Skip to main content
Data & AI
Blog

How Retail and CPG Enterprises in Malaysia Can Fix Fragmented Customer and Commerce Data

 

The campaign is over, but the teams reviewing it still can’t agree on the result. Shopee has the order data. The direct-to-consumer store has the customer records. Retail media has the clicks. Distributors have the shipment figures, while store systems hold the transactions. Each source tells the truth about its own part of the sale. None can explain the whole path to purchase.

In 2026, Malaysian retail and consumer packaged goods (CPG) enterprises need an enterprise data platform that connects customer, product, commerce, media and supply data. Shared definitions and governance hold it together. ADA, The Data & AI Experience Company, builds this layer so enterprises can work from consistent customer, product and commercial views across channels.

The outcome should be more useful than another dashboard. Decision-grade intelligence needs to show what sold, why it sold, who bought it and what the business should do next.

TL;DR: Fix Fragmented Data

  • Malaysia’s commerce landscape is fragmented by design. Marketplaces, social commerce, owned stores, physical retail, distributors and media platforms each hold only part of the picture.
  • Their blind spots differ. Retailers often lose sight of a known customer between channels. CPG companies may have no direct view of the end customer.
  • A data foundation must connect more than customer records. Product, inventory, order, promotion, media and sell-through data need the same treatment.
  • Three connected views matter most. Enterprises need a unified customer view, a unified product and inventory view, and a unified commercial performance view.
  • Governance has to move with the data. Malaysia’s amended Personal Data Protection Act raises expectations for security, accountability, breach response and data portability.
  • The first release should improve one commercial decision. Trying to connect every source before proving value turns a growth initiative into an open-ended infrastructure programme.

Malaysia’s Retail Growth Is Producing More Signals, Not More Clarity

Malaysia generated RM1.288 trillion in e-commerce transaction income during 2024. According to the Department of Statistics Malaysia, that represented annual growth of 8.8%. Domestic transactions contributed RM1.150 trillion.

Behind those numbers sits an increasingly fragmented route to purchase. Ipsos reported that nearly two-thirds of Malaysians shopped online in 2025. Seven in ten had used Shopee during the previous six months. TikTok Shop usage rose from 35% of online shoppers in 2023 to 52% in 2025.

A shopper might notice a product in social content, compare it on a marketplace and inspect it in a store. The purchase may eventually happen through an app. The shopper experiences one connected journey. Inside the enterprise, the same person may appear as four unrelated records.

That disconnect causes four familiar problems:

  • Marketing stops at the platform report. Teams optimize clicks or marketplace revenue without seeing whether the campaign created incremental demand.
  • Commerce teams compare different definitions. Marketplaces, retailers and owned channels calculate traffic, conversion, promotions and attribution differently.
  • Supply teams receive the warning too late. Demand and inventory signals often arrive only after a stockout, overstock position or missed promotional window.
  • Leadership sees several versions of the same period. Finance, sales, media, e-commerce and distributors may all report a different commercial result.

An enterprise data foundation sits above these source systems and creates a governed common layer. The enterprise does not need to replace every application. It needs to connect the records, standardise their meaning, preserve their origin and make the data usable for reporting, activation, analytics and AI.

Retailers See Transactions but Lose the Customer Between Channels

Retailers in Malaysia may own the point-of-sale transaction, loyalty membership, app session and e-commerce order. That still does not guarantee that their systems recognise the same person.

One customer can use an email address online, a mobile number for loyalty and a marketplace profile for delivery. An anonymous payment token may record the physical-store purchase. Unless those identities are resolved, each channel treats a returning customer as someone new.

The error then flows into commercial decisions. A website buyer labelled as new may already be a valuable store customer. A marketplace promotion may appear incremental even if it simply moved an existing buyer away from a higher-margin owned channel. A loyalty offer may reach someone who bought the same product yesterday elsewhere.

