

So the small tasks stay. Nobody owns them, nobody counts them, and when a new person joins they are simply shown how it is done. When volumes rise, the tasks rise with them, and the team experiences that not as one large problem but as a general feeling of being busy.

An illustrative example: a Gauteng distributor of building supplies with about sixty staff processes orders from contractors by email and phone. Before each order is released, someone in the sales office checks the customer's account balance in the accounting system, confirms stock in a separate stock sheet and forwards the order to the warehouse with a short note. Each check takes a few minutes. Nobody had ever added up how many times a day it happens, or how many people do it.

Realisation

The cost of a small task is not set by its duration. It is set by its duration multiplied by its frequency, multiplied by the number of people who do it, plus what it disturbs around it. A task that takes two minutes and happens forty times a day across a team is more than an hour of paid time every day, before counting anything else. Over a working year that is weeks of capacity spent on one step that nobody thought worth examining.

The multiplier is also hidden in three other places. Small tasks interrupt larger ones, so the person who stops to check a payment loses the thread of the quote they were writing. Small tasks done by hand carry a small error rate, and errors multiply with volume too: one mistyped address in a few hundred is a returned delivery, a credit note and an unhappy customer. And small tasks often sit between other steps, so work waits for them, which turns a two-minute action into a half-day delay.

Google's site reliability engineering practice has a useful name for this kind of work. It describes toil as work that is manual, repetitive, automatable, tactical, without enduring value, and that scales linearly as a service grows. The last characteristic is the one that matters commercially. Work that scales linearly with volume means that every new customer, order or employee brings a fixed slice of extra effort with it, so growth never becomes cheaper per unit.

Realisation prompt: pick the three smallest tasks your team does every day. Do you know how many times each one happened last week?

Identification

This is a pattern of unmeasured high-frequency manual work. The visible symptom is a team that is always busy without an obvious reason. The common assumption is that the work is too trivial to matter or too varied to improve. The structural cause is that the business measures work by the size of a single instance rather than by its total volume, so small tasks never reach the threshold where anyone looks at them.

It is related to, but distinct from, two existing Patterns. Why routine admin expands as the business grows looks at how administration accumulates over time. When copy-and-paste becomes part of the job description looks at one specific kind of small task. This Pattern is about the arithmetic itself: how to see which small tasks have quietly become large ones. The broader principle is explored in the Insight why small operational problems become expensive at scale.

Diagnosis

  • Work is sized per instance, not per year. Managers judge a task by how long it takes once. Nobody multiplies by frequency, so a two-minute task never competes for attention with a two-day project.
  • The task compensates for a missing link. Most small repeated tasks exist because one system does not tell another something. The account check exists because the order system cannot see the accounting balance; the status update exists because two tools both track the job.
  • The work is spread thinly. When five people each do a task a few times a day, no single person feels its weight, and no single person has the authority to change it.
  • Interruption and waiting are invisible. The minutes are visible to the person doing the task. The lost focus and the queued work behind it are not.
  • Errors are treated as individual mistakes. When a small manual step goes wrong, the person is corrected and the step stays. The rate of error, which is a property of the step, is never measured.

The Lean tradition calls activity that consumes resources without creating value for the customer muda, and distinguishes waste that cannot yet be removed from waste that can be removed promptly. Many small administrative tasks are of the second kind: they survive not because they are hard to remove, but because nobody has looked.

Commercial impact

The commercial value at stake is labour capacity, productivity, margin, error cost and scalability.

Capacity is consumed in fragments that never add up on any report, so the business hires to relieve pressure instead of removing the work that creates it, a route described in scaling by adding headcount. Productivity falls because skilled people are interrupted many times a day. Margin narrows when every additional order carries the same fixed handling effort. Error cost grows in step with volume. Scalability suffers most: a business whose small tasks scale linearly cannot make each extra customer cheaper to serve than the last.

Common misidentification

The usual response is to ask people to be quicker or more careful, or to write a better procedure. That treats the symptom. A faster manual step is still a manual step multiplied by volume.

The opposite mistake is to buy an automation tool and point it at whatever is most annoying. The most annoying task is not always the most expensive one. Without counting, a business can spend effort automating a weekly irritation while a forty-times-a-day check continues untouched.

Possibility

In a better state, the business knows its high-frequency tasks by name and by number. It has a short list of the small steps that run most often, roughly how long each takes and how often it goes wrong. The most frequent ones have been removed at source: the order screen shows the account status, the stock figure is visible where the order is captured, and the acknowledgement is sent automatically when the order is logged. People still handle the exceptions, which is where their judgement adds value.

For the building-supplies distributor, that might mean a single order view that shows balance and stock together, with orders released automatically when both checks pass and routed to a person only when one fails.

Intervention

  1. Count before you choose. For one or two weeks, have each team tally the small tasks they repeat. A simple sheet with task name and tick marks is enough. Where systems allow it, task mining tools can record step-level activity, but a manual count is a perfectly good start. The Insight how to find the manual work nobody is measuring sets out other methods.
  2. Multiply. For each task, estimate minutes per instance times instances per day times working days. Add a note on how often it goes wrong and what it holds up.
  3. Ask why the task exists. For the top five, identify the missing information or link that makes the task necessary.
  4. Remove, then automate. First ask whether the step can be dropped or merged. Only then consider automating what remains, using the approach in how to calculate the real cost of a repetitive task to decide whether it is worth it.
  5. Keep the exceptions human. Design the change so that routine cases flow through and unusual cases reach a person with the context they need.
  6. Re-count. Repeat the tally after a month to confirm the task has actually gone, not simply moved to someone else.

Diagnostic questions

  • Which tasks in your business happen more than twenty times a day?
  • Could anyone tell you how many times a specific check was performed last week?
  • How many of your small tasks exist only because one system cannot see another?
  • When volumes rose last year, did handling effort per order fall, stay the same or rise?
  • When a small step goes wrong, do you correct the person or examine the step?

Bottom line

Small tasks become expensive because their cost is multiplied by frequency, people, interruption, errors and waiting, while attention is allocated by the size of a single instance. A business that counts its smallest repeated steps usually finds that a handful of them consume more capacity than any single large process. Removing those few tasks at source frees time, reduces errors and lets growth become cheaper per customer rather than proportionally more expensive.

Sources and further reading

Related Mellorca reading

Article summaryA two-minute task looks too small to fix, but its cost is set by how often it runs, who does it and what it interrupts. Counting the multiplier shows which small tasks have quietly become large ones.
Explore our solutionsStart a conversation