Choose accounting software by writing down the transactions your business actually performs, then finding the cheapest product that handles all of them without add-ons. That sounds obvious, and it is the opposite of how most decisions get made. The usual approach starts from a feature list or a recommendation, which is how businesses end up paying for capability they never use while still keeping stock in a spreadsheet.
Start with your transactions, not with features
Spend an hour writing down every kind of money movement in your business. Not categories like 'invoicing', but actual events: a customer pays half now and half in thirty days; a supplier delivers short and we return two units; a staff member gets a salary advance; stock gets damaged in the warehouse. This list is your requirement specification, and it is far more useful than any vendor's feature matrix because it is specific to you. Any product that cannot record one of these events cleanly will generate a spreadsheet workaround, and workarounds are where books go wrong.
- List real events, including the awkward ones
- Include partial payments, returns, write-offs and adjustments
- Note which events happen weekly versus once a year
- Flag anything you currently handle outside the ledger
Compare total cost, not licence price
The headline subscription is rarely the real number. Add every module you would need to license separately, every integration subscription, any implementation or consultant time, and the per-user cost at the team size you expect in a year. A cheap ledger plus an inventory add-on plus a payroll service plus an email tool frequently costs more than a single product that includes all four, and it certainly costs more in administration. Build the comparison as an annual total for your actual team.
- Add inventory, payroll and reporting add-ons into the base price
- Include integration subscriptions and any sync tooling
- Price it at next year's headcount, not today's
- Count implementation and training as real costs
Test with your own data, never with the demo
Vendor demos are built to look effortless and use data shaped to flatter the product. During a trial, import your real product catalogue and your real customer list, then enter a genuinely messy week of transactions: the partial payment, the short delivery, the return, the correction. This is the only reliable way to discover the friction you would live with daily, and it usually takes an afternoon.
- Import your actual catalogue and contacts
- Enter a real week including the awkward transactions
- Have the person who will use it daily do the testing
- Check the reports you actually need, not the dashboard
Check the exit before you sign up
Ask how you would leave. A vendor that makes export difficult has every incentive to keep it that way once your books are inside. Confirm that record lists export as CSV, that reports export as PDF, and that you can do it yourself without raising a support ticket. This is a five-minute check during a trial and it is the single best predictor of how you will be treated as a customer later.
- Confirm CSV export of products, contacts and transactions
- Confirm PDF export of the core financial reports
- Verify you can export without vendor assistance
- Read what happens to data after cancellation
The four mistakes that force a second migration
Almost every repeat migration traces to one of four errors. Buying for today's size and outgrowing the user limit or the feature set within a year. Ignoring inventory because you only hold a little stock, then discovering the spreadsheet has drifted from reality. Treating payroll as a separate problem and ending up with labour cost invisible from the accounts. And choosing on price alone, then paying for three add-ons that together exceed what a complete product would have cost.
- Buying for current size instead of next year's
- Deferring inventory until the spreadsheet breaks
- Leaving payroll outside the accounting system
- Optimising licence price and paying for add-ons instead
A decision sequence that works
Write the transaction list. Shortlist to three products that handle all of it without add-ons. Trial each with your own data for a week. Compare annual total cost at next year's headcount. Verify the export path. Then pick the cheapest of the remaining options, because at that point they all do the job and the difference is preference rather than capability.
- Write the transaction list first
- Shortlist on complete coverage without add-ons
- Trial with real data, one week each
- Compare annual totals, verify exports, then choose on price
FAQs
How do I choose the right accounting software for my business?
Start by listing every type of transaction your business actually performs, including partial payments, returns and adjustments. Shortlist products that handle all of them without paid add-ons, trial each with your own imported data, then compare total annual cost at next year's team size.
What features does small business accounting software need?
At minimum: double-entry bookkeeping, sales invoicing with receivable tracking, purchase recording with payables, categorised expenses, bank and cash accounts, and standard financial reports. Add inventory if you hold any stock at all, and payroll records as soon as you employ anyone.
Should I pick accounting software my accountant recommends?
Weigh it seriously but do not treat it as decisive. An accountant's preference genuinely reduces friction and sometimes their fee. It should not override a requirement they do not deal with, such as inventory or payroll, which you live with daily and they may never touch.
Is the cheapest accounting software a false economy?
Often, yes, because the cheap option usually excludes inventory and payroll, and those add-ons together frequently exceed the price of a complete product. Compare the total annual cost of everything you would need rather than the headline base price.
How long should I trial accounting software?
At least one full transaction cycle, which for most businesses means a month. A week is enough to judge the interface but not enough to surface month-end reporting and reconciliation problems, which is where products usually disappoint.