Marketing segmentation concept. Businessman looking at customer profiles on a tablet
PHOTO: Shutterstock

This article is part 4, the final article, in this series on customer data platforms sponsored by Jahia.

The value a customer data platform (CDP) can bring to your company is extensive, and a must-have for any company serious about running customer-centric, data-driven programs. By serving as the hub of your martech stack, a CDP can help to create a more complete view of an individual customer by bringing in data from a variety of sources, sorting it, and providing actionable direction to marketers and CX professionals to help drive modern personalization campaigns.

Best Practices for CDP Success

The first way you can head off any potential problems and concerns while setting up and getting your CDP operational is by doing the proper planning and following some best practices, both which can help mitigate risk further down the line. These best practices will help set the stage for your CDP at your company, align management to your core KPIs for success, and ensure that the CDP you do get has future-proof functionality that will allow your system to grow as your business and customer behaviors evolve over time.

  • Get Executive Support – Since CDPs are intended to connect into multiple data sources across business units and silos, having all internal partners on board with the project, from evaluation through installation and operations, is critical to success. And the only way to get all those stakeholders on board and committed to the project is if you get senior executive buy-in and support first.
  • Define Your Measurement for Success – CDPs may not generate revenue directly, so it is important you provide a business case and rationale up front to senior management that helps tie the value that a CDP can bring to your business and customers overall.
  • Define Roles and Responsibilities – Defining who internally manages your CDP from a technology side is important. In fact, according to data from The CDP Institute website, organizations that have a defined owner manage martech (either marketing or IT) are more successful than one that have the responsibility shared.
  • Use Schema-less Data Ingestion – Using schema-less (a.k.a. databases without a fixed structure) ingestion to aggregate data from different sources is another must-have feature in a CDP. Without schema-less data ingestion, all data that comes into the CDP will need to be formatted and structured; limiting what data you can take in.
  • Use Dynamic Profile Segmentation – With dynamic profile segmentation you can more effectively target specific groups of customers with personalized content and offers.
  • Keep Your Data Persistent – Look for a CDP that keeps data indefinitely, so you can get a persistent and more complete view of your customers, across time and multiple interactions.
  • Use Machine Learning and AI – Some CDPs offer machine learning that can help sift through your unstructured data and deliver insights and analysis. An AI system could use those insights to deliver actionable guidance or automation for your personalization tools.
  • Don’t Forget Security – Ensure the CDP you choose has security features and governance rules that meet compliance and regulatory standards like GDPR and the The California Consumer Privacy Act.

Related Article: Tips for Evaluating, Setting Up and Supporting Your CDP

Open Source and Development Approach

As an alternative to a proprietary CDP, an open source CDP may be the right solution for your organization depending on your requirements. Open source systems offer a more modular approach than proprietary systems, allowing them to be more flexible and scalable, as well as connect easier to other platforms through open standards.

Most importantly, as opposed to a proprietary offering, the open source approach helps you maintain complete control over your system as well as its managed data, no matter where the CDP is hosted.

Beyond data ownership, a central feature of any CDP is the ability to collect and use data to and from other applications, and allow for that data to be modified. Finding an open-source CDP with open APIs that allow for such modularity is paramount.

As an example, here are two key initiatives that focus on open source and open standards:
  • OASIS Context Server (CXS) - The OASIS open source interoperability standard is an initiative to advance standards for personalized user experiences online. The Context Server allows for simplified management, integration, and interoperability between services like WCM, CRM, Big Data, machine learning, digital marketing and DMP platforms.
  • Apache Unomi - Apache Unomi is the first Java-based open source CDP server that is designed to manage customer, leads and visitor data for personalization while offering GDPR compliance. Apache Unomi is designed to allow for extended and easier integration of external data sources and help standardize personalization and offer experience management and user privacy controls.

