Data Governance & Data Quality Stewardship JudgmentData Governance & Data Quality Stewardship Judgment. The hard part was never the framework.
It is the table nobody can name an owner for, the headline number two teams define differently, and the urgent extract with a plausible reason and no stated purpose. A scored diagnostic of the judgment those moments actually need.
Not a certification quiz — a judgment test for the people who hold the data
Governance certifications and vendor academies test recall: name the six quality dimensions, list the pillars, match the term to the definition. None of that tells you what somebody will do when a model is a week from launch and nobody can name the basis on which its training table was collected. This assessment puts you inside 19 of those moments, all tool neutral — no product, no platform, no country's statute.
The Lineage Chain is a new scoring design. Every option carries a hidden provenance grade, from using data whose origin, owner and purpose nobody has established through to establishing all three and recording them where the next person will look, and an enablement weight running from blocking the business without cause, or handing data over with no control at all, to unblocking the request with the smallest control that makes it defensible. Both are scored separately from answer quality, so a blanket refuser and a blanket permitter each score badly for their own reason.
Keying follows published constructs rather than one house framework: data-quality dimensions and the fitness-for-use definition of quality, data-lineage and provenance research, the FAIR principles, data-product ownership, master data and the single-definition problem, purpose limitation and least privilege, dataset and model documentation practice, retention-schedule practice, and the AI-governance work on training-data provenance.
What you walk away with
Five links — source, ingest, model, consume, retire — drawn with the craft you earned at each, and the weakest link named. A chain is only as strong as that link.
How often you operated at each grade: unknown origin, on trust, partly checked, established and recorded.
Whether you unblock the business with the minimum control that holds, or sit at one of the two failure ends: blocking without cause, or releasing with no control at all.
How often you used data whose origin could have been established and was not, and where that puts you: Trusted Enabler, Gatekeeper, Open Tap or Absent Steward.
Provenance and lineage, access and privacy, quality and fitness for use, ownership and retention — with concrete actions for the two weakest.
Inside your report
Illustrative sample — your report is generated from your own responses.
Built for
- Data stewards, analytics engineers, BI leads and data product owners in enterprise data and AI teams
- Privacy, governance and risk partners who sign off on how data is used and shared
- Data and analytics leaders hiring, benchmarking or developing a team before AI systems consume their data
Find out which link of your data chain decides the rest
30 scored exercises · about 35 minutes · full bespoke report with your lineage chain, Provenance Index, Enablement Index, Silent Reuse Rate and Steward Posture.
₹1,199 (incl. GST) · assessment and full report, nothing further to pay
Frequently asked questions
Four competencies: provenance and lineage, access and sharing and privacy, quality and fitness for use, and ownership and retention and change. It measures what someone would actually do in 19 realistic situations — an urgent extract with no stated purpose, a model about to train on a table whose collection basis nobody can name, a metric that drifted after a correct fix — rather than whether they can recall a framework.
Every option carries a graded quality score keyed to published constructs, plus two hidden weights. A provenance grade from 0 to 3 records whether the choice used data nobody had established, took it on trust, checked one thing, or established origin, owner and purpose and wrote them down. An enablement weight from minus two to plus two records whether it blocked the business without cause, released data with no control, or unblocked the request with the smallest control that makes it defensible. Those produce your Provenance Index, Enablement Index, Silent Reuse Rate and Steward Posture.
Data stewards, analytics engineers, BI leads, data product owners and their privacy and governance partners in enterprise data and AI teams, at mid to senior level. Every item is tool neutral: no named warehouse, catalogue or vendor product, and no country-specific statute, so it reads the same for a team in Bengaluru, Berlin or Boston.
About 35 minutes for 30 scored exercises. You get a full report drawn as a chain of five links from source to retire, with the weakest link named, your Provenance Index and its four-grade split, your Enablement Index, Silent Reuse Rate, Steward Posture, four banded competencies and concrete actions for the two weakest — downloadable as a colour PDF.
Rs 1199 in India (including GST) or US$11.99 elsewhere, one-time, for one full sitting and report. Data-governance certification courses run into the hundreds of dollars per seat and test framework recall; vendor academies are tied to a single tool; employer assessment suites are sold as annual subscriptions. Organisations can use AssessAll credits at 40 credits per person.
Each one takes a single capability, puts you inside the situations where it is actually tested, and scores your choices against published evidence — with a report designed for that capability alone, not a template. They span hiring, compliance, education, operations and personal skill.
Browse the catalogue →Methodology: Measures applied judgment in data governance and data quality stewardship through original situational items keyed to published constructs. Sources and constructs operationalised across the assessment: the data quality dimension frameworks used in practice and research (accuracy, completeness, timeliness, consistency, validity and uniqueness) and the fitness for use definition of quality from the total data quality management tradition (Wang & Strong, 1996; Wang, 1998), including the information product view of data; data lineage and provenance research on recording derivation and dependency (Buneman, Khanna & Tan on data provenance; the W3C provenance data model); the FAIR guiding principles for findable, accessible, interoperable and reusable data (Wilkinson et al., 2016); the data mesh and data product ownership literature on domain ownership and data as a product (Dehghani, 2020); master data management and the single business definition problem in enterprise reporting; privacy engineering principles of purpose limitation, data minimisation and least privilege (Saltzer & Schroeder, 1975; Cavoukian on privacy by design; the OECD privacy guidelines' purpose specification and use limitation principles); access control models and the separation of data ownership from data custody (Sandhu et al. on role based access control); dataset and model documentation practice (Gebru et al., Datasheets for Datasets, 2018; Mitchell et al., Model Cards, 2019); the reproducibility and versioning literature on reproducible analysis pipelines; records management and retention schedule practice as described in public records management standards; and published artificial intelligence governance work on training data provenance and documentation, including the transparency and data governance expectations set out in the NIST AI Risk Management Framework. All items are original works. No trademarked instrument, certification syllabus or vendor methodology is reproduced, and no affiliation with any framework owner, standards body or publisher is claimed or implied. This assessment gives skills feedback and is not legal, privacy compliance or regulatory advice.