Customer Data Platform Explained for Online Retailers

A customer data platform (CDP) joins customer records from your store, email, advertising and support tools into one profile that other tools can use. It becomes useful when those systems disagree about who a customer is or what they’ve done. Many smaller retailers can get a workable joined record from their existing email platform.

What a Customer Data Platform Does

A CDP collects customer activity, matches it to the right person and makes the joined record available to other systems. It follows three steps.

1. Collect events from connected sources, such as a product view, an order or an email click.
2. Join those events into profiles using customer identifiers and matching rules.
3. Send segments or profile updates to destinations such as your email platform and advertising tools.

Identity resolution is the matching step. In a hypothetical example, Alex browses anonymously, then signs in and buys a jacket. A CDP can connect that visit to Alex’s customer ID when the store supplies a reliable link between them. Alex’s support email can join the same profile if its address matches the verified customer record.

A shared first name isn’t enough to join two records. Matching unrelated shoppers can send the wrong message or expose another person’s order history.

Vendor feature lists usually include identity matching, segmentation, consent records and connectors for moving data between tools. Data cleaning removes duplicate or malformed records. Some platforms also calculate predictions, such as likely purchase behaviour, from the data you provide. Those predictions depend on the underlying records and need a defined business use.

The CDP organises the data you already collect. The guide to first-party customer data explains what belongs in that data before it reaches a CDP.

Single Customer View

A single customer view is one joined profile showing the customer’s activity across the systems you connect. It gives staff and marketing tools the same record to work from.

Consider Alex again. The store records a jacket order, the helpdesk records a delivery question and the website records an advertising click. The joined profile holds those events under one customer ID, with their source and time. It also keeps Alex’s email permission separate from the order itself.

That record can prevent duplicate campaign emails and remove recent buyers from acquisition advertising. When connected to the helpdesk, it lets service staff see orders from every connected store channel. Each benefit depends on sending the relevant update to the tool that acts on it.

To create a single customer view, work through the following decisions.

1. List the sources you need, including store orders, email activity, advertising events and support records.
2. Choose a matching identifier, such as the store’s customer ID, and define how other systems map to it.
3. Remove duplicates and set rules for conflicting details, such as two different email addresses.
4. Decide where the joined record lives and which tools receive updates from it.

Any solution needs a unique customer ID, consent flags, source history and updates frequent enough for the intended action. Keep a process for correcting mistaken matches. A profile that updates overnight can support a weekly campaign, but it won’t stop an email already queued after a purchase.

A customer database stores the records, but your team must build the matching and connections if they aren’t supplied. A customer relationship management (CRM) system can hold a joined service or sales record. An email platform can hold enough purchase and engagement history for a smaller store. A CDP earns its cost when joining more sources and distributing updates becomes a recurring problem.

CDP vs CRM vs Email Platform

A CDP focuses on joining data across systems, a CRM helps staff manage customer relationships and an email platform sends campaigns and automated messages. Their features overlap, so compare the workflow you need as well as the category name.

Tool	Typical data sources	Main users	Main output
CDP	Store events, orders, email activity, support and other connected sources	Marketing and data teams	Joined profiles and segments sent to other tools
CRM	Customer details, sales activity, orders and service interactions	Sales and customer service staff	Customer records and staff workflows
Email platform	Subscribers, store orders and message engagement	Marketing staff	Campaigns, automated emails and messaging audiences
Data management platform (DMP)	Advertising identifiers and audience data	Advertising teams	Audiences for buying and targeting ads
Data warehouse	Historical records from many business systems	Analysts and data engineers	Queries, reports and datasets
Marketing automation tool	Campaign responses, contact details and connected events	Marketing staff	Triggered messages and campaign workflows

A warehouse can hold customer profiles as well as inventory and financial data. It needs additional matching and delivery processes to turn those records into usable marketing audiences. A composable CDP builds on that warehouse, with separate tools for identity matching and sending data to destinations.

Klaviyo includes profiles, segments and integrations in its standard data platform. Its Advanced Klaviyo Data Platform (KDP) adds further data-management features. An email platform with those capabilities can already cover the work a retailer expects from a separate CDP.

Some vendors use “customer data management platform” loosely. Ask what the product stores and what it sends out. Persistent customer records joined across your own systems indicate a CDP-style function. An audience built mainly around advertising identifiers indicates a DMP-style function. Neither label guarantees the connectors or permissions your store needs.

How Retailers Use a CDP

A joined profile lets marketing tools act on a customer’s activity across connected systems. Three common uses show where the joined profile changes an outcome.

• Exclude recent buyers from acquisition ads. An order event removes the customer from the eligible audience, subject to the advertising platform’s matching and refresh delays.
• Trigger personalised emails. A customer who bought walking boots can receive relevant care advice if their permissions allow it.
• Build audiences from browsing and purchases. Customers who viewed raincoats but haven’t bought one can form a different audience from existing raincoat buyers.

These actions need reliable event names and a destination that accepts the audience or trigger. Adding a CDP won’t repair missing order events or create permission to market to a customer.

Customer Data Platform Vendors and Pricing

CDP vendors include enterprise suites, standalone data tools and marketing platforms with customer-data features. The best fit depends on the systems you need to join and the people who will maintain them.

