Why Data Contracts Don't Solve Data Quality

Why Data Contracts Don't Solve Data Quality
Data contracts promise control, but not quality. I argue, "Data quality isn't about strict rules; it's about trust, monitoring, and adaptability." Rigid, engineering-led contracts slow innovation and create friction. The future belongs to domain-owned, AI-assisted governance, where context, collaboration, and continuous observability drive real business outcomes. Contracts, define expectations. Trust, adaptability, and AI deliver results.
The Problem
The Mirage of Contracts in Data Quality
In the rush to formalize and govern data, many enterprises turned to data contracts as the silver bullet for quality. The logic seemed simple: define strict rules, and quality will follow.
But as I've said before:
"Data quality isn't about strict rules; it's about trust, monitoring, and adaptability."
Contracts don't clean data, they merely document assumptions. If the underlying foundation is broken or inaccessible, no number of contracts will make it usable.
At best, contracts offer structure. At worst, they become bureaucracy layered on top of already strained systems, rules without context, rigidity without meaning.
Contracts: Tools Built by Engineers Without Context
Data contracts were created as an engineering solution to a business problem. They emerged when data was centralized under IT and stripped of its business meaning.
Engineers understood pipelines, not purpose. Contracts became a way to negotiate meaning, to dictate structure in environments where context had been lost.
The rules that underpinned this approach, such as the Bito(l) guidelines later formalized under the Linux Foundation, made sense in a pre-AI world. But those same rules now act as anchors dragging organisations backward.
They enforce control instead of clarity. They assume ownership lies with engineering, when in reality, the business domain is the only place where data truly gains meaning and trust.
The Industry's Contract Obsession, and Why It's Misplaced
Every few years, the data industry rebrands old ideas as new revolutions. From Data Warehouse to Data Fabric to Data Mesh, each cycle promises salvation. "Contracts" are simply the latest headline.
65%
Data-quality initiatives fail
Not because contracts are missing, but because of weak collaboration between data producers and consumers (Gartner, 2024)
26%
Actually trust their data
While 80% of executives call data a strategic asset, only 26% trust their data enough to make critical decisions (McKinsey, 2023)
The failure isn't one of rules, it's one of relationships.
The Control Fallacy of Data Engineering
Many engineers defend contracts because they represent control, something measurable, enforceable, and within their domain of comfort.
But control doesn't equal quality. It equals rigidity.
What engineers perceive as "structure," the business often experiences as "slowdown."
The Data Velocity Trap
Control mechanisms intended to ensure reliability actually reduce time-to-value by 50% (Gartner)
The more layers of validation, schema locks, and ticket queues you add, the further the business moves away from the insight. Gartner calls this the data velocity trap, the point where control mechanisms intended to ensure reliability actually reduce time-to-value by 50 %.
Business leaders don't wake up wanting contracts, they want answers, insight, and growth.
And that's the problem: data contracts optimise for technical governance, not business outcomes.
At Data Tiles, we optimise for the opposite.
Why the AI Era Makes Contracts Obsolete
The rise of AI and natural-language interfaces has made the old engineering-led model unnecessary.
Latttice enables domain owners to create, govern, and share their own data products directly, embedding context at the source.
In this world:
Engineers focus on platform scalability and performance.
Domain owners reclaim control over data quality and trust.
Contracts evolve into lightweight, contextual agreements baked into workflows, not bureaucratic artefacts.
As Reltio wrote in Business Insider (2024), "AI cannot paper over flawed data, and rushing into AI without foundational quality is a dead end."
They're right, but unlike traditional MDM tools that centralize control, Latttice empowers the business itself to define, monitor, and adapt trusted data products without contracts slowing it down.
The Hidden Agenda of the Old Guard
The Old Model
Control represents a way to stay relevant in a world where data was managed, not understood
The Shift
AI has shifted the balance from engineering-centric control to domain-centric empowerment
The Future
Business-owned, AI-enhanced, and adaptive, not contract-bound
It's no surprise some still cling to the contract model. For many, it represents control, a way to stay relevant in a world where data was managed, not understood.
But today, that control is redundant. AI has shifted the balance from engineering-centric control to domain-centric empowerment.
The business no longer waits on IT hand-offs or endless governance debates.
AI-ready enterprises remove those bottlenecks entirely, replacing gatekeeping with guided automation and contextual intelligence.
The future of governance is business-owned, AI-enhanced, and adaptive, not contract-bound.
