How to Choose a POS for a Grocery Store

01 / Checkout and barcode speed

Follow a real basket through checkout.

A grocery-store POS should handle packaged barcodes, weighed items, supplier deliveries, promotions and stock movements without losing the item, unit or location. Start with the jobs your cashiers repeat, then use the general POS selection guide for wider cost and reporting questions.

Demo test: scan a mixed basket on the proposed scanner and till. Include repeated items, a damaged label, an unknown code, a void, a return and a restricted staff action. Check the receipt and stock movements. Observe queue handling with your own basket; a vendor speed claim is not a measured result for your shop.

02 / Weighed items, PLU and scales

Keep the selling unit and stock unit consistent.

Loose rice, produce and packaged goods need different workflows. Ask whether staff enter a PLU (product lookup code), use an integrated scale, or scan a label printed by a separate scale. Check supported units, tare, decimal precision, variable-weight label formats and rounding.

Demo test: sell 0.750 kg from a fictional 10 kg rice balance at 4 currency units per kg. Expect a line price of 3 before tax and stock of 9.250 kg. Return or void it and reconcile the quantity. Confirm compatibility and any local trade-scale approval with the supplier; a USB connection alone proves neither.

03 / Supplier purchasing and receiving

Receive what arrived, preserve what is outstanding.

Demo test: order 24 packaged units and 10 kg of rice. Receive only 18 units and 8 kg. Inspect the remaining order, delivery reference, supplier, destination and staff history. Receive a case and verify its conversion into selling units. Try a damaged delivery, wrong item and supplier return. A purchase order is not stock until your chosen receiving process records it.

04 / Cost changes and margin

Find the margin effect before changing a price.

Demo test: change the next supplier cost from 2 to 2.40 while the fictional selling price stays 3, before tax and other costs. Unit gross profit falls from 1 to 0.60; gross margin falls from about 33.3% to 20%. Ask which inventory costing method the system uses and whether it preserves past transactions rather than rewriting history.

Review freight, tax treatment and case conversions with your adviser. Use the Profit & Stock Calculator with verified inputs to explore margin and stock performance; it does not reconcile a POS ledger for you.

05 / Promotions and pricing

Test the rule, its limits and its end date.

Demo test: set a two-for offer, then try one, two and three eligible items, an excluded item, a return and an expired offer. Check whether loyalty discounts stack, who can override the price and how margin is reported. Compare the shelf label, promotion sign and receipt. Test mixed tax categories in the configuration your local adviser confirms.

06 / Counts, shrink, damage and waste

Give each stock correction a reason.

Demo test: count a shelf while another till sells an item. Ask how the count cutoff prevents double adjustments. Record one damaged unit separately from an unexplained shortage and an expired unit. Check quantity, cost impact, permissions, reason, timestamp and audit history. Confirm waste is removed from availability rather than recorded as a sale.

The dead stock guide helps distinguish slow sellers from stock you should stop buying. A slow-selling item and an item approaching expiry need different decisions.

07 / Reordering and stockouts

Use demand, lead time and usable stock together.

Demo test: reduce a staple below its reorder threshold and inspect the suggested order. Include supplier lead time, minimum order, case size, open orders and stock already reserved. Ask how the calculation treats stockouts, promotions, waste and short shelf life. Replenishing an expiring surplus can make the problem worse.

Use the Reorder Point Calculator to explore timing with your own demand and lead-time inputs. It does not calculate shelf life or automatically place supplier orders.

08 / Multiple tills and locations

Separate stock on hand from stock available here.

Demo test: sell from two tills, inspect the shared balance, then dispatch two units from Store A and receive them at Store B. Total stock should reconcile through the transfer; in-transit stock should remain distinguishable. Try a partial receipt, return at another branch and a lost connection. Check which actions work offline and how duplicates or conflicting updates are resolved afterward. Do not assume an offline guarantee.

09 / Country, payments, tax and hardware

Verify the exact setup in your market.

Software availability does not establish payment or hardware availability. Get written confirmation for your country, business, plan and device combination.

• Software: account eligibility, limits, exports, integrations and ongoing costs.
• Payments: processor eligibility, local payment methods, settlement, refunds and current fees.
• Hardware: tills, scanners, scales, receipt printers, terminals and local replacements.
• Support: training, trading-hour coverage, language and escalation.
• Local and tax fit: receipts, invoices, tax reporting and fiscal requirements confirmed with the provider and a local adviser.

