Small founders ask me what accounting software to use, and they expect me to recommend whatever is most convenient for my own office. I do not, because the tool that suits a CA firm is usually overkill for a business with a handful of transactions a week.
Here is what I actually suggest, by stage, and โ more usefully โ where I tell people not to bother.
Stage one: you have under a hundred transactions a month
Use a spreadsheet. I mean it.
A clean spreadsheet with date, description, amount, category and a link to the supporting document beats a badly maintained accounting package by a wide margin. The failure mode at this stage is never that the software was insufficient. It is that nobody entered anything for four months.
Spreadsheets win early because everyone already knows how to use one, there is no subscription to justify, and the barrier to recording a transaction is close to zero. Your accountant can work from a clean sheet perfectly well. What they cannot work from is a proper accounting system with three months of gaps.
Two rules make this work: record weekly, not monthly, and keep the supporting document at the same time as the entry, not later.
Stage two: you are invoicing customers and chasing payment
This is where a tool earns its subscription, and the reason is receivables rather than accounting. Chasing payment by memory is where small businesses lose real money.
What matters at this stage: invoices that are compliant and numbered in sequence, a view of who owes you what and for how long, automatic reminders, and a clean export your accountant can use. What does not matter: inventory modules, project costing, multi-currency consolidation, or anything described as an enterprise feature.
Pick something mainstream, ideally one your accountant already works with, and use the parts you need. The specific brand matters far less than whether it is kept current.
Stage three: you have staff, GST and regular filings
Now you need a system of record rather than a tool, and the selection criteria change:
- It must handle GST properly and produce the data your returns require without manual rework
- Bank feeds or easy statement import, because manual entry of bank lines is where errors enter
- Payroll that calculates statutory deductions correctly, or a separate payroll tool that does
- A clean export in a format your accountant can actually use
- Your data must be extractable if you leave. Test this before you commit, not after
That last point is the one founders never check and later regret. Being unable to get your own history out of a system is a genuine cost when you change accountant, change software, or go through diligence.
What I tell people to skip
Anything that promises to replace your accountant. Tools produce records. They do not take positions, and they do not sign anything. The software has no view on whether an expense is allowable.
Multiple overlapping tools. I regularly see a small business running an invoicing tool, a separate expense app, a spreadsheet and a bookkeeping package that no longer agree with each other. Reconciling four partial truths is worse than maintaining one imperfect record.
Elaborate categorisation. Forty expense heads on a business with sixty transactions a month produces slow, inconsistent bookkeeping. Ten well-chosen categories give you every insight you will actually act on.
Anything bought because of a feature you will need “later”. Later rarely arrives in the shape you predicted, and you will have spent two years paying for and navigating complexity you did not use.
The habits that matter more than the tool
Separate the business account from the personal one, from day one. This single habit saves more accounting time than any software choice, and mixing them is the most common reason a small set of books takes days to prepare rather than hours.
Reconcile the bank monthly. Not quarterly, not when the return is due. A discrepancy found within a month can be explained; one found eleven months later usually cannot.
Keep documents where the entry is. A number without a supporting document is an assertion, and assertions are what get disallowed.
The honest conclusion
I have seen well-run businesses on spreadsheets and chaotic ones on expensive software. The tool is not the variable. The discipline is.
Choose the simplest thing that captures what you need, use it consistently, and upgrade when the current approach is genuinely breaking โ not because a feature list is persuasive.
If you want a setup reviewed before it becomes a cleanup job, or your books have drifted and need reconstruction, AeTx does both, and will tell you honestly if your spreadsheet is fine as it is.