By all accounts, the CDP has seen an incredible run over the past 10 years. Since then, however, the technological landscape has shifted dramatically, with 3 notable trends emerging:

1. A move towards open source tooling due to the transparency and lack of vendor lock-in inherent to such technologies
2. The rise of the cloud data warehouse, lake, and lakehouse leading to the coalescence of the data stack around these solutions as the organisation-wide single source of truth
3. Increased scrutiny on data privacy given the many examples of Personal Identifiable Information misuse

CDP onboarding used to be seen as a painful but necessary step in the data journey, but increasingly the market has recognized the long-term technical debt that comes with this. The Modern Data Stack allows businesses to build a CDP themselves using a suite of best in class technologies: a “Composable CDP”

We see four challenges presented by the traditional CDP that are best remedied through a composable approach, which include:

• Use Case Gap Risk
• Vendor Lock-in and Future-proofing
• Data Privacy Regulatory Compliance
• Ownership of Key Data Assets

Risks and pains associated with traditional CDPs

1. Use Case Gap Risk

A key risk associated with using a traditional CDP is the potential for a Use Case gap. Traditional CDPs are excellent at their core focus - gathering data from customer facing applications, identity resolution, and then pushing that data to 100s of applications that rely on a unified view of 1) Who a customer is and 2) What they’re doing.

However, the needs and expectations of customers are changing over time towards the more demanding. Those businesses who can actually claim the mantle of being truly “customer-centric” offer a highly fluid experience on their digital applications.

Underpinning this experience are a suite of capabilities to understand and interact with the customer. Two obvious such capabilities are Data Science / ML (e.g. to understand the complex relationships between a customer’s preferences and the products that fulfil them) and real-time behavioural data pipelines (to trigger a set of immediate responses to what a customer is doing at multiple points in their journey).

Based on our conversations with customers, Traditional CDPs are failing on both fronts. The data they provide lacks the structure, richness, and cleanliness to power ML models. Equally, they’re not engineered to provide real-time data streams outside of their core use case and so tend to fall short in the areas of personalisation or recommendation type data applications as well.

As a result, CDPs end up becoming the limiting factor of the data stack - with customers having to work around their limitations versus innovating with new use cases.

This contrasts with a composable CDP, which is built with modularity and flexibility in mind, allowing businesses to mix and match different tools and services to create a customized solution that meets their specific needs. This approach not only reduces the risk of a use case gap but also ensures that the CDP can evolve and adapt over time as the needs of the business change.

2. Vendor Lock-in and Future-proofing

Traditional CDPs raise vendor lock-in concerns, leading to a "Stockholm Syndrome" effect. Despite accumulating engineering and data debt, and use case gaps, businesses fear untangling themselves from CDPs deeply integrated into their processes, as the relationship becomes overly dependent.

Adding to this dissatisfaction is a predatory pricing model that has a tendency to spike aggressively over time. What seemed like a slam dunk in year 1 quickly turns sour when the cost of using a CDP becomes widely decoupled from the value being provided.

This was less of an issue when there was a dearth of viable alternatives but now serves as a significant deterrent to exploring new best-in-class technologies that are emerging in the data sphere.

The philosophy of the Composable CDP, meanwhile, is designed with interoperability in mind, allowing businesses to easily switch out different components as needed. This simultaneously reduces the risk of vendor lock-in through ensuring that the CDP can adapt to new technologies and trends as they emerge.

Moreover, there is the option to build in additional “exit routes” by leveraging open source technologies where possible. This has the first advantage of transparency, since the inner workings of Open Source tech are fully exposed. It also means that a company can revert to the OS if they’re not seeing their required return on investment of the managed service. This is not an option when using a black-box, closed source solution.

Overall, vendor lock-in is a significant concern for businesses looking to build a future-proof data infrastructure. By choosing a composable CDP, businesses can ensure that they control the vendors in their data stack rather than those vendors controlling them.

3. Data Privacy Regulatory Compliance

A key consideration when selecting technology vendors, particularly those that will be storing and processing customer data, is compliance to Data Privacy regulations. The European GDPR has established the precedence with others following such as the US with the CCPA, COPPA, and HIPAA, India with its Digital Personal Data Protection Act, 2022, and Canada’s Consumer Privacy Protection Act to name a few.