Demo test: use the intended devices for a sale, refund and outage/recovery exercise. This guide does not verify product prices, country support, payment support, tax compliance or offline operation.

10 / GS1 2D / Ambition 2027 readiness

Test decoding and processing separately.

GS1’s Ambition 2027 is an industry goal for retail POS to read and process a defined set of GS1-compliant 2D barcodes alongside linear codes by the end of 2027. It is not a universal legal deadline or a promise that every store will transition together.

GS1’s implementation guidance distinguishes product identification from additional uses of batch, expiry and weight data. A scanner beep does not prove that the POS stores those fields or acts on them.

Demo test: ask the provider for valid test labels for supported GS1 DataMatrix and QR Code with GS1 Digital Link syntax, plus a traditional 1D label. Scan a dual-labelled item once and confirm one sale, the correct GTIN/product, and the expected data passed into the POS. Inspect the saved transaction and export. Record unsupported formats, firmware/software upgrades, costs and delivery commitments. An ordinary website QR code is not this test.

11 / Expiry, batch and recall-aware inventory

Know which units need attention first.

In a fictional milk example, 12 units expire on 18 October and 6 on 22 October. A total of 18 hides the earlier batch. Ask the provider to show both lots, locations and dates, then trace a receipt, sale and waste adjustment back to the correct lot.

GS1 describes potential expiry-driven sale prevention, markdowns and targeted recall workflows in its 2D guidance. These need configured data and business processes; barcode compatibility alone is insufficient.

Demo test: identify the 12 earlier units, review first-expiring-first-out rotation and a proposed markdown, and try selling an expired test item. Mark one fictional lot recalled, locate its remaining units across locations and verify the block or staff response. Confirm how recall notices enter the system, who approves actions, and what happens when a label has no lot or date. Keep best-before and use-by/expiry fields distinct and apply your local policy. This is a software evaluation exercise, not a food-safety compliance certification.

12 / Connected shelf / electronic shelf labels

Keep the shelf, offer and checkout in agreement.

FMI describes electronic shelf labels (ESLs) as part of connected store operations, including price consistency and markdown workflows. Treat ESLs as a scale or future requirement when appropriate; a small grocery can use a controlled paper-label process.

Demo test: change one price in the owning product record. Check the shelf label, promotion and checkout, including scheduled start/end times. Disconnect a label or gateway and see how staff detect a stale price, retry the update and reconcile it. Confirm connector compatibility, approval permissions, batteries, maintenance and total cost. An update request is not proof the physical label changed.

13 / Explainable AI forecasting

Ask what decision the forecast changes.

FMI’s supply-chain discussion stresses that forecast accuracy must translate into workable business decisions. Our buying test is to demand an explanation and an owner-controlled action, rather than treating an AI badge as a feature result.

Demo test: request a reorder suggestion with its horizon, data freshness, sales history, seasonality, promotions, stock availability and lead-time inputs. Change a promotion or supplier delay and ask why the quantity changes. Inspect uncertainty, shelf-life limits, case rounding and minimum orders. Compare with a simple baseline using your own history, including stockouts. Require an owner override and audit trail; establish whether suggestions need approval before an order is sent. This guide reports no measured forecasting accuracy or savings.

14 / API and integration readiness

Follow one inventory change beyond the till.

Demo test: request current documentation for ecommerce, delivery, loyalty, accounting and inventory connections. Identify which system owns product IDs, price and stock. Inspect authentication, access scope, rate limits, webhooks, versioning, export fields, plan costs and who supports the connector.

Sell the last unit and watch the receiving system’s available stock. Interrupt the connection, replay an event and cancel an order; check duplicate handling, reservations, failure alerts and reconciliation. An API listing does not prove a working integration. A CSV export can support a manual workflow but is not a live sync guarantee. Reliable product, price and availability data also underpins future shopping channels.

15 / Smart checkout / computer vision — optional

Check exceptions before buying automation.

Traditional checkout, self-checkout and mobile scan-and-go have different staffing and exception needs. As a documented example, Instacart describes Caper carts using computer vision and sensors. That vendor description does not establish accuracy, availability or compatibility with your POS.

Optional demo: add and remove an item, try loose produce and an unrecognized product, then test a price dispute, refund, restricted item and interrupted connection. Review staff intervention, payment reconciliation, accessibility, privacy and support with the provider. Small groceries do not need smart carts to choose a good POS. Larger stores should establish integration feasibility and operating cost before a pilot.

16 / AI shopping / agentic commerce — emerging

Make the basket accountable to the customer.

