Unleashing the Power of Agility: The Benefits of a Composable vs. Non-Composable CDP

For Fortune 5000 companies, the stakes are exceptionally high in harnessing the power of customer data to drive marketing strategies. Customer Data Platforms (CDPs) have become the new go-to solution, offering a centralized repository for customer information. However, as the market dynamics shift towards more agile and adaptive models, new terms have emerged to challenge the status quo of CDP features and tools. So, the CDP as a concept may already be in decline and on the way out in-lieu of newer approaches. Vendors are throwing new terms around to help with differentiation. But, there are CDP vendors who resist some of these new terms and ideas, and their dismissal of them needs evaluation.

The newer messaging and terms are only trying to surface knowledge around the enterprise data platform to enhance the flexibility of the CDP as a complementary component in the enterprise IT/data stack. Enterprises have invested millions of dollars in data infrastructure, and a vendor pitching a monolith CDP will not replace those investments but should complement them.

One of these new terms surfacing this idea more broadly is composability or the Composable CDP. This term is not a groundbreaking idea or a new term. Composable Martech has existed since the beginning; we did not use or discuss this word in a context like we do today. Circa 1998/97 with Webtrends, Omniture, Doubleclick, Eloqua, MySQL(dB marketers), MDM/EDWs, and CRM/email systems all passed data between each other over FTP in CSV files - and some, god forbid, over HTTP via JS calls or RPC/COM/DCOM - REST had not even been invented yet. Skip ahead to today; all we have done is mature the approach into better I/O data transfer, give it a name, and separate features into more discreet micro-data apps that can be sold separately to enhance marketing delivery based on need in the enterprise. AKA Composability.

The composable approach promises enterprises more flexibility and leverage for deep IT capital investments. At the same time, the CDP solutions promise to help the marketer with a higher abstraction UI to interact with and utilize customer data no matter where it originates. The difference is how a CDP vendor will approach the problem to provide a solution for the marketer and how flexible that solution is for a combined data and marketing team. 

Therefore, there are two choices: a non-composable solution unable to connect to the back-office(e.g., Snowflake and Databricks), or has too many non-separable features(aka micro-data apps) or a Composable solution that is flexible, efficient, and interchangeable and can extract or query any data source and abstract it to the end user in a more flexible way(aka micro-data apps).

In most cases, all CDP vendors are already composable - some more than others. The issue is how many can connect to the capital-intensive IT/data investments already established in the enterprise. Unless the CDP vendor wants to own the data in the enterprise end-to-end, then the reality is that all CDPs must become fully composable. Even Adobe and Salesforce message and pitch flexibility to the enterprise. Adobe has been pitching the AEP as composable since the beginning of its release - they price it in this way too. Check out #3 in their help documentation on the AEP Key capability page here: Adobe AEP key capabilities.html

Suppose for a moment that the Martech stack has been 100% composable for decades.

• Why are some CDP vendors resisting the term "composable"?
• Why would a vendor negate or dismiss composability?
• Do they think they will own the entire data management stack?

Due to their static growth, some more prominent vendors could own the enterprise data stream, and some have moved in this direction to grow their SaaS ARR line further north. With its medallion data cloud CDP approach, Salesforce is starting the battle by moving to the back-office taking on Snowflake; even though they are promoting their partnership, smaller CDP vendors need more credibility to move in this direction. Salesforce is still a composable architecture - they can integrate into Snowflake and have separable data app products, sold separately and stand alone like Adobe. Many CDPs can do the same. Therefore, all CDP vendors must be composable if they want to compete and have deep integrations to these back-office origination data sources. So why resist the term, and why not embrace it? Or, why even use the term across the industry when there is a minority of non-composable contrarian vendors?

The same is true for Snowflake and Databricks. Does Snowflake want to begin moving towards creating marketing tools downstream of the data infrastructure and add them in as separable components? They now embrace smaller CDP vendors that can link and integrate into their platform. However, they could easily buy one of these vendors and quickly add the capability.

Questions you need to ask:

• What is the definition of composable? - "Composable refers to a system designed with interchangeable parts combined in different ways for flexible and efficient functionality." 

Rhetorically, this definition is a good thing for most Fortune 5000 enterprises, so why would they buy into a non-composable CDP system that is not flexible, non-efficient, priced in the totality of features, and does not have interchangeable parts to maintain modernity and future-proofing? 

