Data quality problems usually get handed to IT, because IT runs the systems where the bad data shows up. But the definitions that make data right or wrong are set by the business, and that is where the fix has to start.
This piece traces where bad data starts, what treating it as plumbing costs, and what business ownership of data looks like in practice.
Image placeholder: Hero image, 1200 × 630. Devensa AI branded graphic for “Who Owns Data Quality? Putting the Business on the Hook”.
Data quality and data governance both rose sharply as named data integrity challenges between surveys, and 45% of organizations cite inconsistent data definitions and formats.1 Definitions are set by the business, which is why IT cleanup alone does not close the gap.
Data quality belongs to the business because the business creates, defines, and uses the data. IT stores and moves records. Sales, finance, and operations decide what “customer” or “active account” means. In research from Precisely and Drexel University’s LeBow College of Business, 64% of organizations named data quality their top data integrity challenge, and 45% cited inconsistent data definitions and formats.1
Where Bad Data Starts
Most data errors begin at the point of entry. A sales rep types a company name three ways. An intake form accepts “N/A” as a date of birth. An acquisition brings a second customer master with different field rules. A finance team counts revenue by invoice date while operations counts by ship date.
None of these is a technical failure. Each is a business decision made, or skipped, by someone outside IT. The rep chose speed over accuracy because nobody measured accuracy. The form designer allowed free text because no data owner set a rule. The two definitions of revenue exist because no executive ever chose one.
IT sees the results: duplicate records, failed integrations, reports which disagree. IT writes a deduplication script or a transformation rule, and the symptom fades. Next quarter the same reps enter the same records the same way, and the cleanup starts over.
The common misconception treats data quality as a cleanup project, a tool purchase, or a ticket for the data engineering team. Each of those fixes the output. None of them changes how the business defines and creates the data.
Data governance addresses the source. Governance assigns business ownership for each critical data domain, sets definitions everyone uses, and builds quality rules into the processes where data originates. In the Devensa vC-Suite model, this work belongs to the vCDO, the AI-powered virtual officer for data governance structure and data strategy modernization.
Three Costs of Treating Data Quality as Plumbing
Revenue
In May 2022, Unity Software told investors two problems carried an estimated $110 million impact on 2022 results. One was a platform fault reducing the accuracy of the Audience Pinpointer ad-targeting tool. The other: Unity lost the value of part of its training data, in part from ingesting bad data from a large customer.2 The CEO tied recovery to rebuilding data and retraining models ahead of any revenue recovery.2
AI projects
AI raises the cost of bad data because models learn from whatever they receive. Gartner expects organizations to abandon 60% of AI projects through 2026 when those projects lack AI-ready data.3 A Gartner survey found 63% of organizations either lack or are unsure whether they have the right data management practices for AI.3
Gartner’s February 2025 analysis drew on a survey of data management leaders and recommended organizations align data to specific AI use cases, set governance requirements with legal and business leaders, and monitor data quality continuously.3
Trust in the numbers
When three dashboards show three customer counts, executives stop trusting all three and fall back on spreadsheets and instinct. The same Precisely and Drexel LeBow research found 51% of organizations named data governance a top data integrity challenge, up from 27% the prior year. Organizations with governance programs reported improved data quality and improved analytics as leading benefits, each cited by 58%.1
Image placeholder: Supporting chart or diagram for this section. Devensa AI branded.
DAMA-DMBOK: Governance Sits Above Data Management
The DAMA Data Management Body of Knowledge defines data governance as the exercise of authority, control, and shared decision-making over the management of data assets. DAMA stresses an important distinction: governance governs data management, not the data itself.4
The distinction explains why IT cannot fix data quality alone. Data management, the pipelines and storage and integration, often lives in IT. Governance decides who defines “customer,” who approves a change to the definition, and who answers when the number is wrong. Those decisions need business authority.
Devensa AI scores Data Governance as one of six domains on the TruMaturity™ framework. The AI-mediated assessment produces a Data Governance Domain Maturity Score on the CMMI scale, 1 Initial, 2 Developing, 3 Defined, 4 Managed, 5 Optimized, alongside an Overall Effectiveness Score across all six domains. The tenant sets a target level. The gap becomes a prioritized roadmap with named owners, and reassessment shows whether ownership changed anything.
