We have been analyzing the NCR Retail Online (NRO) business and our NCR Industry Solutions Board, an internal team that helps set strategy, has decided to set the NRO product to End of Life on March 31, 2018 . The CPOnline Product was also recently announced with an end of life date of September 30th, 2017 . The End of Life terms indicate that all current customers will need to be transitioned off their respective product and the servers turned off by 9/30/17 (CPO) & 3/31/18 (NRO) . Your NCR Counterpoint business partner has been notified of this decision in advance and has started taking steps to help you transition your eCommerce solution.

Next Steps

As of today, we are encouraging all customers to reach out to your current NCR Counterpoint Partner to begin the transition to a new eCommerce platform. Your partner will be your best resource in planning and transitioning to a new eCommerce solution.

NCR has worked with several partners to create options for your new eCommerce solution. Please refer to the below chart for information about these options. Your partner can provide you with further documentation about these solutions to assist you with the decision process. You can also view a list of FAQ’s about moving from NRO to one of the below options by clicking here .

We will be discussing this transition directly with the users that attend our Synergy User Conference at the end of June. We will be offering a presentation on eCommerce and we will have representatives at the exhibit booth to handle your questions. In the meantime, please reach out to your partner to help determine your next steps.

We appreciate your business and look forward to taking this next, innovative step together.

Recommended eCommerce Solutions

Solution Cost Platform Additional Notes
Commerce5
  • Upfront: Starts at $2500**
  • Monthly: Starts at $495.00 plus hosting
Magento Most tightly integrated with Counterpoint and offers the most advanced features
CP Magento
  • Upfront: Starts at $2,500**
  • Monthly: Starts at $200.00 including hosting
Magento Integrated with Counterpoint and offers features similar to NRO
CP Shop
  • Upfront: Starts at $999**
  • Monthly: Starts at $125.00 plus hosting
Woo Commerce Catalog, Inventory, and Orders are integrated with Counterpoint

Migrating purchase history for personalised recommendations

NCR Retail Online has been discontinued, so retailers that relied on it must move storefront, order, catalogue and customer records to a supported platform such as Magento or WooCommerce, usually with help from NCR Counterpoint partners. This is more than an export: the new system must preserve inventory synchronisation, trading history and the context behind each recommendation.

The work described in Migrating NCR’s Customer Purchase History for Personalized Recommendations turns old orders into useful behavioural signals without exposing sensitive customer information. It also prepares Australian retailers to recommend products accurately across different stores, currencies and fulfilment routes, even when the original NCR web presence no longer exists.

Set the migration scope

Begin by identifying which stores, markets and historical periods are in scope. Decide whether the new system must reproduce customer accounts, purchase history, wish lists and abandoned baskets, or only completed orders for recommendation purposes. This boundary prevents a broad data transfer from delaying the launch and makes completion easier to measure.

Ask NCR or the relevant Counterpoint partner for available export formats, API access and data-retention arrangements. Record the extraction date because stock, prices and promotions change continuously. Keep the original files read-only, encrypt them and create checksums so the team can verify that the historical dataset remained unchanged during transformation.

Define measurable outcomes before choosing algorithms. Useful targets include repeat-purchase rate, basket value, recommendation click-through, conversion and return rate. Separate commercial goals from migration goals: loading every record proves transfer, while a clean customer-to-product history proves that the new platform can act on it.

Normalise customer and order data

Combine customer, order, product and fulfilment exports into a documented model. Preserve stable identifiers where possible, including legacy customer ID, order number, SKU, variant, store, timestamp, quantity, paid amount, discount and return status. Do not use email address as the only join key, since households and shared accounts can otherwise collapse into one profile.

Clean values before importing them: standardise dates and time zones, convert amounts to integer minor units, identify cancelled orders and refunds, and remove obvious duplicates. Keep Australian records in AUD while retaining the original currency and tax fields. A purchase made in a Melbourne store and one placed online through a Sydney warehouse may share a SKU but differ in availability and fulfilment context.

Use old catalogue records to test how categories and variants were represented. A sample such as the cookware catalogue page can reveal whether product names, sizes and ranges were entered consistently, which directly affects attribute-based recommendations. Treat the page as a reference, not as a substitute for the retailer’s authoritative SKU export.

Preserve catalogue meaning

Historical order lines become useful only when product identity survives the move. Create a crosswalk from every legacy product ID to the new catalogue ID, while retaining the original description, category, brand and attributes. Flag products that were deleted, merged or replaced, and decide whether discontinued items should influence future recommendations or remain only in reporting.

