Most growing businesses lose something like 20 hours a week to small, repeated admin: copying data between systems, chasing approvals and rebuilding the same reports. Nobody budgets for it because it never appears as a line item, only as a full calendar. The fix starts with measuring your own number, which takes one week and a shared spreadsheet.
How much time does manual admin cost a small business?
When we run a time audit with a new client, the biggest single bucket is almost never a dramatic bottleneck. It is small, repeated friction: re-entering the same customer details in a second system, pinging someone for a sign-off that could be a rule, rebuilding last month’s report because the data moved. Each task takes five or ten minutes, so nobody flags it, and that is exactly why it compounds.
Treat “20 hours a week” as an order of magnitude rather than a benchmark. Here is how a figure like that forms in a team of six. If each person spends 30 minutes a day on copy and paste work and another 15 minutes chasing approvals or status updates, that is 45 minutes a day, or about 3.75 hours a week per person. Across six people it comes to roughly 22 hours a week, which is more than half of a full-time role.
Now put a cost on it. At a loaded hourly cost of AUD 45 (salary plus super, leave and overheads), 22 hours a week is about AUD 990 a week, or close to AUD 47,500 over a 48-week working year. Your rate and headcount will differ, so run the same multiplication with your own figures: hours per week × people × loaded hourly cost × working weeks.
Where the hours actually go in a growing operation
In a mid-size operation the hours cluster in three places.
The first is moving data between systems. A lead arrives through a website form or WhatsApp, someone types it into HubSpot or Pipedrive, then again into a Google Sheet for the weekly meeting, and later into Xero or QuickBooks when the invoice goes out. Every handoff is a manual copy, and every copy is a chance to introduce an error that someone has to find later.
The second is approvals and follow-ups. Discount sign-offs, purchase requests and expense approvals often depend on one person reading a Slack message. Most of these decisions follow a pattern the approver could write down in two lines (under a certain amount, from a known supplier, approve), yet the request still waits in a queue.
The third is recurring reports. The Monday sales summary or the month-end operations pack gets rebuilt by hand because the data lives in four places and the person who knows the formulas is the only one who can assemble it.
None of it shows up on an org chart. All of it shows up on a calendar.
How to measure your own number in one week
You don’t need a consultant or time-tracking software to find out. You need one week and a shared sheet.
Pick the people who touch the most systems, usually operations, sales support, finance and customer service. Give them a Google Sheet or a simple Notion database with five columns: task, system involved, minutes spent, how often it repeats, and whether it follows a clear rule. Ask them to log anything repetitive that takes more than five minutes, at the end of each day rather than in real time, so the logging itself doesn’t become another chore.
At the end of the week, sort by total minutes and group similar tasks. You will usually find that a handful of tasks account for most of the hours, and that the people doing them already know which ones are pointless. Multiply the weekly total by your loaded hourly cost and you have a number leadership can weigh against anything else in the budget.
You are already paying for these hours; you just never decided what they should buy.
Adding it up is usually the first moment a leadership team takes automation seriously. That is not because the number is shocking, but because every hour on the list is one they already recognise.
What to automate first
Not every item on the list deserves automation. Rank each one by hours per week and by how clearly it follows a rule, and start where both are high.
Copying between systems is usually the easiest win. A workflow in n8n, Make or Zapier can push a new form entry into your CRM, add the row to the reporting sheet and notify the right person in Slack, with no retyping. Approvals that follow a rule can become that rule: requests under a threshold are approved automatically and logged, and only the exceptions reach a person. Recurring reports can be generated from the source systems on a schedule, with Claude or ChatGPT drafting the written summary so the manager only reviews it.
Tasks that need judgment, such as negotiating with a key account or deciding on a hire, stay with people. The goal is to give those people back the hours that were going to work a rule could have handled. One customer-service portal we worked on used to consume hours of manual work; it now resolves routine requests automatically and returns about 30 hours a month to the team.
Where to start this week
Run the one-week log. Choose three to five people, share the sheet on Monday and review it together on Friday. Pick the single task with the most hours that follows a clear rule, automate only that one, and measure the hours again a month later. If you want a second pair of eyes on the list, JARA can run the audit with you and help choose the first workflow.
FAQ
Is 20 hours a week a real benchmark?
No. It is an illustrative order of magnitude based on how small daily tasks add up across a team. Your figure could be lower or much higher, which is why your own one-week log matters more than any published average.
How long does a time audit take?
One working week of logging plus about an hour to sort and review the results. Logging at the end of each day keeps the effort to a few minutes per person.
Which tasks should not be automated?
Anything that depends on judgment, relationships or context that changes case by case. Automate the rule-based steps around those tasks, such as collecting information and routing it, and leave the decision with a person.
Do we need new software to fix this?
Usually not. Most of these hours sit between tools you already pay for, such as your CRM, accounting system and spreadsheets. Workflow tools like n8n, Make or Zapier connect them so data is entered once instead of being retyped.