What Business Ownership of Data Looks Like
A business glossary. One written definition for each critical term, such as customer, active account, revenue, and employee, approved by the business owner and published where every analyst finds it.
Data owners and stewards. An executive owner for each data domain, accountable for quality, and a steward who handles daily rules and issues. Both sit in the business.
Quality rules at the point of entry. Required fields, valid formats, and picklists in the CRM and intake forms. Prevention costs less than cleanup.
Quality metrics in front of owners. Completeness, accuracy, and duplicate rates reported to the owning executive monthly, the same way sales sees pipeline.
An issue path. Anyone who finds a data problem knows where to report the problem, who triages the issue, and how fast a fix arrives.
Five Places to Start
- Pick the three metrics leadership argues about most often. Trace each one back to the source fields and the definition behind them.
- Name a business owner for customer data and for financial data first. Write the names down and publish them.
- Draft glossary definitions for the ten most-used business terms and get owner sign-off.
- Add validation rules to the two highest-volume data entry points, then measure the error rate again after 30 days.
- Before approving any AI project, ask the owning executive to confirm the training and grounding data meets a written quality standard.
How Devensa AI handles this
Devensa AI scores Data Governance on the TruMaturity™ scale and records a named business owner and steward for each critical data domain. A virtual Chief Data Officer (vCDO) on the platform tracks ownership, quality checks, and review cadence.
Frequently asked questions
Who is responsible for data quality?
Business data owners are accountable for the quality of data in their domain, supported by data stewards who manage daily rules and issues. IT and data engineering build and run the systems and pipelines. Responsibility follows the decisions: whoever defines and creates the data answers for its quality.
What is the difference between data governance and data management?
Data management covers the technical work of collecting, storing, integrating, and protecting data. Data governance sets the authority and decision rights over how data management happens. DAMA-DMBOK describes governance as governing data management, not the data itself.4
Why does AI make data quality more urgent?
Models learn patterns from the data they receive, including errors and gaps. Unity Software’s 2022 loss of training data value, tied in part to bad customer data, shows how fast a quality problem reaches revenue.2 Gartner expects many AI projects without AI-ready data to be abandoned.3
What is a data steward?
A data steward is a business-side role responsible for the day-to-day quality of a data domain: maintaining definitions, enforcing rules, and resolving issues. Stewards connect the executive data owner to the people who create and use the data. In most growing companies, the best stewards already exist: the operations analyst everyone asks when a report looks wrong.
How does Devensa AI measure data governance?
The TruMaturity™ assessment scores Data Governance on a CMMI-based scale alongside five other domains and converts gaps into owner-assigned roadmap tasks. Every AI-generated recommendation passes a human approval gate before anything is published or acted on. If your CFO and your head of sales each reported the number of active customers today, would the two numbers match?
See It Running on Your Own Program
Devensa AI scores all six governance domains on TruMaturity™ and keeps the evidence behind each score current. Early access is open now.
Join the early access list: devensa.ai/early-access
Sources
- Precisely and Drexel University LeBow College of Business, 2025 Outlook: Data Integrity Trends and Insights, press release. https://www.precisely.com/press-release/new-global-research-points-to-lack-of-data-quality-and-governance-as-major-obstacles-to-ai-readiness/
- Unity Software, Q1 2022 earnings call, May 10, 2022, transcript via The Motley Fool. https://www.fool.com/earnings/call-transcripts/2022/05/11/unity-software-inc-u-q1-2022-earnings-call-transcr/
- Gartner, “Lack of AI-Ready Data Puts AI Projects at Risk,” press release, February 26, 2025. https://www.gartner.com/en/newsroom/press-releases/2025-02-26-lack-of-ai-ready-data-puts-ai-projects-at-risk
- DAMA International, DAMA-DMBOK2 definition of data governance, as summarized by Data Crossroads, December 2024. https://datacrossroads.nl/2024/12/31/governance-vs-management-clarifying-the-divide