Instacart describes Cart Assistant across online and in-store channels. Conversational shopping is a documented vendor example, not evidence that every grocery POS supports autonomous purchasing.

Future-channel test: ask a connected assistant to build a dinner basket within a fictional budget. Check product identity, pack size, current price, available stock, substitutions and final totals against the source system. Require customer confirmation before purchase and test cancellation, out-of-stock responses, restricted products and a failed payment. Confirm consent and access boundaries. API/data reliability matters more than an AI label; do not grant purchasing authority just to demonstrate a recommendation.

Original ShopsInto evaluation method

Run the 12-step core Grocery POS Test.

This fictional sandbox scenario is a repeatable evaluation method, not a vendor score or a report of hands-on testing. Use the same catalogue and fixtures in every demo. Currency values are illustrative, before tax unless stated. Keep the first packaged item, second 2D-labelled item, rice and milk as separate stock records.

1. Scan a standard 1D barcode. Start with 10 packaged units. Sell one and confirm the correct product, receipt and balance of 9.
2. Scan a GS1-compatible 2D barcode. Use provider-supplied valid test labels for a second item with 5 units. Sell one; expect 4. Inspect GTIN, lot and expiry capture separately, and confirm a dual-labelled pack is not charged twice.
3. Sell a weighed / PLU item. Start with 10 kg of rice. Sell 0.750 kg at 4 per kg; expect 3 before tax and 9.250 kg remaining. Check tare and rounding.
4. Receive a partial supplier order. For the first packaged item, order 12 units but receive 8. Expect 17 on hand and 4 outstanding, with the receipt linked to the order.
5. Process a supplier cost change. Record 2.40 instead of 2 on the new receipt. At a selling price of 3, inspect the 20% gross margin for that cost basis and confirm the documented costing method. Quantities stay unchanged.
6. Record damaged or wasted stock. Remove one damaged unit of the first item, leaving 16. Inspect its reason, cost and staff audit trail.
7. Identify a short-dated item. Load separate milk lots of 12 units expiring on 18 October and 6 on 22 October, using a demo clock before those dates. Find the earlier lot, test an expiry block and a fictional recall in a sandbox, and restore the fixture afterward.
8. Apply and verify a promotion. Preview a two-for offer on the first item without completing another sale. Check eligibility, receipt calculation, return behavior and expiry of the offer; stock remains 16.
9. Trigger a low-stock decision. Set a fictional threshold of 18 for the first item. At 16 on hand, inspect the alert and how 4 outstanding units, reservations and lead time affect any order suggestion.
10. Verify shelf price equals checkout price. Change a regular price, update the paper or electronic label and compare the actual shelf with the till and promotion. Preview without completing a sale.
11. Check stock across tills / locations. Observe the shared balance on both tills, then transfer 2 units of the first item from Store A to Store B. After receipt expect A: 14, B: 2, total: 16.
12. Export or connect inventory data. Reconcile the first item at 16 total, second at 4, rice at 9.250 kg and milk at 18 across its lots. Export identifiers, units, locations, lots/dates where supported, and available versus reserved stock. Never add kilograms to unit counts.

For each step record observed behavior, workaround, required plan/integration, evidence and unresolved questions. Keep vendor statements separate from demo observations. A missing essential workflow needs resolution; do not average it away with a score. For date tests, use a controlled demo clock or equivalent future dates.

Add 4 optional advanced tests when they fit.

These extend the core test. They are not entry requirements for every small grocery.

1. AI reorder suggestion. Explain inputs, uncertainty and shelf-life constraints; override the quantity and inspect the approval record before any order is sent.
2. Electronic shelf label update. Observe the physical label after a scheduled price change; simulate a failed update and recover it.
3. Ecommerce inventory sync. Sell the last sandbox unit, attempt an online order, then cancel and reconcile reservations after an interrupted connection.
4. Smart checkout / external integration. Add, remove and misidentify a test item or replay an integration event; verify staff correction and a single reconciled transaction.

Choose the workflow before the brand.

For a small grocery, prioritize checkout, weighed goods, receiving, accurate pricing and usable stock. For perishables, resolve expiry and lot handling early. Add connected shelves, AI and smart checkout when the scale and operating case justify them. Write down essential jobs, acceptable workarounds, costs and market confirmations.

Use POS Finder to explore options, then bring this checklist to provider demos. Visit Comparisons for product research and its evidence status. This guide does not name a universal best grocery POS.