5 Myths of the Composable CDP

Chances are if you’re reading this blog, you’ve heard of the term “composability”.

Brought to the forefront of IT conversations by Gartner in 2020, it refers to an approach of building out a tech stack or architecture using open, interoperable, modular software.

While this desire for interoperability isn’t new (in fact, it was one of the core tenets on which Segment was founded back in 2011), it can also be seen as a response to an increasing demand for businesses to build their data infrastructure directly on the data warehouse, which for many businesses has become the single source of truth. 

It’s also instigated a fierce debate in the world of the customer data platform (CDP) world. CMOs, CTOs, and – well, anyone who's got a stake in customer data – are all weighing in.

Given the vociferous debate happening online, and the fact that composable means very different things to different people, it’s also led to some understandable confusion around the differences between CDPs and composable CDPs. Some of which is valid and some of which...is not. 

To help separate fact from fiction, we thought we’d break down the five most popular myths we’ve heard from our customers with regards to the composable CDP. 

Let’s get into it. 

Myth 1 - Composability is only developer/programmer centric 

The term composable is often used to reference modular, interchangeable parts of a system, and in more recent times has become a popular approach to software development as a way to maintain flexible and adaptable systems as opposed to monolithic ones. However, it is not exclusive to developers and programmers.

Brought to the forefront of IT conversations b, Scott Brinker at Chief Martech has correctly pointed out that “composability” has taken place for decades. Any martech stack worth its salt is “composed” of modular, integrated components, and is based on the principles of interconnection and integration. 

In other words, if you look under the hood at any modern marketing tech stack today, you’ll likely see composability already in action.

Ultimately, composability is a mindset to improve speed and agility across the business, not just an engineering best practice. 

Myth 2 - A composable CDP is just about architecture

Composability isn’t simply a trend, a methodology, or an architectural approach alone. What truly makes a CDP “composable” is the value it delivers to business users  – customer intelligence tied to an identity-resolved profile and enhanced with data relationships that extend beyond customer data. (e.g knowing detailed product information associated with a customer’s purchase.)

Without this business value, "composable architecture" becomes the end game, with developers breaking down every system to such a granular level you're stuck with a "build everything" approach.

A business uses the customer data from the CDP to do something better than it could if it didn’t have that customer data. The core principles of modularity, extensibility and API-accessibility make the CDP composable – it fits within the composable tech stack as a capability that the business can understand. 

It's important to recognize that architecture is merely a single facet of the broader concept of the composable enterprise.

Myth 3 - Businesses only want a CDP built on top of the data warehouse.

More and more, businesses (understandably) want access to the valuable insights generated and stored within their data warehouse. Which begs the question – if a business is already investing in a cloud data warehouse like Snowflake, Databricks, Google Cloud, and so on, can a CDP be built on top of it?

The stumbling block for this line of thinking is that it sees CDP functionality as add-ons to the CDW. In fact, both fulfill very different jobs within the organization today. 

Much of the data and customer intelligence being collected by a business day-to-day flows in real-time and doesn’t reside in the data warehouse (at least not quickly enough to be actionable for the business). Further, making that data actionable and accessible for business users, again in near real-time, is also not going to surface as a primary job for the data warehouse team to do.

Ultimately, few businesses will be able to meet their business requirements  operating exclusively from their warehouse data, and will be looking for a solution that leverages a data warehouse and capabilities that make real-time actionable and accessible to business users. 

As the below graph illustrates, both the data warehouse and CDP are all part of a broader overview of an organization’s data needs.

Myth 4 - Building a CDP from modular components is more cost-effective

At first glance, taking a DIY approach to building a composable CDP on top of your data warehouse might seem like the dream solution for any businesses looking to cut down on managing spend on multiple different data platforms.

But let’s slow down for a minute. What does it actually take to build and maintain a system like this? It's not as easy as you might think.

First, let's talk about engineering resources. Putting together a data stack from scratch – or even just integrating several tools properly – isn’t straightforward. It’s hard. And it demands a lot of time and expertise. Not every company has the resources to pull this off. 

Next, you have the complexity of engineering tasks and infrastructure management. Managing a collection of specialized tools can be a logistical nightmare. How much time and effort are you willing to sink to ensure everything runs smoothly? Just ask the company who set aside 9 months to build their data stack, only for it to take 3 years.

Then there's the issue of ongoing maintenance costs. CDPs often include maintenance in their pricing. Taking a DIY approach requires continuous upkeep, and lots of companies forget to factor this in when budget planning. At a conservative estimate, Forrester concludes that a minimum of 728 hours a year should be set aside to maintain an in-house system.

That brings us, of course, to the total cost of ownership for your CDP. When you add up all these hidden costs, building a CDPs from modular components turns out not to be as cost-effective as first thought.

Myth 5 – Anything not built on the data warehouse is monolithic and/or legacy

With any new technology paradigm that’s ushered in, from VisiCalc to CRM, and from CRM to CDP, we tend to unquestionably adopt the new with open arms, and label what went before as “legacy” and “monolithic”.

This tendency towards collective amnesia is particularly apparent with regards to discussions relating to the CDP, whereby CDPs have been reclassified in the conversation as monolithic entities that hinder agility and flexibility, and data warehouses as the only platform that supports a flexible, scalable, and modular approach.

This is something of a strawman argument when we consider how the MACH Alliance – a not-for-profit industry body that advocates for open and best-of-breed enterprise technology ecosystems – thinks about the topic of composability.

MACH stands for Microservices based, API-first, Cloud-native SaaS and Headless, and its mission is to protect MACH core principles and support buyers on standards, interoperability, and other important considerations when moving from legacy to a composable technology infrastructure. 

What do they not require to qualify for MACH status? That all your data is stored in a data warehouse. It can reside in a CDP, DMP, CRM, heck even a spreadsheet if that’s what works for your business (not that we’d recommend that)

It could and perhaps should be a critical part of your composable tech stack but not a requirement. That form of standard flies directly against the principles of composability in the first place.

While composable can be used in confusing ways, it should be clear that the CDP sits at the center of the composable experience/marketing stack. The CDP makes customer intelligence accessible via seamless interactions with the tech ecosystem that marketers and business users leverage to deliver on customer experience.

It can and should be interoperable with a cloud data warehouse as well as thousands of other marketing technologies. With a CDP, business users get their customer intelligence where they need it, when they need it, regardless of where it is stored, to deliver on the differentiated experiences that make them stand out.