← all notes

when to stop adding to the spreadsheet

Every growing field business hits the point where the workbook that got it here starts holding it back. Five signs you’ve reached it, what the research says about spreadsheet errors, and what to do next.


Spreadsheets are brilliant. They’re how most field businesses get organised in the first place: the crew schedule, the equipment list, the job tracker. There’s no shame in running on one for years. The trouble starts when a business grows around one and nobody notices it has quietly become the most important system you own.

In June 2003, TransAlta, the Calgary power generator, told investors a bidding mistake had cost it about US$24 million. Its chief executive called it a cut-and-paste error in an Excel spreadsheet that wasn’t caught before the bids went in, and the company took a C$33 million pre-tax charge for the quarter. A routine copy-and-paste went wrong, and nothing in the workbook was set up to notice.

That’s the uncomfortable part. Spreadsheet errors aren’t a sign that someone is bad at Excel. They’re the normal state of things. Ray Panko at the University of Hawaii has studied this for decades. When he pooled field audits of real business spreadsheets, errors turned up in 94% of them, with roughly 5% of formula cells wrong. Human error research generally puts the expected rate somewhere between 1% and 5% of cells, and in his experiments people with hundreds of hours of spreadsheet experience didn’t do meaningfully better than beginners.

So a workbook with a few hundred formulas almost certainly has some wrong ones. The real question is whether your business is set up to catch them. Here are the five signs I look for that it isn’t.

1. there’s more than one “real” copy

The office has one version, the field supervisor has another, and there’s a third on someone’s laptop from before the last job. If people email versions around, or you’ve ever asked “which one is the latest?”, you’re already paying for it in mistakes you haven’t found yet.

Fixing an error in one copy doesn’t fix it in the others. Every extra copy is another place a correction has to be made by hand, and another place it won’t be.

2. one person understands it

When they’re on holiday, things slow down. When they leave, you find out how much of the business lived in their head: what column R means, why row 40 is hard-coded, which tab the field guys actually look at.

In Alberta this isn’t hypothetical. BuildForce Canada expects about 43,400 of the province’s construction workers, roughly 21% of the 2024 workforce, to retire by 2034. Not all of them run spreadsheets, but some of the knowledge leaving with that generation is exactly the kind that only ever got written down in a workbook.

3. you copy the same numbers somewhere else

Quote to invoice. Timesheet to payroll. Job sheet to report. Schedule to the crew. Every retype is time lost and another chance to get it wrong, and TransAlta’s problem was nothing more exotic than this: data moved by hand from one place to another, with no check at the end.

The spreadsheet isn’t the expensive part. The hours spent working around it are.

4. people are scared to touch it

Slow to open, crashes on a big filter, or full of macros nobody dares edit. When the answer to “can we add a column for this?” is “better not”, you have a system that has stopped growing with you. The business keeps changing; the workbook doesn’t, so people start keeping side lists to cope. That’s how sign number one begins.

5. you need to know who changed what

Ticket expiry dates, inspection records, equipment certifications, rates on a signed contract. Once money, safety or compliance depend on the numbers, “someone must have overwritten it” stops being an acceptable answer. A spreadsheet can be made to keep a history, but most of the ones I see don’t, and nobody finds out until an auditor or a client asks.

what it’s actually costing you

The cost rarely shows up as one big mistake. It shows up as hours: checking, reconciling, re-keying, phoning the field to confirm what the sheet should already say. In Statistics Canada’s spring 2026 business survey, 24.4% of businesses expected recruiting skilled employees to be an obstacle over the following three months. If you can’t easily hire more people, the hours your existing team spends working around a workbook are hours you can’t buy back.

If you want a rough number for your own business, the free admin cost calculator takes about a minute.

so what next?

Not necessarily custom software. Start by making one copy the only copy and giving it an owner: one shared file, one person responsible for its structure, and everyone else editing data rather than layout. That alone fixes more than people expect.

Then check whether something off the shelf already does the job. Scheduling, timesheets and equipment tracking all have mature products, and if one fits how you work it will be cheaper than anything built for you.

Build only when the way you work is genuinely specific to you. When you do, build the smallest thing that removes the retyping, and leave the rest in the spreadsheet until it earns a change.

If you’re not sure which of those you need, that’s exactly what a Systems Diagnostic is for.

sources

— Bryan, 05 oct 2026

got a spreadsheet like this?

Thirty minutes, free. We’ll tell you if it’s worth replacing.

Book a free call