It's 10:47pm on a Thursday and Niamh has 47 tabs open. Not browser tabs - Excel tabs. One per engineer, plus three reference sheets, plus a master pivot, plus the consolidated payroll export.
This is what month-end mileage reconciliation looks like in a lot of Irish engineering firms - more of them than anyone wants to admit.
I'd like to walk you through Niamh's Thursday night, because the shape of the evening is the thing that's hard to explain to anyone who hasn't lived it. The shape is the problem - not the individual tabs.
The 47 tabs
The first 30 tabs are one per engineer. Each engineer's tab has three sub-sections: a journey log (date, start, end, distance, project), a calculations table (which band, which rate, which amount), and a running year-to-date column.
The next four tabs are reference sheets. Civil Service rate bands. Engine-CC categorisation. Project codes and which client each maps to. The expense sub-categories (subsistence, tolls, parking).
The next ten tabs are consolidations. One per project, summing the mileage allocated to that project across all engineers, broken down by month. These tabs feed the client invoicing side of the house. (This is also where six-figure project margin leaks tend to hide - the kilometres get coded once, then never reconciled against the project records.)
The last three tabs are the master views. A pivot of all engineers' totals for the month. A reconciliation tab that ties the master pivot to the per-engineer tabs. And the payroll export tab - the one that goes to payroll on Friday morning.
None of these tabs are wrong individually. The problem is what happens when one engineer's tab gets edited.
The Limerick trip Tom forgot to log
Tom is one of the project leads. He drove from Galway to Limerick on the 18th, three weeks ago, to do a site walk for a new project. He forgot to log it. Ran into a wall on the deadline for last month's claim. Mentioned it in passing to Niamh on Wednesday.
"Just stick it on this month's, sure," Niamh said, the way anyone would.
Tom adds the trip to his current tab. 240 km, Galway to Limerick return. Project code: the new one. The system - in this case, Niamh's master pivot - now thinks Tom drove that trip in the current month.
That's the first error. The 240 km belongs to last month's year-to-date, not this month's. It's not a huge error in isolation - 240 km doesn't push Tom across a band threshold by itself. But Tom did three other trips in the gap between then and now, and one of those was a 380 km job to Donegal that pushed him into Band 2. The Donegal trip got calculated with a YTD that didn't include the missing Limerick. So the band split on the Donegal trip was wrong. (This is exactly the missing-trip pattern that throws every subsequent calculation off, and it's invisible until someone reconciles.)
Niamh doesn't know any of this yet. She's looking at this month's tab. The error is in last month's, and the consequence is in last month's payroll export, which has already gone.
The 12-month-old underpayment
It's now 11:30pm. Niamh is reconciling the master pivot against the per-engineer tabs and one number doesn't tie. Eileen's running YTD shows 8,200 km. The master pivot says 8,460 km. A 260 km gap.
This is the kind of thing that happens when someone edits a per-engineer tab and forgets to refresh the pivot. Or when a row was deleted and the YTD column didn't recalculate. Or when a paste-as-values operation last March overwrote a formula somewhere.
Niamh starts at the top of Eileen's tab and walks down. Trip 1 in January, 410 km, fine. Trip 2 in January, 500 km, fine. Trip 3 in late January, 320 km, fine. Trip 4 in early February, 470 km. The YTD column says 1,700 km, which means Trip 4 should have been a band split - 100 km at Band 1, then 370 km at Band 2.
The amount column says €243.81. That's the Band 1 rate for the full 470 km. The trip wasn't split.
Eileen has been underpaid by €144 since February.
It's twelve months ago. The error has been sitting in the spreadsheet for a year. It propagated forward - every subsequent trip's YTD assumed the wrong total, which means every subsequent band split was off by the same 100 km offset. The cumulative underpayment for Eileen across the year is somewhere between €600 and €900, depending on how many band crossings happened.
Niamh closes Eileen's tab. Opens the master pivot. The 260 km gap she came in to investigate is now the smaller of two problems.
The conversation the next morning
It's Friday, 8:45am. Niamh hasn't slept. She catches Eileen by the coffee machine.
"Eileen, do you remember February last year? You did a 470 km trip - Donegal, I think."
"Could be. Why?"
"It looks like that one didn't get the band split applied. I think you've been underpaid since then. I need to work out exactly how much."
Eileen takes a moment. "How much, roughly?"
"Somewhere between six hundred and nine hundred euro across the year. I'll know when I rebuild her - I mean, when I reconcile."
The look on Eileen's face is the thing. It's not anger. It's not surprise. It's something closer to "of course." Engineers tend to be good with numbers, and most of them have suspected for a while that the spreadsheet doesn't quite tie. They just don't have the time to prove it.
The trust hit isn't the underpayment. The trust hit is the realisation that Niamh's process has been quietly wrong for at least a year.
The pattern across firms
I've talked to a lot of finance managers in Niamh's position. Different firms, different engineers, different specific errors. The shape of the problem is the same one every time.
The first contributing factor is that mileage spreadsheets are append-only by design. You add this month's trips to the bottom and let the YTD column do the maths. They're not designed for retroactive editing. When a trip arrives late, the spreadsheet can't recompute the chain backward.
The second factor is that the calculations are technically correct - until they aren't. The VLOOKUPs and CUMSUMs and IF-statements work fine when the data flows in clean and on time. They break silently when someone pastes values, deletes a row, or adds a trip out of order.
The third factor is that nobody owns the cross-validation. The per-engineer tabs are the engineer's view. The pivot is the finance lead's view. The payroll export is payroll's view. When the three views disagree, there's no single person whose job it is to reconcile them. Niamh ends up doing it because nobody else will.
What 6:15pm Thursday could look like
The other version of Thursday isn't dramatic. It's quieter.
The mileage system has been calculating year-to-date in real time for every engineer all month. Every trip submitted recalculates the chain forward and backward. Late trips and out-of-order trips trigger an automatic recalculation of every subsequent trip's band split. The pivot doesn't exist - the master view is just a database query.
Niamh's Thursday job is reviewing the pre-built reconciliation report. The report shows: total kilometres claimed by each engineer, total amount payable, project allocation breakdown, exceptions flagged for manual review (an unusual route, a trip that needs receipts that haven't been uploaded, a vehicle that needs a roadworthiness check renewed).
The exceptions list is short. She works through it in 40 minutes. Approves the lot. Clicks the export button. The CSV arrives in the payroll inbox in the format payroll wants.
She closes the laptop at 6:15pm.
The bottom line
The 47-tab spreadsheet is the symptom of finance teams burning out at month-end - the cause is that mileage is harder than spreadsheets are designed for, and most Irish SMEs grew into the problem rather than choosing to solve it.
If you recognise Niamh's Thursday in your firm's Thursday - the one where reconciling against the master pivot turns into a 90-minute archaeology job - the issue is the tool, not the person using it.
If you'd like to see what the 6:15pm Thursday looks like, book a 20-minute demo and we'll walk through it. (And if you're thinking "we'd switch in January" - a mid-year migration is more straightforward than people expect, with a clear week-by-week shape.)
Last updated: 28 April 2026.