When Contracts Fail in Practice
"The contracts exist in Git … but the reliability exists nowhere."
— SiffletData
Contracts capture intent, not reliability. Reliability lives in the domain, maintained through continuous monitoring, observability, and ownership.
According to Forrester (2024), organisations that embed real-time monitoring improve reliability by 43 %, proving that visibility beats enforcement every time.
The Real Pivot Point: From Paperwork to Practice
The question isn't whether contracts have value, it's whether they scale. As data ecosystems expand and AI accelerates change, static governance becomes a liability.
Rigid contracts can't keep up with fluid data flows. But adaptive, monitored, domain-owned frameworks evolve continuously, detecting anomalies, learning from usage, and improving automatically.
This is the new discipline: computational data governance, living, measurable, and automated.
Where Quality Really Lives: Domain Ownership
Quality lives with the people who understand the data's purpose.
When ownership sits with engineering, quality detaches from meaning. But when domain owners build and manage their own data products:
Trust is embedded from inception
Quality starts at the source, where context and meaning are clearest
Lineage and meaning stay transparent
The journey of data remains visible and understandable throughout its lifecycle
Usability is intuitive
The product was designed for those who use it, by those who understand it
McKinsey (2024) reports that organisations adopting cross-functional data ownership see double the adoption rate of governed data products compared to those with centralised control.
In other words, the business doesn't just need data that's compliant. It needs data that counts.
The Role of AI and Automation
AI has transformed how we achieve trust and quality. Instead of enforcing compliance through scripts, AI automates governance, feedback, and anomaly detection:
AI-based monitoring
Spots schema drift or data decay instantly
Plain-language policies
Replace code-based governance
Continuous learning
Ensures quality improves with use
Deloitte (2024) found that adaptive, AI-assisted governance delivers 30 % higher ROI through faster decisions and fewer manual interventions.
This is where Latttice leads, embedding AI and zero-code access so domain teams can focus on business outcomes, not technical bottlenecks.
Leadership
Enter the Data Catalyst
Transformation needs leadership. The Data Catalyst is that bridge between business and engineering, reframing data quality as a practice, not a policy.
"​The Data Catalyst ensures quality is not an afterthought, it's a living part of the data lifecycle."  Cameron, Data Tiles.
They champion trust, transparency, and measurable improvement. Where the engineer enforces, the Catalyst enables.
The Future: From Compliance to Clarity
Contracts can clarify expectations, but they cannot create quality. True data quality is domain-owned, strengthened by trust, monitoring, and adaptability, and amplified by AI.
As organisations adopt platforms like Latttice, they evolve beyond compliance checklists to living ecosystems where quality is not enforced but experienced.
Final Thoughts
Beyond Contracts, Toward Continuous Clarity
The era of control-driven data quality is ending. Data contracts defined the past. Outcome-driven, adaptive governance defines the future.
The loudest voices often defend control, but control doesn't drive growth. Outcomes do.
The future of data quality isn't written in contracts. It's built in collaboration, measured in results, and owned by those who live its context.
That's how you create data ecosystems that don't just store information, they drive performance, insight, and innovation.
Join a Data Conversation,
Cameron Price.
References
Gartner. Data and Analytics Trends Report 2024. — Finds that 65% of data-quality initiatives fail due to poor collaboration between data producers and consumers, not a lack of contracts.
McKinsey & Company. Global Data Transformation Report (2023–2024). — Reveals that while 80% of leaders call data a strategic asset, only 26% trust their own data to make key business decisions, underscoring the trust deficit in centralised models.
Deloitte. Data Governance and Agility in Modern Enterprises (2024). — Shows that adaptive, outcome-focused governance frameworks achieve 30% higher ROI and significantly faster time-to-insight compared to rule-heavy approaches.
Forrester. State of Data Observability and Automation (2024). — Reports a 43% improvement in data reliability for organisations implementing continuous monitoring and automated observability practices.
BARC Germany. Data Trust and Governance Study (2024). — Highlights that 72% of organisations experience governance misalignment and low adoption due to rigid, top-down control structures.
SiffletData. Why Data Contracts Don't Work. Sifflet Blog (2024). — Argues that "contracts exist in Git, but reliability exists nowhere," capturing the disconnect between documentation and actual data quality.
Reltio / Business Insider. Unleash AI's Potential: Why Transforming Enterprise Data Quality Is Key to Success (2024). — Emphasises that "AI cannot paper over flawed data," reinforcing the need for domain-led ownership rather than centralised engineering control.