Customer Data Platform

What is a customer data platform?

Customer data platform (CDP) software is an all-in-one solution that combines data from multiple sources, organizes that data around individual customers, and enables the streamlined activation of that data within other systems.

Without CDPs, retail systems and teams tend to be channel-based and the data isn’t easily translated across systems. Marketers may have to wait weeks for IT and analytics teams to manually build reports and draw insights—and by that time, the data may be outdated. Additionally, siloed datasets can lead to data discrepancies due to human error, as well as duplicate customer records across multiple systems and touchpoints.

These data challenges make it difficult for businesses to understand their true relationship with each customer, slowing innovation and hampering the overall customer experience. A real-time customer data platform solves these challenges by creating a clean, comprehensive, and persistently-updated single customer view.

Because the data is combined and standardized into an easy-to-understand format, CDPs reduce the amount of time, effort, and resources required to discover customer insights, orchestrate personalized experiences across channels, and quantify the impact of various business activities on actual customer behavior.

Learn how to measure the impact of a CDP, with key CDP metrics and considerations for your team. 

Types of customer data platforms

The four primary functions of CDPs include:

• Data ingestion.
• Unification into a single customer view.
• Analytics and insights.
• Activation across platforms.

However, different CDP vendors have varying strengths and weaknesses in each of these areas. The different types of CDPs can be broken down into four tiers based on their core focus and functionality: data ingestion, unification, intelligence, and activation.

Data Ingestion CDPs

CDPs focused on data ingestion tend to have quick, easy integrations with most common data sources. Pre-built integrations and flexible APIs allow CDPs to ingest data in either batches or streams with little to no heavy lifting required by IT teams. However, key vendors for data ingestion sometimes struggle with actually unifying the customer record and empowering business users to easily analyze and activate data across platforms.

Lexer’s integration process has been described as “the smoothest integration” in one customer’s 10 years of IT management experience.

Unification CDPs

After combining multiple data sources, the CDP also needs to be able to match those different data sources to individual customer profiles to create a “golden record” for each customer, also known as a “360-degree customer view.” This matching can be done using deterministic matching, which matches customers on exact information (such as email and phone numbers), or probabilistic matching, which matches customers on inexact information (such as a combination of last name and postal address).

CDPs with a strong focus in unification typically use machine-learning and artificial intelligence to aid this matching process. Additionally, unification CDPs might strengthen customer profiles by enriching them with data from second-party or third-party sources such as Experian’s Mosaic.

Intelligence CDPs

The single customer view within a CDP can give you an unprecedented look into who your customers are and how they behave across channels and touchpoints. However, simply having a single customer view won’t provide much benefit to your business. You need to pair it with analytics tools to gain sophisticated insights to inform your decisions.

The best CDP vendors for insights and decision-making provide out-of-the-box audience segmentation, predictive analytics, personalized recommendations for next-best actions, and robust measurement and reporting tools to help you track and optimize your performance across every engagement channel.

Activation CDPs

Once you’ve gained the ability to make insight-driven decisions based on comprehensive, real-time customer data, you need to be able to efficiently execute on those decisions. Activation CDPs are the highest-tiered types of CDPs, because they enable businesses to quickly draw insights, build personalized campaigns, and orchestrate those campaigns across systems and touchpoints.

The best CDPs for activation can connect to any channel, including email, social, website, and more. By connecting to your existing engagement channels, CDPs allow you to build, trigger, and automate personalized campaigns from one tool. Although CDPs are generally considered to be a marketer-managed system, CDPs with both analytics and activation capabilities can add value to every business function.

How does a customer data platform work?

Customer data platforms must go through four steps to create the core single customer view that powers every other CDP use case:

1. Integration
2. Data Cleaning
3. Unification and Identity Resolution
4. Enrichment

Once this core single customer view is created, a customer data platform’s capabilities depend on the specific tools offered by the vendor, as well as the strategies you use when analyzing and activating the data. 

Let’s break down each step in the CDP set-up process.

Integration

To achieve a single customer view, you need to bring together all customer, transaction, product, and engagement data from every source, including:

• Ecommerce
• Retail POS
• Service
• Product
• Email
• Website
• Loyalty
• Reviews
• Surveys

Each of these data sources includes its own customer ID and its own record of events, maintained in its own schema, in its own cloud, all with a different API. The information held in each source includes demographics, purchase histories, customer service interactions, web and mobile browsing activities, email engagement, and more.

