Introduction

Customer Data Platforms (CDPs) are gaining prominence because of the immense amount of value that they bring to the marketing organization. As the industry is growing fast with new vendors joining quickly, it’s important for your organization to evaluate and identify the right CDP vendor to work with. Many vendors claim they provide a complete solution, although they may not. It’s therefore upon your organization to evaluate the vendors as they bring distinct capabilities that suit your business needs.

Performing a Proof of Concept (POC) helps provide hands-on evaluation of the products, which can be extremely effective while choosing the right CDP vendor. With many of the cloud-based CDP vendors providing a 1-month trial of their products, leveraging this offer to test the different capabilities that the vendors bring to the table can aid in making an intelligent product decision.

Some of the established CDP vendors are Hull, Lytics, FullContact Enrich API, Segment, Blueshift, BlueConic, Listrak, Ensighten Pulse, Evergage, V12 Data, Tealium AudienceStream, Arm Treasure Data, Radius, NGDATA Lily, RedPoint, Umbel, Signal, Zaius, SessionM, mParticle, QuickPivot and Lemnisk.

Investing in a CDP should be done with along-term aim – various systems from which the data is imported and systems to which data is exported may change over time, but the CDP becomes the master repository of data and should be leveraged by the marketing organization for a very long time.

Through a well-defined evaluation process including vendor demos and proof of concept, you can aim to select a vendor that can meet all your business’s current as well as future needs. The next few sections help explain various aspects of the CDPs and the type of questions that you need to ask the vendors.

 

Data Ingestion and Storage

An essential part of building a robust data platform is to ingest the data from various systems into it. Many CDPs provide built-in plugins to quickly import the data from other systems. Test the key plugins to see how the data is imported and maintained as this will impact the data model you’d want to create within the CDP. Check if the CDPs import from standard file formats and data streams. Additionally, it’s important to check if the data can be imported on a recurring basis or only as a one-time activity.

For use cases such as web personalization, it’s extremely important that the CDP import real-time data and you should be able to act upon it immediately. Some of the systems take a significantly large amount of time to import data, which means that web personalization using such systems would be difficult. It’s, therefore, important to understand when the data is available in CDP after the import is complete.

Data Ingestion and Storage

Different CDPs provide different ways of saving and retaining the ingested data. Based on these models, the retraction / roll-back of data (e.g., via an erroneous import) and how the data change is written in the database is managed differently across vendors.

 

Questions To Ask:

  • Can the CDP ingest data real time when an event occurs?
  • How does the CDP support structured and unstructured data?
  • Does the CDP support both anonymous and named user data?
  • Can the data be imported as a one-time event or on a recurring basis?
  • How quickly is the data available for use after import?
  • How is the data stored?
  • Can raw data be read from the system?
  • Which systems does the CDP connect out of the box to import data?
  • Does CDP persist old values or replace the values with new data upon any change?
  • How does retract/roll-back of an erroneous data import work?
  • Is it possible to re-create data at a given point in time?
  • Is it possible to define a persistence schedule of how long the data needs to be retained?
  • Is it possible to re-create data at a given point-in-time?
  • Is it possible to define a persistence schedule of how long the data needs to be retained?

Data Quality And Enrichment

Data quality is a key part of any data platform. As the data gets imported into the CDP, it needs to be validated and cleaned as needed. Different CDPs provide different built-in capabilities around how the data can be cleaned. The more the number of sources (especially external ones) that you bring the data from, the higher the probability of data being messy while importing and combining them.

Many CDPs provide a very thorough data cleansing capability built into the product, but some lack this functionality altogether and a separate data cleansing tool/product may be needed to perform this activity before the data gets imported.

To increase the value from your data, you can also enrich it with additional information provided by systems such as DMPs (Data Management Platforms) or other 3rd party repositories. Some CDPs provide plugins for DMPs which could be especially useful for data enrichment. Although data quality validation/cleansing in CDPs is currently manual (typically through rules definitions and scripts), as the products start maturing, you’ll see more systems introducing AI and machine learning based data cleansing and validations.