What are the other questions to ask that would predicate a more composable CDP system:

• Do you have a considerable capital investment in back-office data platforms like Snowflake, EDW, and Databricks?
• Do you need every feature a CDP promotes - will the vendor turn features on and off based on SKU type - Micro-data apps? e.g. you could have already deployed an ID resolution in your Snowflake or Databricks investment with add-on tools like Zingg.ai
• Do you use any new modern data platform tools because of rapid digital transformation initiatives? e.g., Apache Open source tools, Kafka, Keeboola, Fivetran, dBT, Atlan, Zingg, et al.? If so, you should continue down the composable path or figure out a way to consolidate if that is the initiative.
• Do you want to remove your investments in data platforms or leverage them further?
• Do the new CDP platforms you evaluate connect easily to your enterprise data, like your Snowflake investment?
• Do the CDP vendors you evaluate have separable micro-data apps priced and with the ability to stand alone?
• Are you repeating your data infrastructure with a CDP vendor?

Would you have an architecture that has some or all of this illustration?

Most enterprises today have some if not most of this stack in-house to manage their data and all the facets of their businesses. Guess what – this is composability.

Another way to describe the two types of CDP vendors is to bucket them: Composable or non-composable. I wouldn't say I like this as I feel even the non-composable vendors are nevertheless composable in certain aspects even though they refuse to embrace the concept.

Therefore, the advent of Composable CDPs marks a significant departure from the rigid structures of Traditional or non-composable CDPs. Where non-composable CDPs offer a one-size-fits-all solution, The Composable CDP brings fresh air with their modular and flexible architecture with flexible SKUS and separable features that can be added as micro capability and when based on need or usage. This approach meets the unique and evolving needs of Fortune 5000 enterprises, allowing for a more dynamic interaction with customer data. By breaking free from the limitations of conventional systems, Composable CDPs empower companies to construct a customized tech stack that aligns perfectly with their marketing objectives. This bespoke configuration ensures greater alignment with business goals and fosters a more intuitive and responsive marketing ecosystem.

The edge provided by Composable CDPs in the competitive world of Fortune 5000 enterprises cannot be overstated. These platforms are not just tools but strategic assets offering unparalleled adaptability and efficiency. By embracing Composable CDPs, companies position themselves at the forefront of marketing technology, ready to capitalize on emerging trends and customer insights. They essentially are future-proofing their infrastructure for more nimble interchangeable tools when needed.

Below, I explore the benefits of adopting a Composable CDP. I list some brief benefits of enhanced customer insights and operational efficiency to scalability and long-term cost savings. While not a comprehensive list, I touch on light cost-benefit to highlight why Composable CDPs and various add-on services and tools vendors offer are the future of marketing technology for Fortune 5000 enterprises but also for their data teams to exploit the capital investments they have made with their current infrastructure.

II. Understanding Composable CDPs

Modular Architecture and Customization: The core of Composable CDPs lies in their modular architecture. This design principle allows businesses to handpick components like data processing tools, analytics modules, and customer engagement interfaces from different vendors and integrate them into a cohesive platform. Some vendors, like Adobe, offer these tools as separate data products and a huge comprehensive API library where you can build and host your own applications in a serverless environment. Census, a smaller player, has many composable capabilities but does not message their go-to-market as a CDP in the martech stack but supports the features and abstractions marketers need and want like their "Audience Segmentation" tool. They also use a more headless API approach like Adobe with their embedded system but seem to have a broader horizontal opportunity due to not being 100% tied to the CDP moniker and more agile in scope. This starkly contrasts Traditional non-composable CDPs, which typically offer a one-size-fits-all package mentioned above and with limited scope for customization. Whereas the Composable CDPs' modular nature ensures that businesses are not tied to a single vendor's ecosystem, allowing for seamless upgrading or swapping of components as technologies advance. Such flexibility is pivotal for enterprises aiming to stay agile and responsive in a fast-paced market.

Enhanced Flexibility and Responsiveness: Composable CDPs redefine flexibility in customer data management. They allow businesses to swiftly respond to new data sources, customer interaction channels, and analytics tools. For instance, if a new social media platform emerges as a key customer interaction point, a Composable CDP can easily integrate data from this new source without overhauling the entire system or throwing tons of staff resources into creating the integration. In contrast, Traditional CDPs might require extensive customization or even a system upgrade to accommodate such changes, leading to potential downtime and increased costs. Furthermore, Composable CDPs can quickly adapt to changing data privacy regulations, a critical consideration in today's data-driven business landscape.