In order to create the best possible relationship with your customers, you need to collect this valuable data and pull it together using common linkage variables such as name, email address, and phone number.

Data Cleaning

Once you’ve gathered all this data, you need to prepare it so that it’s fit for calculation. In other words, you need to remove errors and duplicates across systems and data formats to ensure the highest-quality information in your single customer view.

In data science, it’s common to spend about 80% of your time and effort in data preparation and analysis before you can move onto higher-impact activities like predictive modelling. Streamlining this process is one of the most important benefits of CDPs.

The process for data cleaning involves:

1. Validation: Ensure all provided data is correct, consistent, and reliable.
2. Unification: Link records and remove duplicates.
3. Normalization: Transform all data into one consistent, straightforward format.
4. Categorization: Categorize data for easy segmentation, insight, and activation.

Clean data makes it easier to compare records and get the best results.

Unification and Identity Resolution

Data unification is also known as deduplication or identity resolution.

The three key principles of data unification within a CDP are:

1. Identity graph: A collection of known customer identifiers that can be associated with one another. Think of this as a really big space with every record of every customer represented as individual points in that space. The goal of identity resolution is to draw links between each of those points.
2. Deterministic matching: Exact matching on known fields such as email and phone number.
3. Probabilistic matching: Matching using fields that aren’t exact matches, such as a combination of last name and postal address.

A CDP can take a deterministic, probabilistic, or combined approach to identity resolution, and it’s up to you to define how you want your data linked. The choices you make around identity resolution are central to how you understand and communicate with your customers, but they also need to work with your existing operational systems and business reporting.

Data Enrichment

Finally, once you’ve cleaned and matched your data into a standardized format, you need to transform it into information that’s easy for business people to understand. Predictive metrics and calculations help you create more value from your data than ever before.

These CDP metrics and calculations can include:

• Inferred gender
• Total spend
• Average order value
• Average time between orders
• Sales channel preference
• Predicted customer lifetime spend
• Predicted orders
• Churn risk
• Next best product recommendations

… and more!

Once your single customer view is created, you can use it to fuel every other CDP use case using built-in customer data platform tools such as audience segmentation, cross-channel campaign automation, surveys for data collection, and measurement and reporting dashboards.

Customer data platform implementation

Major IT projects such as martech integrations are notorious for scope creep, dragging timelines, and heavy lifting by IT teams—but a customer data platform implementation doesn’t have to cause headaches, as long as you approach it properly. 

Customer data platform requirements for a smooth implementation include:

• Clearly defined goals and objectives.
• A carefully-chosen CDP vendor.
• A data-driven, customer-first culture.

Goals and Objectives

CDPs can solve multiple problems across the business, so it’s important to be specific about the challenges you’re experiencing, your top business goals and objectives, and key CDP use cases you’re interested in prior to the implementation. Clearly articulating these details ahead of time will help ensure an accurate scoping process and a well-managed implementation down the line.

CDP Vendor Comparison

The CDP landscape can be complex, with a variety of vendors offering different tools, integrations, and specializations. It’s important to carefully evaluate each potential vendor to ensure you’re choosing the right CDP for your business. By aligning your outlined goals and objectives with the vendor’s ability to deliver on those goals and objectives, you can choose a reliable vendor who will help you manage a smooth implementation process.

When evaluating different vendors, be sure to ask about:

• Current customers and relevant industry expertise
• Onboarding, support, and strategic consulting services
• Out-of-the-box analytics and customer intelligence capabilities
• Activation, automation, and measurement tools
• The level of technical expertise required to use the CDP
• Available integrations and APIs
• Security features and compliance certifications

Data-Driven Culture

CDP technology will only get you so far. The people and processes driving the use of that technology are what will create the truly transformative, revenue-generating impact of a CDP. The most effective CDP implementations are always accompanied by a customer-centric culture across the business and an adaptable team that’s eager to test and learn. You need to train your team to use the new platform effectively, fold insight-driven decision-making into existing business processes, and encourage a test-learn-optimize mindset to ensure continuous improvement.