A retail data platform should link these records only when consent, available identifiers and business rules allow it. The result is a persistent profile that changes as fresh signals arrive. Within a wider governed data foundation, ADA’s Customer Data Platform capabilities support identity resolution, profile unification, segmentation and activation.

That customer view can improve:

  • Cross-channel frequency and recency analysis
  • Customer lifetime value across online and offline purchases
  • Next-best-product recommendations using current behaviour
  • Loyalty segments based on total value rather than one channel
  • Suppression of duplicated or irrelevant promotions
  • Measurement of movement between marketplaces and owned channels

A unified customer view does not give an enterprise permission to use every signal it can collect. Consent, purpose, access and retention rules must remain attached to each record throughout identity resolution and activation.

CPG Enterprises Have a Harder Problem: The End Customer Is Often Invisible

CPG companies usually know what they sold to a distributor, retailer or marketplace operation. They have far less visibility into the consumer purchase, the influence behind it or whether the product moved at full price.

This is the difference between sell-in and sell-through. Sell-in records product shipped into a channel. Sell-through records product purchased by the consumer. A strong sell-in figure can conceal weak demand if stock is accumulating further down the chain.

The missing evidence is spread across distributor files, retailer reports, marketplace seller centres, retail-media platforms, promotion calendars and field-sales systems. SKU names and pack sizes may not match. Reporting periods can differ. One retailer may provide daily data, while another sends a weekly spreadsheet.

A CPG leader may not need an identified customer profile as the first result. A dependable view of product movement by SKU, channel, location, promotion and period can resolve more immediate decisions.

It can show:

  • Whether a media campaign increased sell-through or only marketplace traffic
  • Which promotions created incremental volume after discounts and media costs
  • Where distributor inventory is moving more slowly than forecast demand
  • How pack sizes perform across marketplaces, modern trade and regional stores
  • Whether stockouts are suppressing sales in locations with strong demand
  • Which retailer or marketplace contributes most after fees, discounts and advertising

ADA’s consumer goods solutions connect trade, media and operations to improve visibility into channel profitability. For CPG businesses, this commercial view bridges the gap between aggregate shipment reporting and better consumer-demand decisions.

The Three Views a Retail and CPG Data Foundation Must Create

An enterprise data platform for Malaysian retail and CPG should create three connected views. Building only one leaves an important part of the commercial picture unresolved.

A Unified Customer View Connects Identity to Permitted Action

A unified customer view links known and anonymous activity using transparent identity rules. It should reveal what the enterprise knows, where the information came from, the customer’s current consent status and which teams can act on it.

Retailers gain continuity between digital and store behaviour. CPG companies may instead combine consented direct interactions, campaign responses, product registrations, messaging activity and retail-partner audiences rather than pursue a complete transaction-level identity.

A Unified Product and Inventory View Creates One Commercial Language

A unified product and inventory view gives the business consistent identifiers for SKUs, packs, categories, locations, prices and promotions. It can reconcile “500 ml,” “0.5 L” and a marketplace listing code as the same product.

Demand models become unreliable when product hierarchies change from one source to another. The same inconsistency also forces commerce teams to correct listings, pricing and stock reports manually.

A Unified Commercial Performance View Connects Spend to Sell-Through

A unified commercial performance view brings media exposure and promotional investment together with orders, sell-through, margin and inventory movement. It extends measurement beyond a platform’s isolated return-on-ad-spend figure.

The foundation will not make attribution perfect, because marketplaces and retailers will continue to withhold some information. It does make the available evidence comparable and shows teams exactly where the remaining measurement gaps sit.

What the Foundation Can Change in 2026

Connected data matters when it improves a decision that teams make repeatedly. Four use cases provide realistic starting points for retail and CPG enterprises in Malaysia.

Retail Personalisation Can Respond to the Whole Relationship

Retail personalisation becomes more useful when recommendations consider current stock, recent purchases, loyalty status and cross-channel behaviour together. The system can stop recommending an unavailable product or repeating an offer after the customer has already purchased.

