There’s a quiet truth emerging across many data teams — one that rarely makes it into strategy decks:
Most “data products” aren’t products at all.
They’re dashboards with better branding.
Pipelines with an owner.
Or governance artefacts dressed up in modern language.
And while the terminology has evolved… the outcomes often haven’t.
Beneath the surface, many organisations are discovering that despite significant investment, their data product initiatives are underdelivering — struggling to gain traction, adoption, or trust.
The question, then, isn’t whether data products are the right idea.
It’s why so many organisations are finding it so difficult to make them real.
The idea is right – but the execution is misunderstood.
At its core, the concept of a data product is compelling.
It promises a shift away from fragmented, poorly governed data assets towards something more intentional:
- Clearly owned
- Designed for specific consumers
- Measurable in terms of value
- Built to be reusable and scalable
In theory, it represents a long-overdue evolution in how organisations think about data. And yet, in practice, many organisations haven’t changed how they operate – only how they describe what they already do.
This is where the disconnect begins.
Adopting the language of data products without adopting the discipline behind them creates an illusion of progress, without the substance.
Why Data Products Fail: The 5 Disconnects
- Ownership Disconnect: No real decision rights
- Value Disconnect: Defined by data, not decisions
- Trust Disconnect: Governance applied after the fact
- Resilience Disconnect: Not designed for failure
- Operating Model Disconnect: Built outside control frameworks
Figure: Why Data Products Fail – The 5 Disconnects
01. Ownership Disconnect: Ownership in name, not in reality
One of the first places this becomes visible is in ownership.
Many organisations have introduced the role of “data product owner”. On paper, this suggests accountability. In practice, however, the role is often constrained.
These individuals rarely control:
- The roadmap
- The funding
- The prioritisation of trade-offs
Instead, they sit in a coordination layer — managing stakeholders, aligning delivery, and navigating competing demands.
The result is predictable.
Decisions are diluted.
Priorities shift constantly.
And accountability becomes collective rather than individual.
But a product without clear ownership is not a product — it is a shared responsibility problem.
And shared responsibility, in complex organisations, is often another way of saying no one is truly accountable.
02. Value Disconnect: Value defined by data, not by decisions
Even where ownership structures exist, another challenge quickly emerges: a lack of clarity around value.
Too often, data products are defined in terms of what they contain:
- datasets
- integrations
- pipelines
Rather than what they enable:
- decisions
- actions
- business outcomes
At first glance, this may seem like a subtle distinction. In reality, it is fundamental.
When products are defined by data, they tend to grow in scope. More sources are added. More use cases are accommodated. Complexity increases.
But adoption does not. Why?
Because consumers are not looking for data, they are looking for clarity. They are looking for confidence in the decisions they need to make.
If a data product cannot clearly articulate the decision it supports, then it is unlikely to be used consistently, no matter how technically sophisticated it may be.
03. Trust Disconnect: Trust is discussed, but rarely engineered
From here, the issue becomes one of trust.
Trust is frequently cited as a critical success factor in data strategies. But in many organisations, it remains aspirational rather than operational.
Governance is still applied after the fact.
Lineage is incomplete or inconsistent.
Quality is monitored, but not designed into the product itself.
This creates an all too familiar dynamic.
When something goes wrong:
- root cause analysis is slow or inconclusive
- explanations are partial or uncertain
- confidence erodes quickly
In highly regulated environments or those operating at scale, this is more than an inconvenience — it becomes a material risk.
A true data product should not rely on trust as an assumption. It should earn it.
It should be able to explain:
- where its data originates
- how it has been transformed
- what controls have been applied
Without that transparency, the product may function technically — but it will never achieve the level of trust required for critical decision-making.
04. Resilience Disconnect: Designed for success, but not for reality
Another pattern begins to emerge when data products are tested under real-world conditions.
Many are designed with an implicit assumption that systems will behave as expected.
But in practice:
- upstream data changes unexpectedly
- pipelines fail
- definitions drift over time
When this happens, organisations often lack:
- clear visibility of downstream impact
- defined ownership of incident resolution
- mechanisms to communicate issues to consumers
At that point, the limitations of the “product” become clear.
It behaves less like a resilient product and more like a fragile dependency.
Mature data products take a different approach.
They anticipate failure.
They define failure modes.
They make impact visible.
In doing so, they create confidence — not because nothing goes wrong, but because when it does, it is understood and managed.
05. Operating Model Disconnect: The deeper issue
Perhaps the most significant challenge, however, lies beneath all of this.
In many organisations, data products are being developed in isolation — while governance, security, and risk functions operate in parallel structures.
This separation creates a fundamental misalignment.
Data products evolve without embedded controls.
Governance frameworks exist without direct enforcement.
Lineage is attempted retrospectively, rather than designed upfront.
The result is a system where:
- products are delivered, but not fully trusted
- controls exist, but are difficult to operationalise
- accountability is diffused across multiple functions
The organisations beginning to move ahead are addressing this more holistically.
They are no longer treating data products as standalone artefacts.
Instead, they are positioning them as the execution layer of a broader data control plane, where governance, quality, and accountability are built directly into how data is created and consumed.
This pattern is reinforced by broader industry trends:
- According to Gartner, through 2026, over 80% of data and analytics initiatives will fail to scale due to governance and organisational challenges
- McKinsey highlights that data-driven organisations are 23x more likely to acquire customers, yet most struggle to operationalise data at scale
- MIT Sloan research shows that only a small percentage of organisations successfully translate data into measurable business outcomes
Interpretation:
The issue is not investment – it’s operationalisation and trust.
What success starts to look like
When organisations begin to close this gap, the characteristics of effective data products become clearer.
Not as abstract principles, but as observable practices:
- A single, accountable owner with genuine decision rights
- A definition grounded in business decisions and outcomes
- Embedded lineage, controls, and quality mechanisms
- Transparency that allows the product to explain itself end-to-end
- Designed resilience, with clear and visible failure modes
Perhaps most importantly, these products do not exist in isolation.
They are part of a wider system in which governance is not something applied to data after it is created, but something designed into the data lifecycle from the outset.
The real problem
In many ways, the challenges organisations face with data products are not new.
They reflect long-standing issues in data management:
- unclear ownership
- weak alignment to business value
- fragmented governance
- lack of accountability
What is new is the expectation that these issues can be solved through a different paradigm.
But a new paradigm requires more than new terminology.
It requires a shift in how organisations design, build, and take responsibility for data.
In reality, most organisations don’t have a data product problem.
They have:
- A design problem
- An accountability problem
- An operating model problem
Until that changes, the gap between ambition and reality will persist.
But when it does change — when ownership is real, value is clear, and trust is engineered , something more fundamental happens.
Data products stop being a concept.
They become the mechanism through which organisations actually trust, govern, and act on their data.
Data Products – That’s Quaylogic
At Quaylogic, we are seeing a clear shift in the organisations that are moving ahead.
They are:
- Embedding governance into delivery
- Aligning ownership with accountability
- Designing for trust from the outset
- Focusing on outcomes – not artefacts
We’re working with organisations to redesign how data products are owned, governed, and delivered.
👉 If this reflects challenges you’re seeing – let’s compare notes.
About the Author
Professor Geoff Smith, Partner – Quaylogic
For more than 30 years, Geoff Smith, has led commercially focused data privacy and empowerment functions, developing, operating, and transforming data systems and culture. He is a Visiting Professor at Loughborough University focusing on Digital Ethics and Trust Economies, Innovation, Leadership and Empowerment Technology.
Geoff is an advocate for sustainable development he is an ethical advisor on the UK Government National Digital Twin Program and a member of the working group for digital decarbonization.
