Where there is hype, there is curiosity. But in marketing technology, hype is often accompanied by confusion. Martech buyers continually strive to stay on top of what’s new and exciting. This requires going through product marketing vocabulary and understanding what a particular topic or trend can do to create value and how it applies to your organization’s needs.
Composable CDP is another such topic, which sparks various interpretations and unique positionings from different vendors trying to stand out from the noise. Buyers should be wary of this trend, as some details are not fully explained by sellers in their stories.
Brutal Truth 1: Composable CDP is not a thing
As momentum builds around composable architectures, messaging aligns with the enthusiasm. CDP Providers quickly adopted this model and even described their platforms as composable. However, a tool itself is not composable. Architectures are composable. Tools that support interoperability and MACH principles are good citizens within a composable architecture.
CDPs are particularly useful when organizations cannot implement their own cloud-native customer data store and can more easily implement a commercial off-the-shelf solution, or “packaged” CDP, that helps them leverage benefits of cloud technologies by managing and making decisions on customer data.
As organizational cloud capabilities have matured, brands are building their Single Customer View (SCV) with cloud-native applications and services that solve for identity, modeling and data quality. They no longer need the CDP to solve the SCV. Some CDPs are adapting to how brands are changing their approach to managing customer data in the cloud. This allows for a modular approach to integration into the martech stack and rebranding into “composable CDP”.
The Key to Buyer Success: Taking a Capability-Driven View
Once the SCV is migrated to a cloud database, the software provided is simply no longer a CDP. The software has abandoned responsibility for identity resolution, data integration, and data storage capabilities.
In a composable architecture, what remains of the “CDP” functions above all as an orchestration platform. Its main capabilities include:
- Audience management.
- Travel management.
- Develop predictions with machine learning.
- Activation of customer data.
Ultimately, they help you make decisions and deliver experiences based on those decisions.
These are essential capabilities for personalization and customer experience, which are not trivial. Finding the best solution for your organization’s needs requires a thorough analysis of these capabilities.
Brutal Truth 2: You Don’t Need a ‘Client 360’
In a composable architecture, orchestration vendors (calling their platforms names like “data enablement,” “reverse ETL,” and “composable CDPs”) will tell you that you just need a “Client 360.” residing in a cloud database.
For too long, the industry focused on creating this 360-degree view of the customer – creating a longitudinal view of all of a customer’s online and offline interactions – before it could truly understand who a customer is and how to have a conversation with them. them. This approach requires:
- A comprehensive approach to identity resolution.
- Significant data collection and, therefore, data quality effort.
- Extremely complex data integration implementation.
In reality, 80% of organizations fail to achieve Customer 360, so pursuing it is not a wise investment in time and budget.
Key to Buyer Success: Start with “Customer 101”
Organizations today collect enormous amounts of data in their data lakes. What is needed to get started is “Customer 101” – the introductory level or beginner’s perspective in understanding the customer domain. Building a client 101 requires focusing on the right data – the data that matters most and will power your early machine learning efforts.
If you start by truly understanding the questions that different parts of the organization want to ask of the data, you’ll likely find that to answer 80% of those questions, you only need 20-30% of the data. This is your customer 101.
Over time, you will enrich your understanding of customers with more data. This should be accomplished by:
- Start with less data.
- Test and learn.
- Fail fast.
- Adding data based on these learnings.
Introducing machine learning capabilities is the best way to accelerate your learning and get more from less data by starting with the right data: your Customer 101.
Dig Deeper: Gartner: Give up the complete view of the customer
Brutal Truth 3: Composable CDP is not a new paradigm
Proponents of Composable CDP might have you believe that it’s an exciting new approach to managing customer data and your marketing programs. However, this is just another example of “what’s old is new again.”
Direct marketers have long created centralized databases for marketing execution, analysis and measurement, connecting an orchestration tool to create audiences and distribute them to activation platforms. We used to call this composable architecture a marketing database with campaign management, statistical modeling, and reporting tools.
Key to Buyer Success: Choose Your Modules Wisely
Things have evolved since the original composable martech stacks. Real-time data pipelines and API integrations are now supported. Channels have become more numerous and more complex, beyond the traditional direct outbound marketing touchpoints of telemarketing, direct mail and email. So we need more modern tools.
Understanding the capabilities and level of sophistication you need is the first problem to solve. For example, marketers want to be able to do more in real time, but can they really understand and explain what that means and how it can be enabled?
Real-time capabilities are not just a single concept. And real time isn’t always a strict requirement. There is a place in marketing for real-time, relevant data: real-time data processing that takes context into account.
Dig Deeper: Limitations of “real-time” CDP use cases
Brutal Truth 4: A composable CDP approach is difficult to deploy
Taking a modular approach with tools focused on easy integration offers clear efficiencies, but let’s not forget where the real hard work happens and where many composable CDPs have abdicated all responsibility: data management.
No single platform or tool solves the data quality problem, and it will never go away. Your machine learning, measurement, and targeting are only as good as the data that powers them.
In addition to solving data quality, you need to ensure full control of your data with governance, traceability and cataloging while ensuring compliance and confidentiality of your customer data.
Key to Buyer Success: Consider All of Your Client’s Consumers 101
When packaged CDPs handled data management, they focused largely on orchestration and activation use cases. Composable CDPs don’t just require you to solve these use cases.
Centralized customer data stored in a data cloud must consider all consumers of that data, including marketing reporting and analytics. More data consumers mean more questions to ask about the data to understand what a customer data store should contain.
The future is bright
Vendors that describe themselves as composable CDPs, such as ActionIQ, Simon Data, and Lytics, are adapting their platforms to the modern data stack. Making these adjustments is important and helpful for organizations that are ready to adopt a modern data stack, and these vendors will be well positioned for the market in the future.
What doesn’t help is the additional confusion they create in an already confused CDP market by calling themselves “composable CDPs.” These vendors must embrace their scalability and emphasize the mature capabilities they offer to create and derive value from data in the customer data store. Let’s just come up with a new name or stick to a conversation around capabilities.
Get MarTech! Every day. Free. In your inbox.
The opinions expressed in this article are those of the guest author and not necessarily of MarTech. Staff authors are listed here.