ADA’s retail capabilities bring together unified customer intelligence, AI-powered experiences and commerce execution. The commercial aim is to improve conversion, basket size and purchase frequency using decisions informed by the whole relationship.

Retail Media Can Be Measured Against Commercial Outcomes

Retail-media reporting usually ends with impressions, clicks, attributed marketplace orders or platform return on ad spend. A connected foundation can bring product margin, promotional cost, repeat behaviour and downstream inventory into the same assessment.

The discussion can then move from the campaign with the most platform revenue to the investment that produced profitable, incremental sell-through.

Demand Forecasting Can Read More Than Historical Sales

Demand forecasting improves when models consider promotions, media activity, local events, prices, weather, stock and channel demand alongside past sales. Inventory status is particularly important because zero sales may signal no demand or simply no product available to purchase.

The forecast must also enter planning and replenishment workflows. A more accurate number that remains inside a dashboard cannot prevent a stockout.

Trade Promotion Analysis Can Separate Volume From Value

A CPG promotion may lift unit sales while reducing contribution after discounts, retailer funding and media costs. Connecting promotion, sell-through and margin data allows teams to compare incremental volume with the full investment required to generate it.

Commercial teams can use that evidence when negotiating with retailers, setting promotion calendars, planning pack strategy and allocating products across channels.

Malaysia’s PDPA Makes Governance Part of the Architecture

Malaysia’s Personal Data Protection (Amendment) Act 2024 came into force in three phases during 2025. The amendments introduced direct security obligations for data processors, higher penalties, breach-notification duties, data-protection-officer requirements and a right to data portability.

Governance cannot sit in a policy document disconnected from the platform. Retailers and CPG enterprises need those rules to operate through the data pipeline.

The foundation should record:

  • The source and permitted purpose of customer data
  • Consent status and any subsequent change
  • Data ownership and accountable business stewards
  • Role-based access to identifiable and sensitive information
  • Retention, deletion and portability requirements
  • Transformations between source records and business metrics
  • Logs showing which person or system accessed protected data

These controls are not limited to compliance. They determine whether teams can trust, reproduce and responsibly use a marketing audience, customer profile, forecast or AI recommendation.

Build Around One Decision, Then Expand

Connecting every source or replacing every system before producing a result is the slowest path to value. ADA recommends beginning with the commercial constraint, then working backwards to the minimum data needed to improve it.

1. Name the Decision That Is Currently Weak

Select a decision with an outcome the business can measure. Marketplace media allocation, priority-SKU replenishment, loyalty reactivation and promotion effectiveness are practical examples.

“Create a single customer view” describes a platform ambition. “Reduce duplicated offers to active loyalty customers” describes an operating decision whose result can be measured.

2. Map the Evidence and the Gaps

Document the systems, owners, update frequencies, identifiers and quality problems behind the decision. Promotion analysis may require media spend, campaign taxonomy, orders, discounts, fees, product cost, inventory and returns.

3. Establish Shared Customer, Product and Metric Definitions

Agree on the authoritative identifiers and the calculation of core measures. Active customer, completed order, available stock, net sales and incremental volume must mean the same thing across teams.

4. Connect and Govern the Minimum Viable Data

Build ingestion, validation, identity or SKU matching, quality monitoring, lineage and access controls for the selected sources. Use batch, near-real-time or streaming pipelines according to how quickly the decision must be made.

5. Put the Output Inside the Workflow

Deliver the result into the environment where the responsible team plans, activates, approves or replenishes. Depending on the decision, that could be a campaign platform, planning tool, marketplace operation, CRM or supplier workflow.

6. Prove the Commercial Result Before Adding Sources

Measure the baseline, the change in the decision and the resulting business outcome. Once the first use case works, its governed data products can support related decisions.

ADA estimates that an initial data assessment can be completed in two to four weeks. Priority sources may then be activated within eight to twelve weeks. Broader platform programmes often run for four to six months, depending on the number and condition of source systems.