Streamlined Integration and Future-Proofing: Composable CDPs facilitate streamlined integration with existing enterprise systems, such as CRM and ERP. this leverages the vast amounts of data and capital inside the enterprise vs just digital web/app data. This integration is often more fluid and less disruptive than Traditional CDPs, which may require extensive backend work to ensure compatibility. Additionally, the forward-thinking design of Composable CDPs positions them as future-proof solutions. As technologies evolve and customer behavior shifts, enterprises can adapt their Composable CDPs with minimal disruption, ensuring they are always at the cutting edge of customer data management and engagement strategies.

Composable CDP types

Composable CDP with Full data platform with componentized data apps approach: This CDP has a data platform that stages and stores data. e.g., it probably uses Databricks, Snowflake, AWS, Azure, and Apache as the underlying infrastructure or is natively built and hosted. It includes traditional data staging like a medallion datalake or lakehouse approach. Salesforce is a good example. This vendor approach may sell an entire enterprise data management infrastructure solution and have the ability to consume ERP, OMS, WMS, CRM, and marketing data. Eventually, this software company would buy or build these source systems to sell and manage. This approach was Oracle's original vision and didn't quite happen, and now you see the repeat in the Salesforce vision. Snowflake may move toward this vision in a few years because of the ARR flat-lining.

Composable CDP with partial data platform and componentized data apps: This CDP has a data platform that stores data but does not stage the data. This approach is the most common among nearly all CDP vendors. It can handle capabilities like enrichment, data value standardization, multiple datasets, and cataloging. It will potentially include tools and features like ID Resolution, analytics, ML tools, or journey tools, which are often premium add-on features - aka micro-data apps. These micro-composable features are often what sales teams call add-on or premium features internal to the platform but are the composable tools we see being discussed on social forums like LinkedIn, Medium and Substack. Adobe, mParticle, Amperity, and many others are good examples of this type of CDP. These vendors use Databricks, Apache, Snowflake, AWS, Azure, or all the above in a custom format as their underlying data infrastructure foundation - nearly identical to what the enterprise has deployed but abstracted into domain-specific UIs, easier to use for the marketer.

Non-Composable / Monolithic CDP: This CDP has the same data infrastructure as the above, but you cannot separate different capabilities as standalone tools, features, or data apps like in the above choices. The underlying infrastructure is typically custom, with hard-coded integrations into other systems like AWS or pure-play marketing sources. In many cases, they have opted not to directly integrate into the Databricks or Snowflake data platforms or systems due to cultural competitive perceptions. This limits them in many ways, reducing their ARR scale and deep integrations into the enterprise, subjugating them to web and app-only type data. Often, they make integration decisions to platform systems driven by their clients - therefore, everything they do is very custom and contrarian. These CDP platforms could revert and message composability but are resistant. Ironically, they are architecturally composable even though they do not go to market in this regard. These are old thinking vendors and not embracing modernization. They can be hard to identify, but you will understand the nuanced approach once you dig into their capability and pricing.

The last type I am calling the Agile Marketing Platform" (AMP) or the Headless CDP - you could even call it Agile Integration Manager (AIM). I am very excited about this category because it is not really a CDP in a traditional marketing sense. These tools typically fall into the reverse ETL, Zero-copy, or generic ETL genre - but often, they are much more and have not been given full opportunity to come into their own yet as they are still infants in the landscape. Give them more time, and they will eclipse the other categories for flexibility. Census (API docs), Hightouch, and Simon Data are all in this category. They connect directly to the sources, Enterprise Data infrastructure, and other data platforms. It utilizes the resources of these capitalized data infrastructure investments vs. investing in more of the same as you do with the above CDPs. You have much more flexibility in approaching the data pipeline, schema data models, sources, audience segment management, and destinations. Each API of the AMP/AIM solution is a composable feature. In a sense, it becomes a headless CDP and can be embedded anywhere. It also scales your partners to create new UI abstractions for any solution they need on top of this type of solution, which scales their salesforce horizontally. Embedded is the KEY with this category. Here is a list of some application examples this type of tool could support along with its base tools and UI. This is not an exhaustive list, just a few ideas as an embedded model.

