Customer 360 in retail: how to connect CRM, ecommerce, ERP and POS without duplicating customers

In retail, the same customer usually exists several times without anyone noticing: one version in the CRM, another in the ecommerce, another in the physical store’s POS. Each system knows a part of that person, their online history, their store visits, their support complaints, but none has the full picture. Customer 360 is exactly that: a unified customer view combining all those sources into a single profile.

This article explains how to build that view by connecting CRM, ecommerce, ERP and POS, and above all how to do it without ending up with more duplicate customers than you had at the start.

Quick summary

• Integrating more sources without identity resolution does not improve the customer view: it multiplies the existing duplicates.
• Identity resolution is a set of business rules, not a function that comes included in a tool.
• The hardest case is the physical store customer, who often buys without identifying themselves.
• It is better to start with two or three high-value sources and sustain the syncing, instead of integrating everything at once.

Why the same customer ends up duplicated

When a person registers in the ecommerce, buys in the physical store and then calls customer support, it is common for each channel to register them separately with slightly different data: an alternative email, the name spelled differently, a phone number updated in one system and not in another. Without an identity resolution process, each system assumes these are different customers.

There are causes that repeat in almost every operation:

• Each channel defines its own key. The ecommerce identifies by email, the ERP by tax ID, the POS by phone number or by nothing.
• In-store purchases are usually anonymous. If there is no incentive to identify yourself, that transaction is not associated with any profile.
• Data is entered by hand at the counter or over the phone, with the expected variations and errors.
• Previous migrations dragged in duplicates that were never cleaned up.

What data to integrate to build the Customer 360

Source	What it contributes to the profile
CRM	Commercial interactions, opportunities, campaigns, contact history
Ecommerce	Online purchases, abandoned carts, browsing preferences
ERP	Invoicing data, commercial terms, credit limits in B2B
POS	Physical store purchases, which usually stay isolated from the rest
Customer support	Complaints, queries and returns
Loyalty	Points, benefits and tier within the programme
Logistics	Deliveries, shipping incidents and usual addresses

A prioritisation criterion that works: start with the sources that contribute reliable identity data (ERP and ecommerce usually have tax ID and verified email) and add afterwards the ones that contribute context (support, loyalty, browsing).

How to avoid duplicating customers: identity resolution

The technical process behind this is called identity resolution and consists of defining rules to determine when two records that look different actually correspond to the same person, and merging them into a single profile.

There are three levels of rules, and a serious implementation uses all three:

1. Deterministic matching. Two records share a unique, reliable identifier: tax ID, verified email, loyalty programme member number. It is the safest rule and the one that should be applied first.

2. Probabilistic matching. There is no shared identifier, but several attributes coincide: similar name, same address, same phone number. It requires defining a confidence threshold and deciding what happens below it.

3. Manual review. The cases that fall in the grey zone are routed to a review queue instead of being merged or discarded automatically. Mistakenly merging two different customers is a harder problem to reverse than leaving them separate.

What to do when data conflicts

Two records for the same person with different phone numbers raise a question that has to be answered before integrating: which one wins? The usual policies are to prioritise the most recent data, prioritise the source most reliable for that field (the ERP for tax data, the ecommerce for email), or keep both as primary and alternative.

What matters is that the policy is defined per field and documented. Without that definition, each sync produces a different result depending on the order in which the data arrived.

Stages to build a Customer 360 in retail

1. Identify the customer data sources

CRM, ecommerce, ERP, POS, support and loyalty, prioritising the two or three that contribute the most value to start with. Record for each one which identity fields it contains and how reliable they are.

2. Audit existing duplicates

Before integrating, measure the problem: how many duplicate records exist in each system under the rules you are going to apply. This number usually comes as a surprise and is what justifies the project.

3. Define the identity resolution rules

Which fields determine that two records are the same person, what threshold is accepted in probabilistic matches, and what to do in the face of data conflicts.

4. Unify and expose the profile

Make the consolidated profile available to the teams that need it, sales, marketing, support, as a single source of truth, not as an isolated report generated once. If your CRM is HubSpot or Salesforce, the profile syncs with the CRM without manual uploads and is visible to sales and marketing.

5. Maintain the syncing over time

Updates in any of the source systems should be reflected in the unified profile. A Customer 360 built once and not maintained goes stale within weeks.

6. Close the identification loop

Solve the root cause, not just the symptom: encourage identification at the point of sale, unify capture forms across channels and validate data at the moment of entry.

Frequently asked questions

Do you need a CDP to have a Customer 360? Not necessarily. A CDP is a specialised and useful tool, especially for marketing cases with a high volume of behavioural data. But a unified customer view can be built with an integration layer that applies identity resolution over the existing systems. The decision depends on the use case, not on the product label.

What about physical store customers who do not identify themselves? It is the structural limitation of Customer 360 in retail. It is mitigated with incentives to identify, loyalty programme, digital receipt, extended warranty, but there will always be a percentage of anonymous transactions. It is worth measuring that percentage and treating it as an indicator to improve, not as a failure.

Is it risky to merge profiles automatically? Yes, if it is done without thresholds and without reversibility. Merging two different customers creates privacy and commercial data problems. The recommended practice is to merge automatically only on deterministic matches and route the rest to review.

How does personal data regulation affect this project? It has a direct impact. Consolidating data from several sources requires reviewing the legal basis for processing, informing appropriately and being able to handle access, rectification and deletion requests on the unified profile. It is best to involve legal from the design stage, not at the end.

How long does it take to build a Customer 360? The technical integration is usually not the longest part. What dominates the schedule is cleaning up existing duplicates and reaching internal agreement on the identity and data precedence rules.

Checklist to build a Customer 360 without duplicates

• Are there clear, documented rules to identify when two records are the same customer?
• Was the number of existing duplicates audited before integrating?
• Is the precedence policy defined field by field for data conflicts?
• Do ambiguous cases go to manual review instead of being merged automatically?
• Does the unified profile include online and physical store purchase history?
• Do the sales, marketing and support teams access the same profile?
• Are updates in one system reflected automatically in the unified profile?
• Was regulatory compliance for the processing of consolidated data reviewed?

You may also be interested in reading:

• “Master data in ecommerce: what it is and how to integrate it for AI, reporting and automation” • “How to integrate POS, ecommerce and ERP without breaking inventory, pricing or invoicing” Weavee connects CRM, ecommerce, ERP and POS through its systems integration for retail to build a real Customer 360 view, with configurable identity resolution rules and continuous syncing across every channel.