How to Run Old and New Retail Systems Together During Migration

When NCR announced that NCR Retail Online was being discontinued, retailers across Australia suddenly faced a decision they had not budgeted time for. Whether the destination is Magento or WooCommerce, the technical move is only one part of the picture. The weeks leading up to cutover are where most data errors, stock discrepancies and staff confusion arise. A controlled parallel run gives a business the breathing room to compare two live systems, isolate faults and protect revenue while staff adapt to the new environment.

A parallel run means operating the legacy platform and the replacement platform at the same time, processing the same transactions through both, then comparing the results. Done properly, it converts a high-risk leap into a measurable, auditable handover. Understanding the discontinuation timeline helps frame how much runway a retailer actually has before legacy access ends.

Planning the scope and length of the dual-run window

The length of a dual-run window should be set by transaction complexity and regulatory obligation. Australian Consumer Law guarantees, GST reporting under the ATO and the five-year record-keeping rule all favour longer observation windows rather than shorter ones. A Brisbane-based apparel retailer selling through both a physical store and an online channel needs to test GST-inclusive pricing, freight calculations and click-and-collect paths. By contrast, a Sydney grocer may need to test loyalty point redemption and daily stocktakes in real time. These differences dictate whether a two-week window is enough or whether a month is more appropriate.

Retailers should pick a representative slice of the business rather than the entire catalogue. Twenty percent of SKUs that generate roughly eighty percent of revenue is a practical starting point. This keeps the test measurable and avoids overwhelming warehouse staff who are simultaneously learning a new back office. Each SKU included should have a known price, a known supplier cost and a known shelf location in stores such as Melbourne's flagship sites or Adelaide warehouses. That way, variances found during reconciliation can be traced back to a clear source.

Communication also matters during this phase. A simple internal memo explaining what is being tested, who is responsible for reviewing the results and when the next status update will arrive prevents the project from drifting. Australian retail operators are accustomed to weekly huddles, and the same habit works well here. Standing meetings on Monday mornings to compare Friday's variance report keep the migration on track without consuming management bandwidth.

Mapping data flows and reconciling inventory between platforms

Inventory is the most sensitive data type during any platform transition. A single SKU counted twice, or worse, sold twice, creates customer disputes that are difficult to recover from under Australian Consumer Law. Before the parallel run begins, the data team should map every field that flows from the legacy system to the new platform: SKU codes, barcodes, supplier codes, bin locations, batch numbers for perishable lines and country-of-origin labels required by Australian import rules. Even seemingly minor fields often turn out to be critical during reconciliation.

Daily reconciliation becomes the heartbeat of the parallel run. At the close of business, both systems should be compared on three core metrics: total units on hand, total value of sales for the day and total number of transactions. Any variance above a pre-agreed threshold, say fifty dollars or three units, should be logged and investigated before the next trading day begins. Tools such as spreadsheets, simple SQL queries or dedicated reconciliation software can do the job, depending on the size of the catalogue. Many Australian retailers pair this work with their existing stocktake habits used before EOFY, treating the reconciliation as a discipline rather than a one-off exercise.

A useful bulk product upload reference explains how CSV imports handle large catalogues without losing field mappings, reducing the chance of discovering upload limits only after the parallel run has begun.

Configuring test transactions and verifying payment flows

Test transactions should mirror real customer behaviour as closely as possible. Australian shoppers frequently use buy-now-pay-later providers such as Afterpay and zipPay, alongside regular credit card payments, POLi internet banking and Apple Pay. Each of these payment methods must be exercised during the parallel run, including the failure path, such as a declined card or an abandoned Afterpay instalment. Verifying only the happy path leaves blind spots that surface at the worst possible time.

Refunds, partial refunds and store credits also need testing. The Notifiable Data Breaches scheme under the Privacy Act 1988 means that any customer data mishandled during a refund flow could create legal exposure if it leaked. Running a refund in both systems and confirming that the receipt, the email notification and the ledger entry match exactly is non-negotiable. Tax invoices must show GST correctly, with the ten percent component displayed separately as required by the ATO.

Ecommerce-specific flows deserve equal attention. Discount codes, free-shipping thresholds, regional rates for Western Australian customers and gift card balances all behave differently across platforms. The chosen destination theme also influences how these flows appear to customers. Reviewing WooCommerce theme guidance helps ensure the storefront being tested reflects what shoppers will actually see once the legacy system is retired.

Training staff and documenting operational differences

A parallel run is as much a people project as a technical one. Cashiers in Perth, floor staff in Hobart and warehouse pickers in regional Queensland will each encounter the new system in their own context. Training should happen before the dual-run period starts, so that the comparison data reflects competent usage rather than learning-curve noise. Short, focused sessions of forty-five minutes each, repeated across shifts, tend to work better than a single three-hour workshop for retail teams.

Documentation matters just as much as training. Each operational difference between the two systems should be recorded in a shared document that staff can consult during the parallel run. How do you process a manual refund in the new platform? Where is the lay-by module? How is a gift card activated for a customer who lost their original card? These answers must be written down, tested and confirmed before they are relied upon during peak trading.

Finally, the documentation should be reviewed by management every week during the parallel run. Any process that requires more than two extra clicks, or any report that takes noticeably longer to generate, is a candidate for redesign before the legacy system is retired. Australian retailers have learned the hard way that operational friction shows up first in staff turnover, then in customer reviews, and only later in the financial reports.

Setting the cutover moment and fallback triggers

The cutover decision should be evidence-based, not calendar-based. Define in advance the criteria that must be met before the legacy system is retired: zero unexplained inventory variances for five consecutive trading days, ninety-five percent of test transactions matching exactly, all payment providers confirmed and staff comfortable with the new workflows. When those conditions are met, the cutover can proceed with confidence. When they are not, the parallel run simply extends another week.

A fallback plan is also essential. If the new platform fails on cutover day, can sales continue on the legacy system for another forty-eight hours? If customer data becomes inaccessible, is there a manual ordering procedure that can be invoked? Documenting these fallback paths in advance prevents panic decisions during a real incident. Australian retail calendars, with Boxing Day sales in late December and EOFY promotions in June, are particularly unforgiving of unplanned downtime, so resilience matters as much as performance.

Once the cutover succeeds, keep both systems accessible for at least one more full trading cycle. This grace period allows any delayed transactions, end-of-day reports or after-the-fact stock adjustments to be captured correctly in both ledgers before the legacy system is archived or decommissioned. Treating this final week as insurance rather than overhead is the mindset that separates a clean migration from a costly one.

The most valuable habit during a system migration is treating reconciliation as a daily discipline rather than a one-time event. Australian businesses that document variances every morning, retest failed paths the next day and refuse to cut over until the data agrees across both platforms consistently arrive at the new system with intact customer trust, accurate stock records and a team that has already absorbed the change.