The drive to comply with these regulations is two-fold. Firstly, the direct cost of an infringement can be substantial - for example, up to 4% of annual global turnover. A second more latent cost is reputational damage. High profile breaches such as the Cambridge Analytica scandal have led to an increased awareness and sensitivity at the consumer level to practices around the handling of personal data. Although hard to quantify, this cost should not be taken lightly.

Although the intention of these regulations is broadly the same - to protect the privacy and personal information of individuals in the face of massive technological change - the devil is in the detail. Each ruling will have its own nuances, underlining the need for any multinational business operating across these regulatory spheres to be adaptable to change.

A data strategy that emphasises flexibility is therefore a given to secure compliance. A second consideration is around ownership, which can be defined as retaining as much control of where and how data is processed as possible.

These two aspects highlight the drawbacks of a packaged CDP. Firstly, they are often designed to meet general compliance needs, which can be a problem for businesses that operate in highly regulated industries or jurisdictions with specific data protection requirements. Moreover, these regulations are fighting to keep up with technological developments and are by no means static. In the same vein as “Use Case Gaps”, using Packaged CDPs can make a business vulnerable to “Compliance Gaps” brought about by regulatory change. Secondly, in relation to ownership - given that customer data is sent off site, businesses relinquish control over its processing and as such cannot guarantee to their customers that it is being managed in accordance with their own policies.

Composable CDPs, meanwhile, offer businesses the ability to retain full ownership of their customer data infrastructure and hold a unique copy of their Customer 360 within their own data warehouse or lake. When, for example, a GDPR data subject rights request (such as for deletion or rectification) is received - there is no external dependency. Further, a Composable CDP provides flexibility, allowing businesses to tailor their data infrastructure to their specific compliance requirements by modularly selecting solutions. This reduces concentration risk, in the sense that if one component of the data stack cannot cope with a regulatory requirement, it can be substituted out in isolation rather than entailing the wholesale replacement that typically arises with an end to end solution.

4. Ownership of Key Data Assets

Outsourcing the ownership of customer data has a secondary impact, which is that you are effectively handing over the keys to the kingdom - a customer 360 is arguably a business’s single most valuable data asset. It forms the foundation of many mission critical use cases - Attribution and ad optimization, measurement of CLTV, and personalisation & recommendations all rely on accurate user identification. Allowing a third party “black-box” autonomy over this means that they determine the Identity Resolution logic - i.e. the rules to consolidate a number of user identifiers from different platforms into one user. Given that many CDPs are priced on Monthly Tracked Users (MTUs) an immediate cause for concern is that they’re effectively marking their own homework. Additionally, your view of the customer becomes confined to what is available to the CDP, meaning data collected from customer facing digital applications. This presents a huge blind spot consisting of everything else you know about that customer - demographic, transactional, and offline data sources being obvious examples.

There are immediate technical and cost-related implications of offloading the responsibility of your CDP to a vendor. However, we should also apply a long-term strategic lens, too. The needs that a business has from their customer data solution will naturally change over time as it matures and evolves. This will translate into new use cases and new logic to underpin identity resolution - needs that Packaged CDPs are poorly equipped to serve, given their inherent lack of flexibility arising from their focus on a single purpose.

The alternative is to build your CDP around a data warehouse or lake, which becomes your single source of truth for all things customer. This way, you retain full control over how a user is formulated, which can be adjusted over time as customers interact with your business in different ways and through different channels. Other data sources can be incorporated to enrich your Customer Behavioural Profile. For example, having a full view not only of how a customer engaged with your website but incorporating their full transactional history (through ETL’ing from your transactional database into the warehouse) to link specific types of engagement to propensity to convert. This then starts to look like more of a strategic data asset than a tactical data silo.

Conclusion

The emergence of the Composable CDP comes at a time when its packaged counterpart faces scrutiny for its limitations and drawbacks. The Composable CDP offers a modular and flexible approach that allows businesses to mix and match different tools and services to create a customised solution that meets their specific needs. 

Such a solution offers the following benefits:

• Reduces the likelihood of Use Case gaps whilst ensuring that the CDP can evolve and adapt over time as the needs of the business change 
• Promotes interoperability, allowing businesses to easily switch out different components of their CDP as needed, mitigating the risk of vendor lock-in
• Allows retained control and ownership of customer data and the Customer 360 - crucial to respond to regulatory pressures but also to build a future-proof strategic data asset