Separate genuine preference signals from operational noise. A large appliance bought once for a renovation is weaker evidence than repeated purchases of compatible kitchen products. Apply time decay, account for promotions and returns, and retain the distinction between a stocked item and an unavailable one. Test imports should also use a secondary reference dataset in a sandbox after checking its rights and quality; malformed titles, missing images and inconsistent categories can expose weak mappings without touching production records.

Model events in a way that can support basket analysis, item-to-item affinity and next-best-product logic. Relevant events include product views, searches, carts, purchases, cancellations and post-purchase returns. Where the old platform did not capture an event clearly, do not manufacture certainty; label the gap and use only the strongest available signals.

Select the target platform

Magento is suited to retailers needing granular catalogue control, complex pricing and a broad extension ecosystem. WooCommerce often offers a lighter transition for smaller operations because core ecommerce functions sit within WordPress and specialist plugins add recommendation features. Both can support APIs, automation and integrations with inventory or business-management systems.

The main differences can be summarised across the decisions that affect a migration:

Consideration NCR Retail Online legacy Magento WooCommerce
Availability Discontinued; archive and export transition required Active commerce platform Active commerce platform
Core model Retail inventory and business-system integration Flexible commerce stack with extensions WordPress-based stack with plugins
Recommendation approach Primarily rules and available legacy data Extensions, APIs and custom ranking Plugins, APIs and custom ranking
Integration NCR and Counterpoint transition support Broad ERP, PIM and extension ecosystem Plugin and API-based integrations
Best fit Historical data source during migration Complex catalogues and deeper customisation Faster, lighter implementations

Cost is not the only deciding factor. Compare hosting, plugin or extension licences, security maintenance, data-modelling effort and the skills required to operate the platform. A simple rule-based recommendation may be easier for a small Australian retailer to maintain than an elaborate model, while a catalogue with thousands of variants may justify deeper platform functionality.

Create recommendation-ready profiles

Translate transactions into customer-level features: preferred categories, brands, price bands, purchase frequency, average basket size, recency, seasonal patterns and complementary items. A shopper who repeatedly buys coffee accessories may respond to related grinders and filters, while another customer who buys giftware once may need a different strategy. Make each feature explainable so merchandisers can distinguish useful patterns from accidental correlations.

Start with transparent methods. Product affinity and frequently bought-together rules are often sufficient for an initial launch; collaborative filtering and ranking models can follow once the data is dependable. For example, category and accessory relationships visible in barware product lines can support sensible cross-sell rules, while stock status and margin controls determine what is presented.

Test recommendations by segment, channel and lifecycle stage. Compare them with popular-product and no-recommendation baselines, and review click-through, conversion, average order value, margin and return rate. Use historical order replay to catch broken SKU joins, then run controlled live trials before changing all product pages. Relevance matters more than the number of recommendations shown.

Protect privacy and local operations

Where the Privacy Act 1988 applies, handle customer data in line with the Australian Privacy Principles. Collect only what the recommendation purpose requires, secure it against unauthorised access, and explain how purchase data is used. Profiling that reveals or implies sensitive information needs particular care. Marketing messages must also meet the Spam Act 2003, and customers should have a practical way to opt out.

Apply role-based access, encryption in transit and at rest, audit logs, retention limits and tested recovery procedures. Separate identifiers from behavioural features where possible, and restrict exports of email addresses and order histories. Security should cover the recommendation feed itself: manipulation, stale recommendations and leaked personalisation can expose data even when the underlying order database remains protected.

Local fulfilment affects both relevance and trust. Major customer concentrations around Sydney and Melbourne, long routes to regional buyers and separate stock pools can make a frequently purchased item unavailable in a particular store. Synchronise inventory in near real time, show availability by location, and suppress recommendations when fulfilment cannot be promised. Record GST-inclusive prices and promotions consistently so shoppers do not see an offer that changes at checkout.

Cut over and remember the essentials

Run a staged cutover with reconciliation checkpoints. Compare customer, order, product and permission totals between source and target; test sample accounts through search, checkout, refund and recommendation paths; and retain a recoverable copy of the source export. Run old and new recommendation logic in parallel where feasible, then switch traffic by store or customer cohort rather than all at once.

Assign ownership for data quality, platform configuration, privacy, merchandising and stock accuracy. A weekly review should catch duplicate profiles, broken product mappings, declining click-through and unexpected return rates. Recalibrate around Australian trading peaks such as EOFY sales, Black Friday and Christmas, when promotion-driven baskets can otherwise be mistaken for lasting preferences.

The essential takeaway is trustworthy context, not the archive itself: a complete, lawful and well-identified dataset stays useful only when it remains aligned with live inventory and genuine customer choices.

After you have completed your move to a new eCommerce platform, don’t forget to submit the Store Closure Request form to close your NRO site and cancel your billing subscription.