Why clean data matters before moving to a new platform

Moving from NCR Retail Online to a replacement ecommerce system is more than a technical export and import. Product records, customer accounts, stock levels, prices, images, tax settings and order history all carry assumptions from the old environment. If those assumptions are copied without review, the new store can reproduce years of small errors at a much larger scale.

A data cleanup plan gives the migration a controlled sequence. It identifies what should be kept, corrected, combined, archived or discarded before records reach Magento, WooCommerce or another platform supported by an NCR Counterpoint partner. The result is a cleaner catalogue, fewer customer service problems and a more trustworthy connection between online sales and the shop floor.

Australian retailers have specific details to protect. GST-inclusive pricing, ABNs, Australian postcodes, state-based delivery rules, AUD currency and stock held across metropolitan and regional locations can all affect how data behaves after import. A rushed migration can turn a harmless spreadsheet issue into an incorrect price at checkout or an order sent to the wrong side of the country.

Start with a complete data inventory

Before editing anything, list every source that may contain retail information. This usually includes the ecommerce database, Counterpoint, vendor spreadsheets, point-of-sale exports, email marketing tools, accounting software, image folders and manually maintained CSV files. The inventory should record the owner, format, last update date and business purpose of each source.

This step often reveals duplicate systems. A product may have one description in the online shop, another in Counterpoint and a third in a supplier catalogue. Customer details may also exist in several versions, with different spellings, old addresses or multiple records for the same household. An inventory makes those conflicts visible before a migration team treats one version as automatically correct.

Retailers selling wine, specialty foods or other regulated products should give extra attention to attributes such as vintage, vintage year, bottle size, pack quantity, allergens and age-related sales rules. Product structures discussed in resources about food and wine retail can help highlight fields that a generic product export may overlook.

Define what clean data means

Data cleansing is not simply deleting blank cells. It means setting practical rules for accuracy and consistency. For example, a retailer might decide that every product needs a unique SKU, a clear title, a GST treatment, a brand, a category, a selling unit, an image and a current stock status before it can be published.

Create a data dictionary that explains each field in plain language. “Weight” should specify whether the value is grams or kilograms. “Cost” should state whether it includes GST. “Available stock” should clarify whether it means physical stock, sellable stock after safety buffers or stock in one particular store. These definitions prevent different teams from correcting the same file in contradictory ways.

Australian addresses deserve their own validation rules. Keep the leading zero in postcodes, separate suburb from state, and standardise state abbreviations such as NSW, VIC and QLD. Confirm whether a customer’s delivery address is a residential, business, PO Box or remote-area address, since carriers may apply different service conditions and fees.

Remove duplication and outdated records

Duplicate products are a common source of overselling. The same item may appear under a supplier SKU, an internal SKU and a barcode, while variations such as colour, size or flavour may be recorded as unrelated products. Establish a master identifier and decide which records are parent products, variants, bundles or discontinued items.

Customer deduplication requires more care because two people can share an address, while one person can use several email addresses. Match records using a combination of email, phone, name and address rather than a single field. Preserve useful order history and loyalty information, but avoid importing obsolete marketing permissions or contacts who should no longer receive promotional messages.

Old products should be archived rather than silently erased when they have reporting or warranty value. However, they should not remain accidentally visible for sale. Add clear statuses such as active, draft, discontinued and archived, then map those statuses to the new platform’s publication and inventory rules.

Map fields before importing records

Field mapping is where business meaning is translated into the new platform. A source field called “department” might become a category, collection or product tag. A “web status” field may control visibility, while “available quantity” may need a different treatment from the physical quantity recorded in the shop.

Vendor data needs special handling because supplier files are rarely consistent. One spreadsheet may use millilitres, another litres; one may include GST in cost and another may not. Guidance on syncing vendor catalogues is relevant when product updates must move between Counterpoint and a non-NCR ecommerce platform. Importing a supplier file without normalisation can overwrite carefully maintained titles, prices or stock rules.

Validate imports in a safe environment

Never make the live store the first place where migrated data is tested. Use a staging site or a restricted import area and load a representative sample. Include simple products, variants, bundles, sale items, products with missing images, customers with multiple addresses and orders with refunds or shipping charges.

Check the customer experience as well as the database. Search for products by SKU and barcode, open category pages, add variants to a cart and test GST calculations. Compare displayed prices with approved source files. For Australian stores, test delivery to a major city such as Melbourne and to a regional location such as Cairns, including postcode validation and shipping-rate selection.

Reconcile totals after each test import. Count products, active variants, customer accounts, open orders and inventory units before and after the transfer. Investigate differences instead of accepting them as normal migration noise. A small mismatch in a test environment is much easier to explain than hundreds of missing records after launch.

Control the cutover and future quality

A clean import can deteriorate quickly if new records continue to enter the old system without rules. Set a cutover date, freeze or limit changes during the final export, and assign owners for products, customers, inventory and orders. Keep a dated backup of every source file and an import log showing what was loaded, rejected or amended.

After launch, monitor practical indicators: overselling, failed product searches, duplicate accounts, incorrect tax values, broken images, abandoned carts and support requests about delivery. Schedule a review after the first week, then again after the first full trading month. Seasonal Australian retail periods, such as Christmas, Boxing Day sales and end-of-financial-year promotions, can expose issues that a quiet test week misses.

The strongest migration process treats cleanup as an operating discipline rather than a one-off spreadsheet exercise. Keep naming conventions, mandatory fields and approval rules documented, and make suppliers follow a usable template. Before the final import, preserve the original exports, approve a reconciled sample and sign off each critical field. The practical takeaway is simple: correct the source, define the meaning, test the result and only then make the new platform live.