MDM and the basics of Customer data platform

Customer data platforms(CDP) have evolved from a mixture of distinct technology stacks. How does a CDP platform differ from a combination of Customer MDM, Data Quality, CRM, Customer Data Mart/Data Warehouse, Customer support/ticketing, Web analytics, and other applications?

Well, functionally not very much except the heavy lifting of stitching together the data from all the different applications and creating the unified data platform gets handled by the CDP platform with a minimum amount of coding. It becomes easy for organizations to analyze customer behavior and respond to customers quickly with the appropriate responses.

What is needed to develop a customer data platform?

1. Customer Event Tracking

2. Data Collection and Transformation

3. MDM with strong Identity Resolution

4. Flexible database storage

5. Reporting Analytics

6. Machine Learning Libraries

7. Data Delivery

Let's look at each piece of this stack:

Customer Event Tracking:

Customer event tracking enables the tracking of the various activities a customer/visitor performs on an organization's website, app, e-commerce site, blog, etc. Event tracking is usually a snippet of code that resides in each of the applications and captures the interactions of the users and sends this back to a CDP data collection and transformation module.

Data Collection and Transformation:

Data received by the CDP can get processed in realtime and batch from different applications; these can communicate with the CDP using different protocols. After the data is collected, it needs to get transformed into a standard predefined data format.

Master Data Management and Identity Resolution:

The single most crucial step allows the CDP to group customer events to a given customer. Master data management is required to build a single version of the customer. Without a single version of the customer, your CDP implementation will suffer as events do not get attached to the correct profile because of duplicate or missing information. Identity resolution can get complicated when unique identifiers for a customer/visitor are not available. Most CDP platforms depend on the availability of a unique identifier to perform Identity resolution.

Flexible Data Storage:

Another critical piece in the CDP platform, flexible data storage is needed to handle massive volumes of data and different data schemas that need to be stored and retrieved. Support for analytical queries and retrieving specific records from hundreds of millions of records efficiently and quickly is a key requirement.

Machine Learning Libraries:

ML libraries surface the intelligence and the insight from the customer data captured by the CDP platform. Real time personalization, dynamic segmentation and interactive web engagement are some of the key benefits delivered by the ML libraries in a CDP context.

Reporting Analytics:

Reporting analytics provide the visualizations and easy to read data for users to understand trends, summarize information and gain actionable insights. Collaboration among different teams becomes easy with shared reports and story boards.

To summarize, while customer data platform is something that has been around in the marketing and technology worlds albeit as siloed applications with limited integrations, a true CDP provides organizations a quick, dependable path to provide their customers with superior experience.

Check back soon for posts about the reference implementation of OpenDQ Customer data platform, this will be a zero license cost solution with option to run in the cloud or in-premise datacenters.