Composable CDPs: revolution or hype in customer data management?

Composable customer data platforms (CDPs) have recently attracted a lot of attention in the marketing and technology world. They promise a flexible, modular approach to customer data management. This should enable companies to modernise their data architecture with as little redundancy as possible. But is this more than just another buzzword to breathe new life into tried and tested concepts?

The following article provides an overview of how composable CDPs work and highlights the differences to conventional CDPs.

Conventional CDPs?

The term ‘conventional CDPs’ raises the legitimate question: What exactly is ‘conventional’ about it? To understand this, it is worth taking a look at the history of these platforms and the challenges they promised to solve.

Origin and evolution

Customer data is as old as business models themselves and existed long before the introduction of CDPs. However, it was usually spread across different systems such as CRM, marketing automation and e-commerce platforms. With increasing digitalisation, companies have realised that they need to bring these isolated data sources together to leverage synergies and get a holistic picture of the customer – this was a driving factor behind the development of the first CDPs in the early 2010s.

The first CDPs promised a centralised database that would enable companies to

• Better understand customer interactions across different channels
• Gain a single view of the customer (often referred to as ‘golden record’, ‘single point of truth’ or ‘360-degree customer view’)
• Make data-based decisions in real time
• Overcome the problem of isolated data silos

Further development of requirements

Over time, the requirements for CDPs have evolved:

• Flexibility: companies expect not only centralised data collection, but also responsive, flexible systems.
• Scalability: With the exponential growth of data volumes, CDPs must be able to process enormous volumes of data.
• Integration: CDPs should be able to integrate seamlessly into existing technology stacks.
• Data protection and compliance: With the introduction of regulations such as the GDPR, CDPs must fulfil stricter data protection requirements.

Limitations of conventional CDPs

Despite their advantages, conventional CDPs can reach their limits when it comes to integration into modern, agile technology environments:

• Lack of flexibility in adapting to specific business requirements
• Difficulties in seamless integration with existing systems
• Potential creation of new data silos

What do composable CDPs do differently now? A short trip to the modern data stack

The rise of cloud computing and affordable data storage solutions has opened up revolutionary data management opportunities for organisations. These technological advances have paved the way for the so-called ‘modern data stack’ – a concept that has fundamentally changed the way companies collect, store, analyse and activate data.

The Modern Data Stack enables companies to access a broad ecosystem of specialised solutions. Instead of investing in expensive, monolithic systems, companies can now combine modular, cloud-based tools that are tailored to their individual needs. This development has not only simplified data processing, but also made it considerably more efficient and scalable.

Cloud data warehouses such as Snowflake, Google BigQuery or Amazon Redshift are a central component of the Modern Data Stack. These form the foundation for advanced data analyses and enable companies to store and process large volumes of data cost-effectively. Building on this, various data integration, transformation and visualisation tools can be seamlessly integrated.

However, the major cloud providers such as AWS, GCP or Azure are not the best or most cost-effective solution for every company. There are examples of companies that have taken alternative paths:

• 37signals, known for Basecamp and HEY, reported savings of around 2 million dollars by moving to their own servers.
• The German company everysize.com switched from AWS and GCP to Hetzner, a data centre based in Germany, thereby reducing its costs from 50,000 to 10,000 euros.

So there are viable alternatives to the big three cloud providers that can be considered depending on requirements and scaling needs.

Companies should review their options and consider hybrid or multi-cloud strategies to take advantage of different providers and optimise costs.

These developments in the area of the modern data stack form the context in which composable CDPs have emerged and can realise their full potential.

Core functions and integration

Essentially, composable CDPs share the basic goal of conventional CDPs: the collection, processing and activation of data from various sources. The key difference, however, lies in their architecture and integration method:

• Modular design: composable CDPs are characterised by their high level of modularity and flexibility. Companies can select and combine specific components to fulfil their individual requirements.
• Seamless integration: These platforms can be easily integrated into existing data infrastructures. Their modular approach enables better customisation to existing company technology.
• Zero data copy: A key principle of Composable CDPs is the avoidance of redundant data copies. Instead, centralised data management is made possible, which is more efficient and saves resources.

Data collection and integration: a new approach

A key advantage of composable CDPs is their flexibility in data integration.

While both traditional and composable CDPs integrate data from various sources, there is a fundamental difference in the way this data is managed:

• Traditional CDPs work with an isolated, proprietary database.
• Composable CDPs, on the other hand, build on existing data warehouses (DWH) – this is also referred to as ‘warehouse-native’

A special feature of composable CDPs is that they do not require rigid data schemas from a provider, but can be easily adapted to the company’s existing data structures.

