The concept of composable software is that systems are built and connected with modules. By deploying a composable architecture, companies can repurpose existing code to streamline their toolsets and stay ahead of their competitors.
The idea of composable software is beginning to make waves in the customer data platform (CDP) market. Composable CDPs, also known as unbundled CDPs, are being touted by CDP vendors that want to sell their systems as components. But, is this a valid selling point, or are buyers being led down a path that integrated CDPs may better serve?
Today, companies are gravitating toward integrated, purpose-built platforms that deliver on long-standing CDP promises, including real-time identity resolution, seamless activation, personalization, decisioning, and governance—without the operational headaches of a fragmented tech stack.
What is Composable Software?
Composable software components can be swapped easily when needed. This allows code to be written once and reused across multiple instances by breaking core applications into specialized microservices. Microservices make applications easier to scale and faster to develop, which can help to improve innovation and speed up time-to-market. They typically communicate through APIs.
Composable systems are distinguishable from platform architectures in that every component can be replaced, and no component is required for the entire system to function. In a platform architecture, replaceable modules depend on a shared core system that cannot be moved.
A CDP can be looked at as a composable module of the larger marketing technology stack, something that can be easily switched out for new software when needed, and that allows other software modules to be swapped out easily too.
Since CDPs are built from the ground up with built-in connectors or APIs, they can help companies make their technology stack easier to integrate and interconnect. A CDP makes it easier to collect and share data between different platforms, making it more practical for companies to use multiple best-of-breed Martech platforms. This allows brands to reduce dependence on a single vendor when building an ideal tech stack.
What is a Composable CDP?
Proponents of the composable CDP see the CDP itself as a software platform that could be broken up into components. They argue that companies should not buy an all-in-one CDP to handle everything from data collection to integration to profile unification, but instead purchase separate components for each function.
This is a similar argument to the overall all-in-one vs. best-of-breed discussion – that a CDP that is built from components can be easier to deploy, maintain, and operate.
But is this true? According to the CDP Institute, a CDP is “package software.” Meaning the CDP itself is a full suite, not a collection of modules. In that case, can a composable CDP really be called a CDP?
Composable CDP advocates argue that the core of the CDP, the customer database, does not need to be there, as many companies already have various data warehouse solutions. They argue that brands should supplement their legacy data warehouse solutions with modules to expand their functionality.
Unfortunately, most companies do not have a data warehouse, and even fewer have data warehouses with complete customer information. Even less have data that is appropriately cleaned, unified, and formatted in the way that marketers need. Only 39 percent of businesses have a data warehouse in place, according to the Harvard Business Review.
Unless a data warehouse was created specifically for marketing operations, it’s unlikely it will have the necessary functionality to flag inactive customers and calculate customer value.
The Challenges of Composable CDPs
As data volumes surge, composable CDPs falter under the weight of manual oversight and ad hoc scaling. Performance degrades, costs balloon, and teams can be left scrambling. Further, updates across the composable stack don’t necessarily translate to other systems, yielding broken models, delays, and often a full stop of operations.
Composable CDPs tout cost savings by leveraging existing data warehouses. But rather than eliminating compute expenses, they shift the burden upstream to platforms like Snowflake, Databricks, BigQuery, or Redshift. These systems end up handling intensive tasks like data transformation and identity resolution, running longer and racking up heftier bills. What appears as a vendor cost cut often masks a spike in overall expenses, leaving enterprises questioning the promised savings.
| High risk of failure | Hidden costs | Compliance chaos | Vendor fatigue |
| Too many components introduce a high risk of system failures | Staff costs: Forced to fund a large team to manage: – Integrations – Interoperability – Maintenance – Compliance wrangling- Development – Training | Difficult to maintain and audit global compliance for 5-7 vendors | Different vendor cultures |
| Composable architecture is still a relatively new concept with very little evidence it can handle scale | Exposes multiple security vulnerabilities across vendors, increasing breach risks | Disjointed Support, Sales, and Success teams | |
| Interoperability | Variable compliance across vendors | Strain on Procurement with multiple contracts | |
| Different roadmaps | Hidden usage/feature fees | Compromises regional compliance with gaps in global datacenter presence | Forced to negotiate multiple SLA environments |
| Extended deployments | Burdens IT with maintenance instead of driving innovation | ||
| Inconsistent AI strategy |
The Integrated, All-in-One CDP
Regardless of the current state of CDP vendor market positioning, most companies do not have data warehouses, and building out a CDP into modules may not make a lot of sense.
For the majority of enterprise customers, an integrated CDP will be just what they need, giving them the data collection and integration capabilities that are required in order to tailor the customer experience with personalized messaging.
An integrated CDP allows companies to capture customer data across any channel. It then can ingest that data, clean it, combine it into unified single customer view (SCV) profiles, orchestrate the customer journey, and store it indefinitely and persistently over time. Data warehouses simply cannot achieve this type of functionality, unless they are specifically built that way. And then, maybe they would be called a CDP?
Read More: CDP vs. Data Warehouse: What’s Best for Your Business?
Composable CDP vs. Integrated Customer Data Platforms
Having unified customer profiles that can be shared as a single source of truth across an enterprise is critical to brands to be able to differentiate themselves and deliver a personalized customer experience at scale. This need led to the development of the integrated, all-in-one CDP that combines all the functionality needed to create and maintain those profiles.
Composable CDP vendors are proposing breaking up the all-in-one CDP into components that can be purchased through separate vendors. But is this best for most organizations, or is it just another marketing tool?
Some organizations may want to go beyond composable and look into building a CDP from the ground up instead of buying a fully realized platform. According to David Raab, of the CDP Institute, building a CDP is a serious technical challenge, and for the greatest chance of success on this path, you need to find the optimal mix of both built and purchased components.
“Whether the composable approach actually offers this advantage depends entirely on the details of your company’s situation. It’s certainly not obvious that specifying, buying, integrating, and maintaining as many as a dozen new components would be simpler than deploying a single integrated CDP.” – (Composable or Integrated CDP? – CDP Institute/Treasure Data)
For most companies, an integrated CDP will be the right solution to help them understand their customers better so they can serve up data-driven, personalized customer experiences across any channel.
In the end, most companies are still beginning on their journey of leveraging customer data for business value. Trying to piece together a solution for customer data is not for the faint of heart. Look for an enterprise-grade CDP that offers all the features you need to construct unified customer profiles to benefit your business.
