It’s an age-old story: CDP implementations start with momentum, only to slip into endless scoping, months of integration work, shifting timelines, and no live use case in sight.
In a recent episode of Couch Confidentials by Martech Therapy, Zeotap CEO Elad Simon joined Matthew Niederberger to unpack why this happens, where CDP projects typically go off-track, and what actually leads to faster, more effective outcomes.
If you’re navigating your own CDP decisions, these five takeaways can shed some light on the underlying forces that often determine how CDP projects play out:
1. Start With Impact, Not Infrastructure
Throughout the discussion, a clear pattern emerged: CDP implementations succeed when they start with business outcomes, not with trying to perfect all the data foundations upfront.
Trying to “get all the data ready” often turns CDP rollouts into long, drawn-out implementations that quickly lose momentum. Instead:
- Define your top 20-50 use-cases
- Prioritise the first wave of use-cases to activate
- Reverse-engineer the process with the minimum data needed to get them live
This shifts the implementation from an engineering-heavy exercise to a business-led one, and in turn can bring timelines forward by months.
2. There’s No “Right” Deployment Model, And That’s the Whole Point
There’s no single “correct” approach to CDP deployments. It isn’t purely packaged, and it isn’t purely composable. The conversation highlighted a practical divide that many teams recognise:
- Organisations with strong engineering capabilities often lean toward Composable or zero-copy patterns
- Organisations with fewer engineering resources often look to Packaged solutions to reduce technical lift
But there are many additional levers to take into account when evaluating a deployment model, such as the ability to support key use cases, the resources required, or warehouse readiness.
3. CDPs Don’t Fix Silos. They Reveal Them
Matthew made a sharp analogy: a CDP is a mirror.
It reflects how your organisation actually works and whether teams are aligned or operating in silos.
If CRM, email, app, onsite personalisation, and paid media all sit in separate teams with different KPIs and no shared view of the customer, the CDP will simply make those disconnects more visible.
This is why many “CDP failures” are not strictly technology failures; they are change-management challenges that expose the lack of alignment across teams.
4. When the CDP Becomes the Bottleneck
The issue isn’t just complexity. It’s rigidity.
Suite CDPs require teams to rework their existing data to fit their fixed schema, creating bottlenecks and slowing down even the simplest use cases.
And that challenge compounds with a second reality: many enterprises operate across multiple clouds, multiple warehouses, or multiple activation tools, far beyond what any single Suite fully covers.
When a CDP is tightly tied to one ecosystem, connecting to external data or activating in non-suite tools becomes slow, fragile, or simply very, very hard.
5. AI Will Un-Commoditise the CDP Market Again
Looking ahead to 2026, the conversation centred on how AI is reshaping the CDP landscape.
Two areas stood out as the big leaps in advancing CDP performance:
- Natural-language “talk to your data” tools: Allowing marketers to explore segments and performance simply by asking the platform questions.
- AI-assisted data onboarding: Automatically mapping fields and reducing schema alignment pain to dramatically cut setup time from days to hours and minutes.
Put simply, AI is poised to become the defining factor in how quickly and effectively CDPs can deliver value.
Closing Thoughts
The conversation made one thing clear: CDPs succeed when they fit the organisation’s reality: its goals, its teams, its architecture, its use cases. Getting that fit right is what ultimately shapes outcomes.
