Why Composable CDPs Are the Missing Piece in Modern Commerce

Over the years I’ve guided retailers and e-commerce players through migrations and transformations, and whether you sell to consumers or to businesses, you need to know who you are selling to. You need to understand your audience, shape your content and your products around their needs, and keep a channel open between what your brand promises and what your customers actually expect.

It still surprises me how many companies manage to operate successfully without really knowing their customer base. It almost makes me question my own pitch for CDPs. But the moment they begin to collect and unify this information, the moment they actually see a single picture of who they are dealing with, they usually realize how much they’ve been missing out on. It’s a wake-up call. Suddenly there is an awareness that opportunities were lost, that the potential lifetime value of their customers was reduced simply because they didn’t know enough about them. The other problem is fragmentation. Different departments each hold part of the puzzle. Analytics knows a piece, sales another. KPIs don’t align, metrics tell different stories, and no one can claim ownership of the whole.

This is why Customer Data Platforms are such a critical piece of a composable architecture. The rise of the composable CDP marks a shift away from bundled, monolithic solutions towards a model where data is unified at the data warehouse or data lake level and activated flexibly across the business. Traditional CDPs promise an all-in-one solution: they collect data, store it, resolve identities, build audiences and push activations. That may sound convenient, but in reality it locks you into a vendor’s ecosystem, limits your choices, and often leaves you paying for features you don’t need. In the worst cases, you become dependent on a technology that ages poorly and becomes more expensive to maintain over time.

The composable approach gives you something different: freedom. It lets you decide where to put your money and your resources, and it allows you to build incrementally, use case by use case. You can start small, with the data you already have, and prove value quickly rather than committing to a long, costly implementation before you see any results. Many of the newer solutions are warehouse-native, meaning they work directly on the data you already store in your data warehouse or data lake. That reduces redundancy, avoids unnecessary pipelines, and makes governance and compliance much easier because you keep control of your own data. It also reduces the risk of vendor lock-in.

When I talk to leaders about this shift, I get very different reactions depending on the role. Executives want ROI. They want proof. And while the exact calculation can be difficult, sometimes it is enough to hold up the mirror and point out that they don’t really know their customers. That often gets their attention. Product leaders are more interested in usability and speed. They want to give their teams tools that actually work, that save time and frustration, so that energy can go into building better products instead of fighting bad systems. Letting them experiment with an accelerator project or a proof of concept has always been a powerful way to make the potential visible. Engineers are usually the easiest to convince. Most of them want to work with modern toolchains and APIs that are robust and scalable. MACH principles give them exactly that, and composable CDPs fit naturally into that ecosystem.

At the moment everyone is talking about AI and automation. Every conference keynote seems to start there. But very few organizations have the data foundation to back it up. This is where I see the real urgency. If you want to build meaningful recommendations, if you want insights that matter for your business, then you need to start with the basics: understanding customer behavior and customer interest. You don’t get there by chasing the next hype. You get there by laying the groundwork. In my experience businesses often fall short on clean, connected data, available where and when it’s needed.

Work in commerce has always been about knowing the customer. The only difference is that today, we finally have the tools to do it at scale. The companies that take this seriously now, that make customer data a cornerstone of their composable architecture, will be the ones who not only talk about AI, but actually use it in ways that matter.