This means that the raw data can initially be loaded into the data warehouse untransformed and the transformation is carried out flexibly and as required downstream – depending on what the data is to be used for. This decoupled approach not only offers companies faster access to their data, but also the freedom to transform data in such a way that it can be optimally utilised for analytical or operational purposes. As a result, data integration and transformation processes can be customised to the company’s specific requirements.

While this approach can speed up implementation, it also involves greater technical complexity.

Event collection of behavioural data

Another aspect of data collection in composable CDPs is the collection of behavioural data in real time. There are various approaches for this:

• Specialised composable CDP solutions: Some composable CDP providers offer their own event tracking functionalities and SDKs that can be seamlessly integrated into the existing data infrastructure.
• Integration of third-party tools: Companies can integrate specialised event capture tools such as Snowplow or Fivetran into their Composable CDP architecture. These tools are designed to collect behavioural data from websites, mobile apps and other digital touchpoints and process it in real time.

The choice of the right approach depends on the specific requirements of the organisation, the existing technology infrastructure and the desired use cases.

Data activation: Reverse ETL as a game changer

A key feature of composable CDPs is the ability to activate data stored in the data warehouse (DWH) in real time and transfer it to various operational systems – a process known as reverse ETL.

This approach offers several advantages in customer data activation:

• Efficient use of existing data infrastructures: Companies can optimally utilise existing data warehouses without building redundant databases.
• Real-time activation of customer data: Customer data can be fed more quickly into relevant systems (e.g. CEP, CRM, marketing automation tools) to be immediately available for personalised campaigns or analyses.
• Avoidance of additional data silos: The direct use of the DWH eliminates the need to maintain several isolated data stores.

Composable CDP ≠ Composable CDP

Not all composable CDPs are the same. Under the umbrella of the composable approach, the platforms offer a variety of functionality.

• Some composable CDPs have their own event processing systems that allow data to be collected and analysed directly.
• Others, however, focus exclusively on data activation using reverse ETL and leave data collection and integration to specialised third-party providers.

However, this versatility also makes the term ‘composable’ more complex, as not every system does the same thing.

Challenges / The downside of flexibility

While composable CDPs can offer advantages in terms of flexibility and adaptability, it is important to also consider the challenges.

• Increased technical complexity
  The modular nature of composable CDPs requires a deep understanding of different technologies and how they interact. Organisations must not only master individual tools, but also manage their integration and orchestration. This can lead to a significantly higher demand for technical expertise and increase dependence on specialised IT staff.
• Integration challenges
  Linking different tools and systems can lead to compatibility problems and data inconsistencies. Seamless integration of all components requires careful planning and continuous maintenance, which ties up time and resources.
• Increased administrative effort
  As the number of tools used increases, so does the administrative effort. Each component must be maintained, updated and monitored separately. This can lead to a fragmentation of responsibilities and affect the overall efficiency of the system.
• Cost aspects
  Although composable CDPs are often presented as a cost-effective alternative, the overall costs can increase due to the need for specialised staff, additional integration work and potentially higher infrastructure costs. Particularly worth mentioning here are the hidden warehouse costs, which should not be ignored, especially with more complex segmentation approaches: (see CDP 2.0 – Why Zero Waste Is Now from mParticle).

A careful balance between the benefits of flexibility and the challenges of complexity is crucial. For many organisations, especially those with limited technical resources, a traditional, integrated CDP solution may be the more practical and efficient choice, despite its limitations.

Status quo of traditional CDPs

Interestingly, established CDP vendors are also adapting to this composable trend. Tools traditionally known for their all-in-one solutions are increasingly opening up to more modular and flexible approaches and more of these vendors are now offering reverse ETL capabilities to enable DWH data. Examples of this are Segment and mParticle, both of which are now part of the MACH Alliance. This alliance is committed to modern architectures based on microservices, API-first, cloud-native and headless technologies. You can find out more about the MACH Alliance here

Summary

Composable CDPs represent more than just a single technology – they embody a holistic approach to customer data management. These systems form a comprehensive architecture consisting of different, harmonised solutions – also known as customer data infrastructure.

So are composable CDPs a revolution or a hype? The answer is probably somewhere in the middle: they represent an evolution driven by technological advances and changing business requirements.

For organisations looking for a flexible, scalable and future-proof solution for their data management, composable CDPs could be the answer. They make it possible to fully utilise the benefits of modern data technologies while at the same time managing the complexity of the data landscape. However, this approach also presents challenges – the necessary computing power for complex real-time requirements must be taken into account, as well as potentially non-transparent warehouse costs.

The choice between traditional and composable CDPs or a hybrid approach largely depends on an organisation’s specific requirements, technical capabilities and strategic goals. Carefully analysing your own needs and capabilities is essential to find the optimal solution.