• B2B Demand-Gen Engine
• AI Personalization Enabler powering a reinforcement engine
• Omnichannel Journey Mapper powering a live Real-time visualization
• Privacy Compliance Suite - enhances consent and governance
• Brand Sentiment Analyzer
• Predictive Budget Allocator - augments media mix and acquisition ROI
• Customer Lifetime Value Predictor
• LLM Command and Control Enabler
• Media Governance engine

You could derive many more ideas for augmented tools or add-ons that you could develop and create independently. The real strength of these vendors is that they can sell to ISV and agency partners who can use the headless AMP/AIM framework to develop the tools mentioned above, but the salesforce GTM of these vendors just expanded exponentially.

III. Benefits of Composable CDP

This section applies to all types except the non-composable or monolithic type.

Enhanced Customer Insights: One of the most significant benefits of Composable CDP is its ability to provide granular customer insights. By integrating and leveraging various data sources, including CRM, online behavior, and transactional data, Composable CDPs enable enterprises to build a comprehensive view of their customers. This 360-degree view enables personalized and targeted marketing strategies, improving customer engagement and conversion rates.

Improving Operational Efficiency: Composable CDPs offer a modular architecture, empowering marketing teams to optimize their operations efficiently. With the ability to choose and integrate best-of-breed components, enterprises can leverage the most advanced marketing technologies available. This approach minimizes redundant functionalities and streamlines operations, leading to cost savings and increased productivity.

Scalability and Flexibility: Composable CDPs provide unparalleled scalability and flexibility, enabling enterprises to adapt to changing business demands. Traditional CDPs often require significant investments and time-consuming processes to scale or accommodate new functionalities. In contrast, Composable CDPs allow businesses to add or modify components as needed, ensuring seamless growth and agility.

Cost Savings: While the initial investment in implementing a Composable CDP may appear higher than that of a Traditional CDP, the long-term cost savings outweigh the short-term costs. The flexibility of Composable CDPs enables enterprises to avoid vendor lock-ins and tailor their tech stack to their specific requirements, eliminating unnecessary expenses for redundant functionalities. Moreover, Composable CDPs promote operational efficiency, reducing the time and resources required for managing disparate systems.

IV. Costs and Risks of Composable CDP

Initial Investment: Implementing a Composable CDP could involve higher initial costs than a Traditional CDP - this is still debatable as many enterprises have struggled for years with their traditional CDP implementations, leveraging a fraction of the capabilities. So, these costs should be considered within the context of the long-term benefits and cost savings. While Traditional CDPs may seem more affordable upfront, they often require additional investments for customization or integrating third-party tools, and these brittle integrations and add-ons are using older, less modern methods. This can be costly long term due to a lack of skills or sheer time and maintenance required.

Integration Complexity: The flexibility offered by Composable CDPs can also present challenges in terms of integration complexity - this is getting much easier due to Co-pilots and low/no code UI. Since multiple components and technologies are involved, proper planning and expertise are required to ensure smooth integration with existing systems. However, this challenge can be mitigated with the appropriate implementation partner or an in-house team.

Learning Curve: Composable CDPs require a certain level of technical expertise to fully harness their capabilities. Training marketing teams to effectively utilize the modular architecture and leverage various components may take some time. This is no different with Traditional CDPs. Therefore, the potential gains in operational efficiency and enhanced customer insights outweigh the learning curve associated with Composable CDPs.

Vendor Selection: Choosing the right vendors for each component within the Composable CDP can be a critical decision. It is essential to thoroughly assess potential vendors' expertise, reliability, and compatibility with existing systems. However, once the right vendors are selected, the overall performance and benefits of the Composable CDP are significantly enhanced. Based on your answers to my questions, I would look closely at the newer AMP/AIM/Headless CDP category.

Conclusion

In conclusion, adopting a Composable CDP presents a wealth of benefits for a Fortune 5000 Enterprise compared to a Traditional non-composable CDP. The superiority of Composable CDPs lies in their ability to provide enhanced customer insights, improved operational efficiency, scalability, flexibility, and long-term cost savings. While initial costs, integration complexity, a learning curve, and vendor selection challenges are associated with Composable CDPs, these risks can be effectively managed with the right strategy and implementation team. Adopting Composable CDPs empowers enterprises to stay ahead of the competition, adapt to market demands, and drive growth by unleashing the power of agility. It's a future-proof approach.