Two different solopreneurs in Vancouver asked me on the same day what number they should be watching. A transaction count, a revenue figure, something that would clearly tell them when to get the bookkeeping software.
I personally don't think there is one, which makes this the wrong question to ask.
The spreadsheet does not stop working one day. It has never actually been free either. It has just been charging you indirectly. So the better question to ask is whether it costs more than the subscription.
For someone with paying customers and a business to grow, the answer is probably a yes.
The biggest cost is not the money
The subscription is not the most expensive part. A solopreneur's biggest asset is their own time and attention, and a spreadsheet asks for a great deal more of both than the software ever would.
Manual entries take time. So does building a report, tallying a category, or pulling together whatever your accountant asks for at year end. Every one of those is also a place to make a mistake. Then down the line you lose an evening working out whether it is a real error or just how you recorded things back in April.
Plus most of those manual entries don't need to happen at all. Connect the software to your bank and the transactions pop up on their own.
A file you update at a later date can only tell you about a period that has already closed, long after the answer would have been worth something to base decisions on. Properly maintained software, with the bank feed connected and updated as transactions occur, stays current and tells you the state of your books and your business in real time.
The software is built for exactly what you are doing
Small business bookkeeping software is intentionally designed for a non-accountant keeping their own books. It simplifies and even automates a lot of the accounting side of things for this particular reason. And even though that leads to its own problems in the future, those are a lot easier to fix than a manually built Excel file.
The clearest example is one you never notice working. Software will not let you save an entry that does not balance. Every transaction has to account for both where the value came from and where it went. A spreadsheet enforces none of that. Money can easily go missing inside one, or get counted twice, and nothing anywhere flags it.
Human error on a spreadsheet is a matter of when
The spreadsheet has no opinion about what you type into it. It will take a transposed digit, a decimal in the wrong place, or a positive where you meant a negative, and carry all of them forward without flagging any of it.
Then there are the failures that have nothing to do with your entry errors. Delete a row that a formula somewhere else depends on and the total turns into a #REF! error. That version at least reveals itself.
The hidden one is much worse. Add a new row beneath a column being totalled and the new figure can fall outside the range the formula covers, without anyone noticing. Or you accidentally save a number in a cell that used to hold a formula, and it stays hardcoded for the rest of the year.
Invoices are where mistakes reach the client
Errors are bad enough, but it gets a whole lot worse once a client is involved.
This can happen in numerous ways, but possibly the most expensive version is thinking somebody has already paid when they haven't. You notice months later, if at all, and now you are chasing payment a long time after goods or services rendered. That is an awkward conversation, and if you do manage to get a hold of them again, it costs you goodwill with a client who reasonably expected you to be on top of it.
The software numbers invoices automatically and shows you what is outstanding and what is paid in real time.
The move is inevitable if you grow
Everyone who starts on a spreadsheet eventually moves off it if they stay in business, and moving multiple years of history into the software is a real project. Unreconciled spreadsheets are the version a bookkeeper has to reconstruct rather than simply migrate, and that is the main difference between a small fee and a very large one.
So in the end most people decide to just leave it behind and start fresh. Either the cost of bringing the data across is too high, or by the time they decide to switch the old file has become too messy to bring anywhere. Then a year on, when you want to see how this year end compares to the previous ones, you basically can't in any meaningful way.
The receipts end up where they belong
If you are ever audited, the CRA will want to see the receipts sitting behind what you claimed. Attaching them to transactions as you go is a much simpler exercise than digging through archives and storage to try to find the right ones two years later.
It also means a return built from records kept as they happened, rather than one you scramble to assemble afterward.
Everything with a deadline gets easier
Once the business has obligations beyond the annual return, tracking them is no longer optional. GST/HST collected, ITC to be claimed. Source deductions if you have staff, calculated for every paycheck and remitted on a schedule. Amounts paid to contractors, which may need reporting on a T4A. Instalments. And everything your accountant needs in the spring.
A spreadsheet will not automatically update based on a government rule or rate change, and it will not remind you when a remittance is due. Late payroll remittances carry graduated penalties that climb the longer they sit, and they apply whether the delay was deliberate or you simply forgot.
Software keeps a running total of each of these instead of making you assemble one. The GST return comes out of transactions already coded. Payroll works out the deductions and tells you what is owed and when. Contractor payments are already totalled per person by the time you need the slips.
The little expense
Benjamin Franklin has a line that gets quoted at small businesses constantly. "Beware of little expenses; a small leak will sink a great ship." It is nearly always used to argue against spending. Here I'm making the opposite case. The subscription is not the leak.
The leak might be the extra hours spent entering things that could be automated. The evening lost trying to solve the mystery figure months later. The client who ghosted instead of settling the unpaid invoice. The penalty on a remittance never flagged. The credits you never claimed, or had clawed back, because you can't find the receipt. The cleanup fees when somebody finally has to reconstruct all of it. The years you can no longer compare against each other, because the file they lived in was never brought across.
So yes, I advocate moving onto the software as soon as possible. Even the free ones, with all their limitations, will plug most of those leaks before they start.