Where mid-market companies lose money to manual work

Nobody budgets for manual work. It never appears as a line item, because it is paid for through salaries the company would pay anyway. A supplier who overcharged you by $17,000 a month would be gone in a quarter. The same amount lost to re-typing, chasing, and assembling survives for years.

No invoice ever arrives for manual work, which is exactly why it survives every budget review.

In the operations we analyze, the leak concentrates in five places. Below is each one, with worked numbers for a hypothetical 120-person company at a loaded labor cost of $40 an hour, which is roughly what a $65,000 salary costs the business once benefits and overhead are included.

Data re-entry between systems

30 hrs

Manual reporting cycles

25 hrs

Invoice and document handling

20 hrs

Approval chains

15 hrs

Customer follow-up from memory

10 hrs

Data re-entry between systems

The sales system, the accounting system, the inventory system, and the support desk were all bought separately, so people became the integration layer. An order confirmed in one system is re-typed into two others. A new customer is created three times.

The cost hides because each individual act is small. Ninety seconds here, four minutes there, spread across a dozen people, none of whom would describe re-typing as their job. Then there is the second cost: every re-typed field is a chance to introduce an error, and errors create their own downstream work.

Worked example: six people across sales operations, finance, and the warehouse each lose five hours a week moving data between systems. That is 30 hours a week, roughly $5,200 a month, about $62,000 a year, for work that produces nothing new.

The automated version: systems are connected so a field entered once propagates everywhere it is needed. The order exists in accounting the moment it is confirmed in sales. The 30 hours do not shrink; they disappear, along with the transcription errors.

Manual reporting cycles

Every month, managers export spreadsheets, paste them into a template, reconcile totals that disagree, and format the result for a leadership meeting. The numbers are days old by the time anyone sees them.

The cost hides inside management salaries, which makes it the most expensive labor in the building doing the least valuable version of its job. Report assembly feels like diligence, so it is rarely questioned.

Worked example: five department managers each spend three hours a week pulling and reconciling figures, and one analyst spends ten hours a week assembling the monthly pack. Call it 25 hours a week, roughly $4,300 a month. The deeper cost is decision lag: a problem visible on day two of the month is acted on in week four.

The automated version: dashboards draw from live data, and the monthly pack assembles itself. Managers spend their hours on what the numbers mean, not on producing them.

Approval chains

Purchase requests, discount approvals, contract sign-offs. Each one waits in an inbox until someone remembers it, and someone else spends part of their day asking where it is.

This cost hides twice. The chasing is invisible labor, and the waiting is invisible drag: quotes expire, orders ship late, and rush charges appear because a routine approval sat for six days.

Worked example: across the company, staff spend 15 hours a week checking status and nudging approvers, roughly $2,600 a month. Add the hard costs of delay, such as expedited shipping and lapsed supplier quotes, and the real figure is higher.

The automated version: requests route by rule. Anything under an agreed threshold approves automatically, everything else goes to the right person with a deadline, and unanswered items escalate on their own. Approval time drops from days to hours, and nobody chases anything.

Invoice and document handling

Supplier invoices arrive as email attachments. Someone opens each one, keys it into accounting, matches it to a purchase order, and files it. Contracts and delivery notes follow the same path.

The cost hides in a role everyone assumes is necessary, plus a slow drip of penalties: late-payment charges when invoices sit in inboxes, and early-payment discounts missed for the same reason.

Worked example: around 400 supplier invoices a month at eight minutes each, plus contract and delivery-note handling, comes to roughly 20 hours a week, about $3,500 a month, before a single penalty or missed discount is counted.

The automated version: documents are read by software, matched to purchase orders automatically, and posted without human touch. People handle only the exceptions, which is typically fewer than one in ten documents.

Customer follow-up done from memory

Renewal dates, open quotes, promised call-backs. In many mid-market companies these live in the heads and notebooks of account managers, and the system works right up until it doesn't.

This is the leak where labor is the smaller half of the cost. The hours spent manually checking who needs a call are measurable; the renewal nobody called about is not, until the customer leaves.

Worked example: account managers spend 10 hours a week reviewing accounts and reconstructing who was promised what, roughly $1,700 a month. One missed $20,000 renewal a year makes the labor cost look like a rounding error.

The automated version: follow-ups trigger from dates and events, not memory. A renewal 90 days out opens a task automatically. A quote untouched for a week prompts a call. Nothing depends on anyone remembering anything.

How to size the leak in your own business

The example company loses about 100 hours a week across the five areas, the equivalent of two and a half full-time roles, over $200,000 a year. Your number will differ, and you can get a defensible version of it in one week with three counts.

Count handoffs. Pick your three most common processes and count every point where work stops and waits for a person to move it. Each handoff is queue time plus chasing time.

Count re-typed fields. Follow one order, one invoice, and one new customer through your systems. Every field typed a second time is integration work being done by hand, at salary rates, with an error rate attached.

Count report-assembly hours. Ask every manager how long the recurring reports take each month. Sum it. This number surprises almost every leadership team that collects it.

Multiply the total hours by your loaded hourly cost, then by 52. That is the size of your leak, and it is the figure any automation effort should be judged against. Our operational efficiency work starts from exactly this arithmetic, and a free analysis will produce these counts from your own operations. On cost: the analysis and strategic review are free, and beyond that you don't pay unless verified savings exist. The mechanics are on the how it works page.