CDP (Customer Data Platform) is a misnomer

I am not much into controversial writing or challenging the norms. However, my countless discussions with data teams (data architects, enterprise architects, CDO, Head of engineering, Digital Product Owners) over last two years has irked me enough into writing this post. I am probably writing this more for myself so I can share this as a pre-read to anyone who wants to discuss “When we have so many data applications in our business, why do we need a CDP?”. I hope this will also help some of the ‘helpless’ marketers who are trying to build a business case for CDP in their organisation but are getting push back from their data team, who may misunderstand CDP as a cheap imitation of one of their sophisticated data gadgets.

Allow me to begin with a controversial disclaimer a “Customer Data Platform (CDP) is NOT a data platform”. It holds data, but it’s not your traditional data platform.

If I was an influencer or had the aura of a thought leader, I would rename a CDP to a “MAP (Marketing Activation Platform)”. IMHO, CDP (or MAP) is a ‘marketing’ platform built for the marketers so they can do their job more efficiently and independently. More on that later, but let’s first spend a moment on data platforms.

 For better part of my career, I have worked in data roles like DBA, data modeler, ETL/BI developer, data architect. So, when I hear data platform, I think one of the following systems.

• ERP: Most granular and the front-line data collection system designed for specific business applications like HR, Supply Chain, Ecommerce, Finance etc.
• CRM: All customer interactions including lead, contact, account, opportunity, customer support
• MDM: think golden record and data quality
• Data Warehouse: Consolidated organisation data into clean structure for downstream business applications
• Data Mart: Subset of data warehouse meant for a specific business function
• Data Lake: All data (structured and unstructured) aggregated for known and unknown use cases
• Data Mesh: Decentralised, interoperable, domain-specific, API-driven micro-services approach for data management and self-service layered with strong governance, with infrastructure comprising of data lakes and data warehouses
• Data Fabric: Centralised and automated data integration layer including data bus, algorithms, APIs, SDKs providing access to all data applications

 I am sure there might be other more advanced data platforms, but the key takeaway from above mumbo jumbo is that a data platform is a ‘technical’ platform requiring data engineers, architects, stewards, modelers, admins etc.

CDP on the other hand, is an ‘activation’ platform which takes marketing specific subset of pristine data from various data platforms so marketers can create segments and share with marketing channels to run marketing campaigns.

CDP is nothing like a typical data platform. Now, to humour me, let’s relabel CDP as MAP and define what are the functions in MAP, which are not common in data platforms.

Some essential features of a MAP are:

• Data connectors: ability to easily ingest first-party data (there is none to minimal ETL here)
• ID stitching: in-built function to create a single unified profile based on stitching known identifiers (for cookie mitigation)
• Audience Builder UI: The key use case of a MAP is for non-technical marketers to easily create advanced segments in minutes using a drag and drop UI. This alone is a deal-breaker if a MAP does not have this feature
• Marketing Channel connectors: ability to easily share segments with paid, owned and outbound marketing channels
• Customer Privacy/ Preference: Easy UI to manage data privacy labelling, customer preference and opt-out rules

Some non-essential but desirable features of a MAP are:

• Real-time activation: ability to qualify and share segments in real-time
• Business layer: store data in a business-friendly taxonomy hiding the complexity of underlying data relationships
• Data marketplace: ability to ingest second- and third-party data
• Data query/ transformation: basic query and ETL operations
• Reporting UI: segment and marketing performance pre-canned reports, not necessarily your enterprise reporting
• AI mystical powers: anything that helps with task automation or reduces cognitive load on the marketing team like predictive segments, intelligent insights, automated decisioning rules

I have seen data teams in global enterprises trying to build a MAP and end up making a ‘slim’ data lake with high maintenance custom integrations. They generally create a subset of MAP which has primitive ID sync tables, no provision for streaming data, activation only for known profiles, lacking an audience builder UI forcing marketers to create a service ticket every time they need to create or update a segment taking so long, the marketing teams stop using the platform.

I hope some of my views makes sense for my friends in the data space. Next time, you hear ‘CDP’, remember it’s not about building another technical data platform, but instead think ‘MAP’ and consider it as an activation platform to enable personalisation for marketing teams who need to consume data from your traditional data platforms.

Happy Personalisation!

#cdp #customerdataplatform #map #marketingactivationplatform #personalisation #data #marketing