Artificial Intelligence (AI) and Machine Learning (ML) are dynamically changing Customer Data Platform (CDP) analytics by enabling these technologies to access multiple data points, across departmental silos, resulting in a more accurate customer profile and the ability to provide real-time decisioning. This article will look at the ways AI-based CDPs provide the customer with a better experience along the entire customer journey.

The use of CDPs has been growing for several years now. According to Gartner’s 2019 Marketing Technology Survey, 43% of brands polled had a fully-deployed customer data platform (CDP) and 31% were working to implement one. As customer data complexity continues to increase exponentially, the use of CDPs will continue to rise.

What Are the Functions of a CDP?

Gartner’s recent report, the Market Guide for Customer Data Platforms, defines a CDP as being used for "centralizing data collection, unifying customer profiles from disparate sources, creating and managing segments, and activating those segments in priority channels." A CDP is useful to marketers, sales teams, customer service agents, and other customer facing business functions.

Given that customer data is spread across all of a brand’s channels, as well as hidden away in data silos in various departments, in disparate software platforms, the number one function of a CDP is to unify all of that data in one location. CDPs typically have pre-built 2-way connectors which connect them to the most common sources of customer data, such as marketing clouds, tag managers, billing, ecommerce systems, and paid media channels. Brands that use proprietary software are often able to use custom connectors, though this involves additional development time and costs. 

Additionally, a CDP facilitates the creation of customer segments, however it goes further than that and enables the hyper-segmentation of customers. It enables a brand to target very specific groups of customers, and also allows the brand to exclude or suppress other specific groups of customers who, for example, are not likely to be interested in what the brand is offering. For instance, there may be a specific group of customers who are between the ages of 20 and 35 who are Miami Dolphin fans and live in the Miami area — but targeting the males in that group with ads for feminine Miami Dolphin hoodies might not be as effective as targeting the females in the same group. 

Along with data collection and analysis, CDPs are typically able to cleanse the data, removing duplicates and irrelevant characters, standardizing the data to be more effectively used for the creation of unified customer data and unique profiles. 

Sam Ngo, director of product marketing at BlueConic, said that a CDP is able to tame the beast that customer data has become. “Historically, building AI models based on customer data has been time-consuming, labor-intensive, and requires constant handoffs between teams and systems for data collection, segmentation, and activation. By creating a unified profile database that’s connected to a company’s activation channels, a CDP enables data science and marketing teams to work more efficiently.”

By working together using a CDP, different departments are able to create custom predictive models using the unified data collection. “For example, data scientists, analytics, and business intelligence teams can import notebooks and customize predictive models in a CDP and apply these models on top of the first-party database. Models can be set to refresh at a certain interval, and scores get automatically updated in customer profiles which can then be used for segmentation,” said Ngo.

Marketers are able to use a CDP to fine-tune marketing campaigns, through cooperation with other teams within the brand. “These capabilities not only create shared efficiencies between data science, BI, and marketing teams, but also enable companies to create vastly better experiences for customers by shortening the distance between insight and activation,” he said.

Brands can no longer rely on third-party cookies, so the ability to use first-party data has huge implications going forward. “Having a first-party data asset is also a game changer for things like lookalike audiences. As third-party cookies disappear and third-party data becomes less reliable, having the ability to model audiences based on first-party data becomes even more critical,” said Ngo.

Related Article: How Customer Data Platforms Can Benefit the Call Center

Real-Time Decisioning

To be able to recommend the next best action requires the most recent, up-to-date data, which is simply not possible without a platform which unifies the latest customer data from across channels, in real-time. “Take customer churn as an example. If marketing is relying on business intelligence teams or data scientists to collect data from disparate systems, run predictive models, and create segments of customers based on their propensity to churn, the data could be weeks or even months old by the time it’s ready for use,” explained Ngo. “These kinds of operational inefficiencies can have a big impact on short-term marketing goals and long-term business profitability.”

A CDP is uniquely able to provide access to customer data from across all of a brand’s channels, which facilitates personalization, cross-selling, recommendations, and more. “By storing a customer or prospect’s cross-channel behaviors, interests, transactions, preferences, and more in one place, teams can then develop and execute strategies and campaigns to acquire new customers and retain, upsell, cross-sell existing customers much more effectively and efficiently,” said Ngo.

