Spreadsheets scale further than people expect, then stop working fairly suddenly. These are the signals we look for when advising a client it's time for a real system.
Spreadsheets are genuinely good software, and plenty of businesses run on them far longer than a consultant would admit. But they scale in a particular way: fine, fine, fine, then abruptly not fine. Recognising that turn early saves a lot of money.
Someone spends their week copying data between files
This is the clearest signal. When a role exists mostly to move numbers from one sheet into another and reconcile the differences, you're paying a salary to paper over a systems problem. That work is also where most errors enter your data.
Nobody's certain which file is current
Once filenames include dates, initials, or the word 'final', you've lost a single source of truth. Decisions start getting made from different numbers depending on who opened which copy.
Reports take days instead of seconds
If answering 'how did last month go?' requires someone to assemble it by hand, you'll ask the question less often than you should. Businesses quietly stop measuring things that are painful to measure.
One person is the only one who understands it
Complex spreadsheets accumulate logic that lives entirely in one person's memory. That's a real operational risk when they're on leave, and a serious one when they leave for good.
What to build first
The mistake here is trying to replace everything at once. We usually start with the single workflow causing the most manual effort, get that working properly, then expand. It's faster to get value from, easier to fund, and your team learns the new system gradually instead of all at once.
Done this way, the switch is a series of small, reversible improvements rather than a single risky cutover.