Related Article: Why a CDP is a Must-Have for a Unified Customer Experience

Getting Data into a CDP

Since a CDP is intended to be the hub of your martech stack and will bring together data from an assortment of internal and external systems, planning where you will get data from, and what methods you use to load it, is something you want to pay close attention to.

  • From Internal Systems – These are the core customer touchpoints (first party data) that will feed your CDP, and will make up the foundation of that data you feed into it. Internal systems include things like your website, e-commerce and mobile apps, POS, customer support or call centers, billing and loyalty programs. You want to make sure there are connectors that your internal systems can integrate into a CDP, like through an application programming interface (API).
  • From Tracking Tags – A tag management system (TMS) is a great way to get data ready for input into a CDP. They allow you to get your data clean, correlated and standardized for CDP ingestion. Not all CDPs offer the ability to grab tags or interface with a TMS, so find out whether your CDP offers this and gives you the ability to add, configure and remove tags.
  • SDKs – Some CDPs come with software development kits (SDKs) that can be embedded in a variety of mobile apps and IoT systems, allowing you to gather more valuable data from channels that may not be as readily available as your internal systems.
  • External Data – CDPs should be able to bring in data from external sources as well as internal. This would include getting third party data from ad networks and business databases, or second party data from business partners.
  • Methods for Loading Data – Loading data into your CDP can be handled most commonly through an API, but you want to be sure you are aware of what functionality the API offers; whether it offers batch processing; security features; and whether there are any data load limitations in terms of size or frequency.

Related Article: What a Customer Data Platform Can Do For Your Company

Getting Data out of a CDP

Once all that juicy customer data has been piped into your CDP, in order to make it actionable and help to drive and influence your personalization messaging campaigns, you need the proper methods and functionality to export that data into other apps in your martech stack, as well as other external systems.

  • Batch Files – All CDPs should have the ability to create a large batch file to be passed to external systems. This would be the most typical type of data output from a CDP. These types of files are easy for external systems to import, but can be large and take time to process. Other types of output options are necessary if you want to do more real-time processing and get the most up-to-date data.
  • API – API connections are quicker than batch files connections and are the preferred method for real-time access to individual customer data. APIs are flexible though, and can handle groups and segments as well as individual customer data. Important information to always keep in mind when dealing with APIs are questions around supported formats; whether the API is published and has documentation; and are there industry standard connectors.
  • SDK – SDKs are mostly used to export data to mobile apps, and can be used at trigger points (like when a user signs into his account) to deliver highly targeted and timely data.
  • Integration – Integration with the rest of your martech stack and external applications is core functionality of a CDP. Many will have built-in connectors to personalization tools, e-commerce tools, and advertising. These built-in integrations will make it easier to leverage your data and get ROI from your efforts.

CDP Performance Concerns

The performance of your CDP platform is a top concern like any application, but especially so here since there are times when a CDP needs to take in and deliver real-time data of an individual customer. Similar concerns that a web platform would have around response times, latency issues, and scalability are all performance concerns you want to be planning ahead for.

Latency refers to how long it takes for new data that is gathered to be ready for the CDP. The data you may be collecting could be in unstructured form and need to be delivered in a structured format (that is clean, accurate, and standardized) that is useable by the CDP. Also, your source systems may deliver data in real-time, or more periodically over days, weeks or months. Once the data is ingested, the next performance concern is response time, which refers to how quickly the CDP can return data when requested by ad networks, personalization tools, or product recommendations.

Finally, to future-proof your CDP, scalability should be top of mind in your planning. Your CDP needs to be able to handle as much data as your company can throw at it, and while it may be capable now, think down the line to all the types of integration and sources it will need to tap into. Latency and response times can be affected by data volume, so how you scale will directly impact your CDP performance.

Finally…

CDPs are complex projects and operations that require buy-in and close collaboration across your entire organization. Start with some good planning and follow solid best practices to head off any potential challenges during installation and operations, and ensure whatever CDP and vendor you do go with has the type of functionality and scale that will allow you to maximize value for your business and build value and trust with your customers over the long term.