How to Evaluate an Enterprise Data Platform Partner in Malaysia

The right partner is not necessarily the one with the longest list of technologies. It is the one that can connect local commercial realities with enterprise-grade data engineering.

Assess partners against six questions:

  • Can they connect regional sources? Look for experience with APAC marketplaces, payment gateways, retail systems, local CRMs, distributor feeds and legacy applications.
  • Can they resolve both identity and product records? Customer identity alone does not solve the CPG problem. SKU, pack, retailer, seller and location matching also matter.
  • Can they govern data from ingestion to activation? Quality, lineage, consent, access, masking, retention and audit logs should operate as one control chain.
  • Can they work with the current cloud stack? The design should accommodate Snowflake, Databricks, AWS, Google Cloud, Azure and BigQuery rather than force another silo.
  • Can they carry insight into action? Reporting by itself does not change a promotion, campaign, forecast or replenishment decision.
  • Can they prove commercial value? The programme needs baselines and measures tied to margin, conversion, frequency, sell-through, availability or forecast accuracy.

ADA has built more than 500 connectors, including custom integrations for difficult sources. Its Data & AI Foundation services cover integration, governance, platform implementation, customer data, activation readiness and AI engineering.

Frequently Asked Questions

What Is an Enterprise Data Platform for Retail?

An enterprise data platform for retail connects customer, product, inventory, order, media and operational data through shared definitions and governance. In Malaysia, it should support stores, D2C channels, marketplaces, loyalty systems, payment sources and local data-protection requirements.

Why Is Retail and CPG Data So Fragmented in Malaysia?

Retail and CPG data in Malaysia is split across marketplaces, social commerce, stores, owned sites, distributors, media platforms and internal systems. Each participant controls one part of the customer journey, uses different identifiers and reports performance on a different schedule.

How Does a Data Foundation Help CPG Brands Understand End Customers?

A data foundation gives Malaysian CPG brands a way to connect consented direct interactions with retailer, marketplace, media, product, promotion and sell-through signals. It may not identify every buyer, but it provides a clearer account of who buys, where demand forms and what moves products.

Is a Customer Data Platform Enough for a Retail Enterprise?

Customer data platforms support identity, profiles, segments and activation. They do not replace the broader foundation required for product, inventory, finance, supply and commerce data. ADA implements CDP capabilities within a governed data foundation so customer activation does not become another isolated system.

Which Use Case Should a Malaysian Retailer Start With?

The starting point for a Malaysian retailer should be one frequent, measurable decision affected by fragmented data. Promotion measurement, priority-SKU replenishment, cross-channel loyalty reactivation and retail-media allocation are strong options. The right choice depends on the value at risk and the quality of available source data.

Turn Fragmented Commerce Signals Into Decision-Grade Intelligence With ADA

Better decisions become possible when Malaysian retail and CPG enterprises stop reporting demand, product movement, media, inventory and margin separately.

Retailers gain continuity across customer journeys. CPG companies gain visibility beyond sell-in. The same foundation can then support personalisation, forecasting, retail-media efficiency and AI using trusted commercial context.

ADA unifies customer, product and commerce data for activation, analytics and AI. To identify the first commercial decision your data foundation should improve, speak with ADA.

Table Of Content
TL;DR: Fix Fragmented Data
Malaysia’s Retail Growth Is Producing More Signals, Not More Clarity
Retailers See Transactions but Lose the Customer Between Channels
CPG Enterprises Have a Harder Problem: The End Customer Is Often Invisible
The Three Views a Retail and CPG Data Foundation Must Create
What the Foundation Can Change in 2026
Malaysia’s PDPA Makes Governance Part of the Architecture
Build Around One Decision, Then Expand
How to Evaluate an Enterprise Data Platform Partner in Malaysia
Frequently Asked Questions
Turn Fragmented Commerce Signals Into Decision-Grade Intelligence With ADA