Vendor type	Examples	What to consider
Enterprise suite	Adobe Real-Time CDP and Salesforce Data 360, formerly Data Cloud	Fit with the wider suite, implementation work and usage allowances
Standalone customer-data tool	Twilio Segment	Source connections, identity matching and delivery to existing tools
Email platform with CDP features	Klaviyo and Advanced Klaviyo Data Platform	Whether its existing profiles and store integrations already meet your needs
Marketing platform with customer-data capabilities	Bloomreach	Joined data alongside campaign delivery and the contract’s audience allowances

A real-time customer data platform updates profiles as events arrive. For example, a jacket purchase can change Alex’s audience membership during the same visit. The result still depends on event delivery, profile processing and the receiving tool’s update speed. A real-time profile doesn’t guarantee an advertising audience changes instantly.

Pricing can depend on profile count, tracked users, event volume or processing usage. These measures produce different bills. A visitor who browses twenty products creates more events than someone who views one, even though each is one visitor.

Twilio separates its Connections data pipeline from its fuller CDP package. Connections pricing uses monthly tracked users, including identified and anonymous users. The CDP package adds profile unification and audience features, with pricing through sales.

Klaviyo bills Advanced KDP by total profiles, including suppressed profiles and non-subscribers. Adobe’s Real-Time CDP pricing page directs businesses to customised pricing instead of listing a subscription amount. Salesforce lists both profile-based and Flex Credit options for Data 360. Bloomreach’s profile-based Marketing contracts use billable profiles and monthly unique visitors to set the annual subscription tier and included allowances.

For a small ecommerce business, start with the customer-data functions in the email platform you already use. A separate pipeline makes sense when moving data is the problem. A full CDP is a larger commitment when identity matching across sources and audience delivery are both needed.

Include connector setup, staff time, data cleaning and ongoing maintenance in the comparison. A subscription alone doesn’t describe the cost of keeping your customer records usable.

What a Shopping Agent Adds to a CDP

An on-site shopping agent adds conversation activity that order records alone don’t explain. Its records can include chat transcripts, clicked product IDs and the signed first name and loyalty tier supplied by the store.

shoppa is an AI shopping agent that helps customers find, compare and buy products in the chat, using your catalogue and documents. It also handles order support and returns. Its signed personalisation lets the store pass a verified first name and loyalty tier. Those details help personalise the conversation, while a reliable customer ID is still needed to join activity to a CDP profile.

In a hypothetical example, Alex asks for a waterproof jacket and clicks two recommended products before buying one. A retailer’s data connection can attach those events to the same customer ID used for Alex’s order, email clicks and support ticket. The transcript explains the question. The clicked product IDs record which recommendations Alex considered.

Send fields that support a defined customer action, with the necessary permissions. Useful fields include the customer or session ID, event time, product ID and an explicit preference the customer has shared. Keep first name and loyalty tier as store-supplied attributes, with their source recorded. An anonymous session can stay anonymous until a reliable identity link exists.

Detailed answer traces and session performance measures can stay in the agent’s analytics. shoppa’s analytics show conversion, click-through and basket measures for sessions, with a trace of each conversation. They don’t establish that two sessions belong to the same person. A CDP doesn’t need every transcript copied into it to use a product-click event.

Decide which fields move through your data connection and restrict access to transcripts. That design connects the agent’s records to other tools without treating the presence of a shopping agent as an automatic CDP integration.

When Your Store Needs One

Your store needs a CDP when fragmented records repeatedly block actions that your current tools can’t handle. A dedicated marketing team and five or more customer-data sources are useful signals, but the source count is a rule of thumb.

For example, joining store orders, email activity, advertising events, loyalty records and helpdesk tickets can create recurring manual work. A CDP becomes a reasonable choice if the team needs the same audience across several destinations and can maintain the matching rules.

Many smaller stores get enough from their email platform’s order sync and customer profiles. If the platform already delivers the required campaigns with correct exclusions, another system adds cost and maintenance without solving a distinct problem.

Write down one action the disconnected records prevent, which sources it needs and where its output must arrive. Buy a CDP when its matching and connections solve that action better than your existing tools.

Privacy Obligations for Customer Data

The Australian Privacy Principles (APPs) govern how organisations covered by the Privacy Act collect, use and disclose personal information. Joining records in a CDP doesn’t expand the purposes for which you can use them.

The Office of the Australian Information Commissioner’s APP summary sets out the duties. Covered retailers must limit collection to permitted purposes, provide required collection notices and use or disclose data only as the law allows. Direct marketing has its own conditions. Sending records overseas also raises cross-border disclosure duties.

Keep consent and opt-out records with the profile, restrict access and protect the data. Set retention rules and support customer access and correction requests. A stored consent flag needs a stated purpose and channel. It isn’t blanket permission for every connected tool.

Privacy Act reforms are under way. The Attorney-General’s Department describes the reform programme, including enacted changes under the 2024 amendment and further proposals being developed. Further legislation can change retailers’ duties. The guide to Australia’s Privacy Act for retailers covers the broader obligations.

Before joining sources, document what each field is for and which systems receive it. Carry those restrictions into your matching and audience rules so the joined profile remains usable for permitted customer actions.