The Real Cost of Repetitive Work in a Growing Company

Why is the cost of repetitive work so easy to miss?

Because it never arrives as a bill. Nobody raises an invoice for twenty minutes of retyping delivery details, or for the third time this week someone opened a WhatsApp thread to find an address. The work is spread across a dozen people in small pieces, each too small to complain about. It rises quietly with volume, so it feels like the price of doing well.

It is also the work that never gets escalated. A team that is behind on collections will say so. A team that spends the first hour of every morning copying yesterday’s orders into a sheet will simply come in earlier. By the time it becomes visible, it has usually taken the form of a hiring request, a missed month end or a good person handing in their notice.

How do you count the visible hours honestly?

Watch the work rather than asking about it. People underestimate frequency and overestimate their own speed, in both directions and usually by a lot. Sit with one person for a morning, count the times the task occurs, and time the whole loop: opening the file, finding the missing detail, entering it, checking it, chasing the person who did not reply, and doing it again when the detail turns out to be wrong.

Then do the plain arithmetic and label it as an estimate. Suppose four people each spend two hours a day on order entry and follow-up. Over 22 working days that is 176 hours a month, or roughly one full-time person. Multiply by a fully loaded hourly cost and you have a figure worth discussing. The team-capacity calculator on our homepage runs that same calculation with your numbers; it assumes 22 working days a month and shows capacity, not a saving.

Which costs never reach a timesheet?

The hours are the part everyone can see. The rest is spread across the business as delay, error and irritation, and each piece can be observed if you decide to look for it.

Elapsed time against touch time is the most revealing pair. A quotation that takes forty minutes of work but four days to reach the customer is not a forty-minute problem. Most of the cost is in the queue, and the queue is invisible on every timesheet ever written.

What does it cost customers, cash and the owner?

Customers feel repetitive work as waiting. The quotation that arrives on Thursday instead of Monday, the order confirmation that needs a reminder, the delivery date nobody can confirm without ringing the warehouse. You rarely hear about it. People who are made to wait tend to go quiet rather than complain, and the effect appears later as a smaller share of a customer’s business.

Cash feels it too. An invoice raised three days late is paid three days late, and month end becomes a scramble: reconciling, chasing missing documents, correcting entries before filing. The work itself is not difficult. It is simply done by hand, at the worst possible time, by the people who can least afford the interruption.

• Owners become the exception desk, answering the questions the process cannot.
• Growth is met by hiring, because adding a person is the only lever anyone can see.
• Each new person learns steps that were never written down, from someone already busy.
• Senior people spend evenings on work that a junior could do if the process were clear.

How do you measure a baseline you can defend?

Pick one unit of work and stay with it: a purchase invoice, an enquiry, a proof of delivery, a dealer claim. A baseline about “operations” cannot be tested. A baseline about one document can.

1. Volume: how many of these a month, from the system rather than from memory.
2. Touch time: minutes of actual work per case, timed on a normal day and a busy one.
3. Wait time: elapsed time from arrival to completion, which is usually the larger number.
4. Rework rate: how often a case is reopened, corrected or entered twice.
5. Exception rate: how many cases leave the standard path, and for what reasons.
6. Who measured, over which period, and what they excluded.

Include the awkward cases. Every process has a version that takes six times as long, and it is often the reason the team is tired. Then show the numbers to the people who do the work. They will correct you, and the corrected baseline is the one worth carrying into a decision. Our guide to identifying repetitive work worth automating sets out how to compare candidates once you have several baselines.

What does the arithmetic not tell you?

It does not tell you what the freed hours are worth. That depends on a decision you have to make before you start: the time absorbs growth without a new hire, or it moves to work you cannot buy in, such as supplier negotiation and customer relationships, or it ends overtime and weekend catch-up. Say which one you are aiming at, and you can check afterwards whether it happened.

Then subtract the system’s own cost: building it, the software it runs on, the people who review its exceptions, and the maintenance when a form or a rule changes. A system that halves the work but requires every case to be checked has not halved anything. This is why a cost baseline and an operating estimate belong in the same conversation.

Why should some repetitive work stay manual?

Adoption is no longer the hard part. The 2026 AI Index reports organisational adoption of AI reaching 88%, which makes the choosing more important than the adopting. India’s Economic Survey 2025–26 frames the national question as not whether to adopt AI but how to pace its diffusion, warning that rapid, uncalibrated deployment risks displacing people faster than the economy can reabsorb them. Inside a single company the same caution reads as a practical instruction: pick the work deliberately.

• Low volume: twice a month is a checklist, not a system.
• Rules that keep changing: automate a process that has stopped moving, not one still being argued about.
• The task is the control: a person eyeballing every dispatch may be the only check you have.
• The task is how people learn: juniors often learn the business by handling the small stuff.
• The contact is the product: a call to a long-standing customer is not overhead.
• The cost exceeds the benefit: some work is annoying and cheap, and should stay that way.

What do you do once you have the number?

Rank the candidates by three things together: how often the work occurs, how long it takes end to end, and what it costs when it goes wrong. A high-volume task with mild consequences is a better first project than a rare one with severe consequences, because you can test it safely and see the result within weeks.

Then try the cheap fixes first. A shared inbox rule, a form that collects the missing field at the start, a template, a report switched off, or a connection between two tools you already pay for will sometimes remove most of the cost for a fraction of the effort. If the work survives that test and still costs what your baseline says, the AI Opportunity Blueprint is a paid diagnostic that maps the workflow, sets a cost baseline, ranks the opportunities and gives you a recommendation with a 90-day roadmap.