Data Quality And Enrichment
 

Questions To Ask:

  • Can data be checked for consistency while importing the data into CDP, e.g., date/email/datatype validation, etc.?
  • Can physical addresses be validated?
  • Can data values, formats, etc. be changed/cleaned?
  • Is there a rules-engine-based data cleaning that can be implemented?
  • Can tags be applied during cleaning?
  • Are data cleansing/tweaks/enhancements done first time the import occurs or every time the data is modified?
  • Does the CDP crawl across website(s) to capture data / tags based on user visit?
  • Can new fields be created based on certain data conditions / calculated values? (e.g. First Name + LastName)

Identity Unification

As the data starts flowing into the CDP, you need to unify the identities of the datasets to create the 360-degree profile view of the contact. CDPs may provide one of these two key capabilities for identity unification: direct match or fuzzy match. With direct match, the exact field value is used for mapping the different attributes in both the datasets. Whereas ‘fuzzy matching’ provides a more complex data matching using multiple data points and AI/ML (Artificial Intelligence / Machine Learning). By matching at least one field across the records, a chain of linked records can be formed within CDP, leading to a more complete 360-degree view of the contact. Based on your business needs and the tolerance for data errors, the right unification mechanism would make the most sense.

Another important aspect of identity unification is cross-device identity matching which helps tie up identities

Identity Unification

from different sessions from different devices that the same person accesses. While many systems don’t provide this capability as part of the product, some of them provide plugins and capability to integrate with 3rd party identity unification service providers.

 

Questions To Ask:

  • How is identity association done within the CDP?
  • Does the CDP unify identities with DMP systems?
  • Can identity associations be imported from other systems/databases?
  • Does the CDP provide any cross-device identity matching capability?
  • Does the CDP create a single persistent ID per contact?
  • Can the CDP anonymize data when needed?
  • Can you view the complete profile of an individual contact in detail in the CDP?
  • Does the profile show timeline-based details on various events for each contact?

Data Segmentation

With data imported, cleansed and unified, the next step is to group the contacts and create a targeted audience segment to act upon. Many CDPs provide a user-friendly interface to create very sophisticated segments. The more data you have within the system, the more refined segments you can create.

Some CDPs provide AI / ML driven segment recommendations such as potential leads, potential conversions, potential product recommendations, etc. This helps create targeted campaigns for that audience to improve the conversion.

While evaluating the CDP, it’s important to know how the segment data gets updated over time based on web

Data Segmentation

activities. Also, being able to track how contacts are automatically moving from one segment to another segment based on various rules and conditions can help create a better marketing automation workflow.

 

Questions To Ask:

  • How is segmenting done within the CDP?
  • How deep / complex can the data be segmented?
  • Can the CDP create any AI/ML based auto segmentation?
  • Does the segment get auto updated with new profiles over time based on the rules set for the segment?
  • Can a contact belong to multiple segments?
  • Does the system track which segment the contact belonged to and how that changed over time?
  • Can workflow rules be set for segmenting, e.g., is it possible to move a contact from one segment to another based on particular trigger/rule?

Reporting Capabilities

Most of the CDPs provide basic reporting capability built into the system. While many of them provide the ability to integrate with other reporting tools, only a few provide complex reporting capabilities as part of their product offering. Based on the skillsets available within your organization and reporting tools that your team is comfortable with, this might be a criterion for evaluating the CDP.

Through customizable templates and dashboards, you can provide the right visualizations that the leadership teams will require. Some CDPs enable you to create KPIs, trends and drill-down dashboards that can be shared with various teams, who can then view the holistic data and make smarter decisions based on the insights. Additionally, AI surfaced insights are being provided by some of the CDPs as well which can help identify trends that may not be seemingly obvious.

