How CDPs Will Evolve to Offer Full-Service to More Than Just Marketers
What features and benefits are ahead for customer data platforms.
- By Abhi Yadav
- November 8, 2019
Everyone knows that the secret to strong marketing is to know your customers better and to present them with appropriate selections in real time. In recent years, customer data platforms (CDPs) have been eyed as a vehicle for doing just that.
Soon they will do much more.
The evolution of CDPs concerns most businesses. A recent Forbes survey of 400 marketing leaders found that marketers think it takes too much time to analyze the success of a marketing program or to change the customer experience.
Most organizations are aware of CDPs but aren't yet using them to their full potential. The survey found that more than three-quarters (78 percent) of respondents have or are developing a CDP. Unfortunately, companies often think they have a CDP when in reality they don't.
The CDP and its Evolutionary Path
A CDP is a marketing system that pools customer data from multiple sources. In practice, it enables four main marketing capabilities: data collection, segmentation, profile unification, and activation. Gartner notes that features such as prediction modeling and decision management are optional for CDPs.
The raison d'être of CDPs was to democratize customer data. Marketers wanted to do away with separate marketing systems for different purposes and instead aggregate data. Business users soon saw the lure of CDPs: they could present insights from customer data. As CDPs migrated to more users, they became simplified and presented options based only on basic business scenarios. Marketers wanted to integrate disparate marketing and data systems while business users wanted self-service tools that allowed them to gain valuable insights from customer data without having to know much about computer science.
CDPs have progressed to include self-service analytics, but these tools are limited because data preparation is automated based on known use cases and predefined business scenarios. These tools are not capable of advanced analytics and offer users limited options, and basic dashboards and reports.
CDPs are starting to integrate predictions and decisions by adding machine learning model libraries with the promise of customer analytics for the citizen scientist. Solving basic marketing challenges (including churn, micro-segmentation, or prescriptive customer analytics) is complex, and a CDP's ML library doesn't do as good a job as cloud providers such as Salesforce, AWS, GCP, or Azure. AI's promise is appealing but the benefits depend on resources that may be beyond a client's initial budget or expectations.
It's worth pointing out that cloud services are not CDPs, although some cloud services are moving towards including some CDP features. That's why many mistakenly believe they have CDPs when they actually don't.
Where CDPs are Headed
Today, CDPs are defined by the industry based on a specific set of core features. In the future, what defines a CDP will be different than what we have today. I've outlined below where I think CDPs are headed.
Self-learning AI is an umbrella term for technologies that enable a system to become more intelligent over time. CDPs that leverage self-learning AI let organizations achieve much while requiring little user input. For example, marketers could define a campaign's target audience and set a goal for segmentation such as customers who are most likely to buy a category of products. A CDP with self-learning AI would analyze all the available customer data and run tests to find the segmentation that best suits that goal. Self-learning AI also enables automated feature engineering and ML model building. CDPs will eventually include self-learning AI as a key feature.
Leveraging data for analytics and models can be automated using machine learning and self-learning AI. For example, it's possible to entirely automate data prep, enrichment, and segmentation. CDP automation will also allow organizations to handle customer data so that privacy regulations and the privacy expectations of customers are met. Some CDPs already include automated features and automation will play a crucial role when it comes to CDPs in the future.
For example, for each customer record, an enterprise can use persistent ID management to track all raw data or events linked to that customer ID and receive continuously unified data without any ad hoc projects or ETL jobs. Further automated feature engineering can power the embedded analytics to continuously generate dynamic insights across each customer ID. Basically AI can find patterns and insights continuously across ongoing behaviors and intent data.
The most critical piece of a CDP is its ability to enable automated analytics or data science capabilities such as hyper-segmentation, data exploration, and predictions. However, when it comes to gaining critical insights about customers, many CDPs today fall short. CDPs will need to evolve to feature automated data science capabilities which will enable the CDP to provide recommendations and personalization based on the context and intent of each customer at the precise moment.
In the near future, CDPs will be automated such that the platform itself will make many critical decisions. For example, CDPs will determine automatically which data to ingest, which segments will bring the most value, and which customers to target and on what channels. CDPs powered by self-learning AI will be capable of automated decision making and prescriptive analytics.
CDPs Will Evolve to Include Use Cases Beyond Marketing
CDPs were initially designed as tools for marketing. In the future, however, automation and self-learning AI will enable the next generation of CDPs. These next-generation offerings will include many advanced capabilities and cover a wide range of use cases -- from omnichannel marketing campaigns and next best prediction for sales teams to real-time customer outreach and streamlined revenue operations for B2B. When it comes to the future capabilities of CDPs, the possibilities are endless.
Abhi Yadav is the founder and CEO of Cambridge, MA-based Zylotech, a self-learning customer analytics platform that keeps customer data live and enriched while automating the customer life cycle with relevance to continuously produce cross-selling, up-selling, and retention marketing results.