When spreadsheets actually work.
A single store with under 200 SKUs, one cashier per shift, fewer than 50 transactions a day, and an owner who does the books personally — that store can run on Excel forever. The data volume is manageable, the transaction velocity is low, and the owner has full context.
The spreadsheet works because the owner is the redundant memory. They remember which supplier sent the bad batch, which customer pays late, which SKU is actually the low-stock one even though the sheet says otherwise. Their brain is the second source of truth.
The three thresholds that change the math.
Threshold one: more than one cashier. The moment a second person is taking money, the owner is no longer the witness. Reconciliation across people requires structure that Excel does not enforce. You need cashier IDs on every transaction, automatic shift reports, and an audit trail.
Threshold two: more than one branch. Inventory across locations cannot be tracked in a single sheet without enormous discipline. Branch-level stock requires branch-level pricing, branch-level reports, and inter-branch transfers. Threshold three: more than CA$20,000 per month in revenue, where mistakes start costing real money.
- More than one cashier in any shift
- More than one physical branch
- More than CA$20,000 per month in revenue
- More than 1,000 SKUs
- Customer accounts and credit terms (any AR balance)
What a POS gives you that Excel cannot.
A POS gives you transaction-level integrity by design. Every sale ties to a customer (if any), a cashier, a branch, a tax rate, and an inventory adjustment. Excel can model that, but humans typing into Excel will not maintain it. The POS enforces what humans cannot.
It also gives you the ledger underneath. Every sale becomes a journal entry: DR Cash, CR Sales, CR Sales Tax Payable, DR COGS, CR Inventory. In Nonari that double-entry is automatic, and the trial balance is always current. No month-end migration of sales summaries into your accounting software.
The migration that is not as scary as you think.
Most owners delay the move because they imagine three months of pain. The actual migration for a 1,000-SKU shop is two weekends. Export the master list, format the import template, dry-run with last week's sales, then go live on a Monday morning when traffic is light.
The hard part is not the data. It is the muscle memory. Cashiers who have done the same routine for three years will resist the new flow for two weeks. Push through the first two weeks and you will not go back.
What to demand from your first POS.
A first POS for a small retailer needs to do five things well: ring sales fast, track inventory per branch, support multiple cashiers, print receipts customers can read, and produce a usable end-of-day report. Everything else is a bonus.
Skip the systems that try to be everything. Skip the systems that require a server and an IT consultant. Pick a cloud POS with offline support, monthly billing, and a clear export path — so if you outgrow it, your data comes with you.
The compliance forcing function.
In many regions e-invoicing mandates are pushing retailers off spreadsheets whether they like it or not. The CRA in Canada is increasing GST/HST digital filing requirements. Australia's ATO requires Single Touch Payroll for retailers with staff. India's GST e-invoicing applies above a turnover threshold. Each of these is a forcing function.
Below the mandate threshold there is no compliance pressure, but the practical case still applies. SMB retailers everywhere are graduating to cloud POS at scale, because the price point of decent SaaS POS (CA$40-150 per month) is now lower than half a day of weekly reconciliation.
A two-week pilot beats a six-month decision.
Do not spend three months evaluating POS systems. Pick two finalists from a one-day shortlist, run a two-week parallel pilot at one branch (POS plus your existing spreadsheet, both updated), and switch fully to whichever has fewer surprises.
Nonari runs a free starter that lets you do exactly this: bring up one branch, import your current SKU list, and run a real week of transactions. No credit card. The decision becomes empirical instead of theoretical, and you stop second-guessing the call.