Reporting Capabilities
 

Questions To Ask:

  • What reporting capabilities does the CDP have?
  • Does the CDP provide customizable dashboards?
  • Does the CDP integrate with 3rd-party reporting tools?
  • What visualization capabilities does the CDP provide?
  • Does the CDP provide any AI-based insights?
  • Does the CDP provide a pre-built library of useful reports?
  • Can reports be auto generated and emailed to certain users?
Find Out Everything Your Need To Know About Customer Data Platforms (CDP) Download e-Book
 

Data Actioning

Exporting the data and acting upon it is the final purpose of a CDP. As you evaluate the CDPs, create a list of external systems that you need the CDP to integrate with. Some of established CDPs have a pretty exhaustive list of integrations with other systems that becomes very useful as you start actioning on them. Additional effort and cost should be planned if the integrations do not exist.

Check if the CDP provides real-time integration with ad engines. For instance, when a segment is updated, it can be sync’d with Facebook Audience immediately to target those audiences within their social channels in real time. Here are some of the common actioning platforms that you need to consider integrating with while evaluating vendors:

  • Ad servers (DSPs) for online advertising integration
  • Email campaign systems for email marketing
  • Data management platforms for syncing data insights
  • Mobile app notification services to send notifications (on any mobile apps you may have)
  • Other 3rd-party actioning systems that you may use

 

Data Actioning
 

Questions To Ask:

  • What data export options are available for CDP?
  • What are the supported connectors for actioning?
  • Can the data be queried via SQL, Web Services, etc.?
  • Can you create a scheduled batch export of data?
  • Is there a push mechanism (e.g., via webhooks) available for data export?
  • Does the CDP provide real-time integration with ad engines?
  • Can rules be defined around data export?
  • Can data be exported and reports be auto emailed?

Data Security

With ever increasing cyber-hacking, every organization needs to take data security very seriously. Depending on your industry, check if the CDP supports HIPPA, SOX and other industry compliance needs. Additionally, it’s important to define a data governance framework that provides complete guidance on which data can be accessed by whom.

Many organizations allow access of CDP data to multiple 3rd-party agencies and vendors for reporting and actioning purposes. This means defining appropriate roles and permissions to access the right data is critical. Some CDPs provide anonymization of data before actioning upon it. This helps remove personally identifiable information before the data moves from one system to another. Delve into these details if it’s critical for your business.

Data Security

From an infrastructural perspective, understanding how the data is stored, what underlying platforms are used and how the connections between various systems are configured is important to ensure data is secure.

 

Questions To Ask:

  • How are roles and permissions defined within the CDP?
  • What level of data access granularity controls are available within the CDP?
  • Can data access be restricted based on internal and external data sources?
  • Does CDP support data anonymization?
  • Is all user access to CDP logged?
  • Can notifications be set up based on rules around user access?
  • Is there an audit trail of every action performed by a user?
  • How is the data access restricted (if CDP stores data)?
  • What additional measures are taken to ensure that the data is secure?
  • Are there any data governance recommendations?

Performance And Scalability

If you have huge datasets to be imported into CDP or a website with heavy traffic, it’s helpful to perform a proof of concept before finalizing a CDP vendor. Some of the CDPs take a significantly large amount of time (sometimes days to weeks) during the first-time data import (especially ones performing algorithmic processing related to AI/ML). If there’s a big data feed that needs to be ingested into the CDP from a 3rd-party source on a recurring basis, this delay will impact when you can act upon the data after import.

Since a CDP is expected to be a long-term investment, it’s important that it supports a variety of input fields and that the data types are scalable. Similarly, the system should be able to auto scale itself to support the increase in transactions and storage data.

Understanding at what point the system starts to degrade in performance is a good metric to test during a proof of concept. This allows you to make the right decision before you end up investing time and money on an extensive CDP implementation.

Performance And Scalability
 

Questions To Ask:

  • How well does the CDP perform (across various metrics)?
  • Does tag implementation reduce the page load time?
  • Can you share some examples of data import performance metrics?
  • Can you share some examples of data export performance metrics?
  • Can you share some examples of data enrichment performance metrics?
  • Does system auto scale with an increase of data for storage / transactions?
  • Can the CDP handle scaling and updating of fields and data types?
  • How many simultaneous users / requests can be made at once via the API?
  • When does the system start to degrade in performance?
  • Are there any system constraints for the CDP?

