Choosing a stock system for a family hardware shop

Three-year cost, cloud: €95 × 36 = €3,420, plus any card-reader or fiscal-printer integration and the cost of a backup internet connection. Three-year cost, custom: €4,500 plus about €225 a year for three hours of changes = €5,175, plus any office PC upgrade Access needs and the risk of the developer becoming unavailable. On money alone the cloud option fits the €3,000 first-year budget (€1,140 in year one), whereas the custom quote exceeds it before a single change request. But cost is not the only test. The deciding constraint is the unreliable broadband: a till that cannot issue a receipt when the line drops is worse than the present paper system, so the recommendation must be conditional. Choose the cloud package only if it can queue sales offline and synchronise later, and only if its receipts satisfy Maltese fiscal requirements. If the supplier cannot demonstrate offline operation under test, the custom database on the office PC is the safer choice despite costing more. A strong answer also notes that neither option fixes the stocktake by itself: the shop needs a counted opening stock figure and a disciplined goods-in process first.

Non-functional requirements describe how well the system must work, not what it does, and they are where small suppliers are weakest. First, availability and offline operation: the till must keep selling and printing receipts while the internet is down, then synchronise without duplicating invoices. Test it by unplugging the router for a two-hour pilot, issuing at least twenty sales, reconnecting, and checking that stock levels and invoice numbers reconcile exactly. Ask what happens if two tills sell the last item offline. Second, usability and trainability: a part-time assistant with basic Excel skills should be able to complete a sale, a stock receipt and a stocktake count with no more than a one-page prompt. Test it by having that assistant, not the supplier, run ten transactions unaided after one hour of training, and time the process. Record error rates. A third defensible requirement is data portability: a full export of products, customers and invoices in a common format, tested before signing, so the shop is not trapped if it leaves. Testable wording matters - 'fast' and 'easy' cannot be checked, but 'offline for two hours' and 'ten transactions in fifteen minutes' can.

Migration should start with cleaning, not with exporting. First take a physical stocktake and reconcile it to the paper cards, because moving bad figures into a new system only hides them. Deduplicate product lines (the same screw sold under three descriptions will distort reorder levels), standardise units and VAT codes, and decide which customer records are worth carrying: for a cash shop, most sales are anonymous, so only account customers and supplier details need importing. Map each source field to a target field and write the mapping down, then load into a test copy and check a sample of at least fifty lines by hand, including prices and reorder levels. Keep the old workbook read-only for the retention period. For staff, resistance is usually about competence and status rather than the software. Involve the older assistant in the pilot and ask her which screens are unclear; give short sessions on one task at a time rather than a half-day lecture; nominate her as the person others ask, which converts a critic into an owner; and run paper and system in parallel for two weeks while she checks totals. Emphasise the benefit she feels - fewer stock-outs and a one-day stocktake - rather than features.