Data ownership matters more than a data platform
Why companies buy expensive data platforms and end up with the same problems - and what needs to be resolved before choosing a tool.
The conversation about a "data platform" in a company usually starts with pain. Data is scattered across systems, analytics is slow, every report requires manual work. The logical answer is to buy or build a single platform that brings everything together.
I have seen enough of these projects to notice a common pattern. The platform arrives. A year later the pain is back in new packaging. Data is still scattered, only now it is scattered in the platform too.
The reason is almost always the same: the technology was purchased, but the question of data ownership was never answered.
What data ownership actually is
Data ownership is not a legal term and not a box on an org chart. It is an answer to practical questions.
Who is responsible for order data being correct? Who notices if an error appears there? Who decides when the structure changes? Who answers the question "why is the number exactly this value"?
If each of those questions has a specific person or team as the answer - you have ownership. If the answer is "it lives in the system" or "probably the analysts know" - ownership does not exist.
Why a platform does not work without this
A data platform is a tool. It moves data, stores it, transforms it. But it does not make decisions about what the data means, where it comes from, or who is responsible for its quality.
If before the platform, sales data lived in three places and nobody knew which version was correct - after the platform it will live in four places, and the correct version is still undefined. You just have one more system now.
Data without an owner is unclaimed territory. Technology will not claim it.
What minimal ownership structure looks like
For each significant data domain, a few things need to be defined.
First: who produces this data? Which system, which process, which team is the source of truth?
Second: who consumes this data and for what? This helps understand how critical errors and delays actually are.
Third: who is the owner - the person responsible for quality and availability? Not "who technically stores it" but who makes decisions.
Fourth: what does "correct data" mean for this domain? What are the validation rules, what should never appear?
This can be captured in a simple table. You do not need a million-dollar data catalog - you need clarity.
When to think about a platform
A platform makes sense when ownership is defined and working, but scale or complexity exceeds the capacity of manual management. Then the platform automates what already works - and works well.
Starting with a platform when ownership is undefined turns the platform into a place where the problem is stored, not solved.
Questions before choosing a tool
If you are facing a decision about a data platform:
- For each key data domain - who is the owner?
- Is there currently a shared understanding of what "correct data" means in each domain?
- How are data errors discovered today - and who fixes them?
- Is the problem technical - lacking a tool - or organisational - lacking clarity on who owns what?
- What specifically will change about how data is used after the platform is in place?
If the answers are unclear, start with answering these questions - not with choosing a platform. The platform comes later.