Architecture

Architecture and usability of the CDP is important when it comes to acceptance from the IT team and the users of the system. Depending on whether the CDP is a cloud-based solution or on-premises solution, the implementation and planning changes significantly. Based on the organization’s needs and preferences, this will be one of the key criteria for evaluation.

Since many of the CDPs are going through their own maturity cycle, many of the user interfaces are in the early stages of evolution. So, beware of what’s being marketed vs. What’s reality. Having flexibility around types of templates, workflows, automated processes, notifications, report libraries, user permissions, etc., can be extremely valuable in the long run as your organization’s data needs evolve.

Understanding the product and architectural roadmap of the CDPs help map potential future features with your

Architecture

business needs. Additionally, understanding how the adoption of new technologies such as AI (Artificial Intelligence) and ML (Machine Learning) can help you gauge the technology maturity and future readiness of the CDP vendors.

 

Questions To Ask:

  • Can you explain the overall architecture of the CDP product?
  • Is the product cloud based or on-premise?
  • How much of IT involvement is needed during implementation and maintenance of the CDP?
  • Does the CDP provide a GUI interface?
  • Does the system support identity-based tracking?
  • Does the CDP support cookie-based tracking?
  • Can the user interface be customized?
  • Does the CDP provide any of the following AI capabilities?
    • Behavior insights
    • Predictive insights
    • Smart recommendations
  • Can queries and segments be saved and reused at a later point?
  • Can notification emails/text be sent on certain triggers?
  • Can automated processes be created within the CDP?

Business And Pricing Evaluation

The CDP industry is still in its infancy with newer players entering the industry on a regular basis. Here are a few commonly known CDP vendors, Hull, Lytics, FullContact Enrich API, Segment, Blueshift, BlueConic, Listrak, Ensighten Pulse, Evergage, V12 Data, Tealium AudienceStream, Arm Treasure Data, Radius, NGDATA Lily, RedPoint, Umbel, Signal, Zaius, SessionM, mParticle, QuickPivot and Lemnisk. As the industry goes through the hype cycle of growth, consolidation and maturity, selecting the right vendors is important to ensure the platform caters to the long-term need of the organization. Understanding the company’s history, financials, funding stage, number of customers, growth pattern, leadership backgrounds, product roadmap, industry focus, etc., helps you make a more informed decision.

Different vendors have different pricing structures – while a few provide one-time licensing fees, many prefer pricing based on number of profiles or based on the usage. Due to the early stage of the companies and lack of established large players, many of vendors are willing to work on the prices, especially for big brand customers.

Business And Pricing Evaluation

Knowing the future growth of data and possible future systems that would be integrated with the CDP, you should create at least a 3-year projection of costs (since it’s usually recurring license costs) and see how each vendor’s pricing model would differ.

 

Questions To Ask:

  • When was the business formed and how did it grow?
  • How much funding has the company received so far?
  • What's the Company’s revenue from CDP (licenses and services separately)?
  • How many employees do you have?
  • How many customers are using the CDP and how many new customers do you add on a quarterly basis?
  • What is your pricing model?
  • Are there any initial set up / training costs?
  • What's the ongoing cost for support?
  • Would implementation require your assistance?
  • Do you have an onboarding program for customers?
  • Do you have performance and uptime SLAs?
  • What is the future roadmap for the CDP?
  • What are the growth plans and targets for your business?
  • Which industries does your business focused on?
  • Can the company provide us free 1-2-month trial licenses for POC / testing?
  • Can you share customer references?
  • What type of support is provided if we migrate from your CDP to another CDP system? What data can be exported?
Need help evaluating Customer Data Platforms? Contact Us
 

Download With This Form

Download With a Tweet

Download with a Tweet