John Nash, chief marketing and strategy officer at Redpoint Global, thinks that real-time decisioning in a CDP is critical with consumer behaviors changing, as they now expect a brand to know them in real-time, across multistage and multichannel journeys. “There are an increasing number of use cases where this plays out — from BOPIS (buy online pickup in store) and abandoned cart marketing in Retail, to appointment setting and follow up engagement in Healthcare,” said Nash. “The key is to have all of the data current, precise and available for use in the cadence of the customer. This may be 5-minute windows for certain customer experiences (e.g., BOPIS) down to millisecond time frames for web or mobile personalization. By perfecting data in real-time at the point of ingestion — cleansing, matching and integrating it in a way that is fit for purpose — brands are in a much better position to deliver compelling and differentiated customer experiences.”

Learning Opportunities


A report from entitled Brand Loyalty 2020: The Need for Hyper-Individualization revealed that 81% of the 2000 consumers that were polled indicated that they were willing to provide basic personal information in exchange for a more personalized experience. Conversely, a Gartner survey on marketing personalization showed that brands stand to lose 38% of those customers that have experienced poor personalization practices. 

Not only is it more important to provide personalized content today, it must be presented based on the customers’ current state. An AI-based CDP allows a brand to adjust its marketing efforts to the current context of its customers through customer segmentation and real-time decisioning. Because a CDP utilizes first-party data which comes from customers that have purchased from or opted-in to a brand, a holistic view of each customer is developed based on their preferences, past history and current real-time behavior. “One of the biggest challenges to personalization at scale is having access to an actionable single customer view. A CDP like BlueConic collects and consolidates customer data from across systems and sources to create the most comprehensive and up-to-date record of what you know about your customers — including online, offline, behavioral, interest-based, and synthetic attributes,” said Ngo. 

CDPs are also able to incorporate privacy protections based on each customer’s individual preferences. “Perhaps most importantly, today’s era of explicit opt-in and consent requires tools that are built for compliance. A CDP helps companies put consent at the forefront of their personalization efforts by ensuring privacy statuses are captured and stored in each individual profile,” explained Ngo.

Nash told CMSWire that customers have higher expectations for how brands engage with them, and that the true power of a CDP lies in its ability to empower marketers to keep up in a controlled manner. “Despite the growing complexity, consumers expect not only that a brand recognizes them across every channel, but that every interaction is relevant to where they are in a customer journey,” said Nash.

To deliver comprehensive, multi-channel, multi-touch journeys which meet the needs and expectations of individual customers, brands are finding that an AI-based CDP is able to guide such an experience amid growing complexity, Nash suggested. “To do this, there is a growing recognition that an accurate, personalized and real-time customer profile is essential for driving decisions across multiple channels at the cadence of the customer,” said Nash.

Related Article: Customer Data Platforms: Some Assembly Required

How Can AI-Based CDPs Improve the Customer Experience?

CDPs such as Amperity, BlueConic, Redpoint Global, Adobe Real-Time CDP, and ActionIQ have integrated AI into traditional CDP elements to unify customer data and provide real-time functionality and decisoning for marketers, allowing them to gain a deeper understanding of what their customers want, how they feel, and what they are likely to do. 

With this deeper understanding of customers, brands are able to use actionable intelligence to create highly personalized experiences that resonate with customers. Additionally, AI and machine learning (ML) can be leveraged for more precise customer segmentation, as well as to predict which customers are least likely to be retained. This understanding enables brands to proactively take action to keep the customer within the sales funnel until they become a customer. With an AI-based CDP, brands are able to follow customers as they progress through their journey, and can minimize or eliminate any pain points before they become a problem. 

A brand’s customer experience initiative should drive their CDP strategy, as the ultimate purpose for having an AI-based CDP is to improve and enhance the entire customer journey across all of the brand’s channels. “When they are deployed in line with the customer experience and business operations, AI and machine learning are essential to advance toward a market of one — highly personalized messages, offers and content in the context of each customer journey. While it is impossible to make the necessary decisions and next best actions to scale at a purely human level, AI/ML empowers marketers to get to highly granular segmentation and predictive models that are tunable to meeting their priority business objectives,” said Nash.

Final Thoughts

An AI-based CDP can facilitate the collection, unification, cleaning, and standardization of customer data from across all of a brand’s channels, all of which enables the CDP to create hyper-segmented customer profiles. AI enables real-time decisioning based on current customer data, which facilitates the hyper-personalization that customers expect today. Additionally, an AI-based CDP allows brands to provide product suggestions and personalized ads based on both historical customer data and real-time “as it happens” interactions between the customer and the brand.