What Is the Purview Unified Catalog?
Microsoft Purview Unified Catalog is a centralised metadata management service that provides a holistic, organisation-wide view of your data assets — regardless of where they reside. It combines automated data discovery, intelligent classification, business glossary management, and end-to-end lineage tracking into a single, unified platform.
Core Capabilities
- Automated Scanning & Classification — Discovers and categorises data assets across hybrid and multi-cloud environments using built-in and custom classifiers
- Business Glossary — Provides a shared vocabulary that bridges technical metadata and business context
- Data Lineage — Visualises the full journey of data from source to consumption, enabling impact analysis and traceability
- Unified Search & Browse — Empowers data consumers to find trusted, well-described data assets through an intuitive self-service experience
Supported Data Sources
Purview Unified Catalog connects natively to a broad ecosystem of data sources, including:
- Azure Data Lake Storage Gen2, Azure SQL, Synapse Analytics, and Azure Cosmos DB
- Amazon S3, Redshift, and RDS
- On-premises SQL Server, SAP, and Oracle databases
- Power BI datasets, Teradata, and Hive Metastore
- Custom sources via the Apache Atlas API and REST connectors
As of 2024, Purview supports over 200 data source connectors, covering virtually every major data platform in enterprise use today.
Why Unified Catalog Matters
Organisations today face an exponentially growing data estate spread across on-premises systems, multiple cloud platforms, SaaS applications, and streaming pipelines. Without a unified approach to metadata management, data teams waste enormous effort searching for data, verifying its accuracy, and understanding its context. Purview directly addresses these challenges.
🔍 Data Discovery at Scale
Data engineers and analysts spend an estimated 30–40% of their time simply finding the right data. Purview's cataloguing and search capabilities reduce this friction dramatically by providing a single searchable inventory of all data assets, their descriptions, and their ownership — cutting discovery time from days to minutes.
🔒 Governance & Compliance
Regulatory frameworks like GDPR, CCPA, and industry-specific mandates require organisations to know where sensitive data lives and who has access to it. Purview's automated sensitive data classification and lineage tracking provide the audit-ready evidence needed for compliance reporting and data protection impact assessments.
✅ Data Trust & Quality
Data consumers need confidence that the data they use is accurate, up-to-date, and fit for purpose. By enabling data owners to publish descriptions, certify assets, and document known issues, Purview transforms the catalogue into a trust layer for the entire data platform — reducing the proliferation of rogue copies and shadow datasets.
🤝 Collaboration & Culture
Unified Catalog breaks down silos between IT, data engineering, and business stakeholders. When business users can contribute glossary terms and data owners can document assets in plain language, the result is a living, collaborative knowledge base that reflects both technical reality and business meaning simultaneously.
Understanding Your Starting Point
Assessing Your Current Data Estate and Readiness
A successful Purview rollout begins long before any configuration work. A structured readiness assessment ensures you understand the scope of your data estate, identify the highest-priority data domains, and surface any technical or organisational blockers before they become problems mid-implementation.
Technical Readiness Checklist
- Data Source Inventory — Have you catalogued all major data sources, including cloud, on-premises, and SaaS? Do you know which are business-critical?
- Azure Tenant & Subscriptions — Is your Azure environment structured with appropriate Management Groups, Subscriptions, and IAM policies in place?
- Network Connectivity — Have self-hosted integration runtimes been planned for on-premises or private-network sources? Are firewall rules and VNet configurations accounted for?
- Identity & Access — Are service principals and Managed Identities available for Purview to authenticate against data sources?
- Existing Metadata — Do you have any existing data dictionaries, catalogues, or classification schemes that should inform your Purview taxonomy?
Organisational Readiness Checklist
- Executive Sponsorship — Is there clear executive support for data governance? Without it, adoption stalls at the data steward level.
- Data Domain Ownership — Have data domains been defined, and are domain owners willing to take accountability for curation?
- Cross-Functional Buy-In — Are IT, data engineering, analytics, and business teams aligned on the value and their respective roles?
- Governance Policies — Do existing data policies need to be digitised into Purview, or does governance need to be designed from scratch?
Skipping this assessment is the single most common cause of failed or stalled catalogue implementations. Invest two to four weeks here before touching any configuration.
Our Implementation Approach
Phased Rollout Strategy and Guiding Principles
We advocate a structured, phased approach that progressively builds value — delivering early wins that demonstrate ROI while laying robust foundations for long-term governance at scale. Each phase has clear entry criteria, deliverables, and success measures before the next phase begins.
Phase 1
Foundation
Tenant setup, permissions, collection hierarchy
Phase 2
Onboarding
Source connections, scanning, classification
Phase 3
Enrichment
Glossary, ownership, lineage, descriptions
Phase 4
Activation
Search, discovery, self-service enablement
Guiding Principles
Start Narrow, Scale Fast
Begin with two to three high-value, well-understood data domains rather than attempting full coverage on day one. Early success builds momentum and organisational confidence.
Automate Where Possible
Leverage Purview's automated scanning and classification capabilities to reduce manual effort. Reserve human curation for context and business meaning that automation cannot infer.
Governance Follows Technology
Technology alone does not create governance. Pair every technical configuration decision with a corresponding policy, role, or workflow to make governance real and sustainable.
Iterate and Improve
Treat the catalogue as a living product. Plan regular curation sprints, governance reviews, and user feedback loops to continuously improve quality and relevance.
Phase 1 — Foundation
Tenant Configuration, Permissions, and Collection Hierarchy Design
The foundation phase establishes the structural and security bedrock upon which everything else is built. Mistakes made here — particularly in collection hierarchy design and permission assignments — are expensive to undo later and can undermine governance integrity across the entire catalogue. Allocate sufficient time and involve both IT security and data governance stakeholders in these decisions.
Purview Account Setup
- Provision the Microsoft Purview account in the appropriate Azure subscription and region, aligned to your data residency requirements
- Configure diagnostic settings and audit logging to a Log Analytics Workspace for operational visibility
- Enable Managed Identity and configure network access (public endpoint vs. private endpoint with VNet integration)
- Register Purview's Managed Identity with each data source's IAM policies to grant read-only scanning access
Collection Hierarchy Design
The collection hierarchy is the most consequential design decision in the foundation phase. Collections define how assets are organised, who can access them, and how permissions are inherited. A well-designed hierarchy should reflect your organisational structure — typically by business domain, geography, or data classification level.
- Root Collection — Represents the entire organisation; restrict root-level admin access to a small, trusted group
- Domain Collections — One per major data domain (e.g., Finance, HR, Marketing, Operations)
- Sub-Collections — Subdivide by source system, environment (prod/non-prod), or sensitivity tier where needed
Collections cannot be restructured without re-scanning assets. Design with a 3–5 year growth horizon in mind and validate the hierarchy with all domain owners before finalising.
Roles and Permissions Framework
Collection Admin
Manages the collection structure, assigns sub-roles. Typically the data governance office or IT governance lead.
Data Source Admin
Registers and manages data sources within a collection. Typically data platform engineers or source system owners.
Data Curator
Reads and writes to all metadata in the collection. Typically data stewards and governance analysts.
Data Reader
Read-only access to metadata. Typically data analysts, scientists, and business users consuming catalogue content.
Phase 2 — Data Source Onboarding
Connecting, Scanning, and Classifying Key Assets
With the foundation in place, Phase 2 focuses on bringing your data estate into the catalogue through structured source registration, scan configuration, and automated classification. This phase is where Purview starts delivering immediate, tangible value — transforming an empty catalogue into a searchable inventory of your organisation's most important data assets.
Source Registration
Register each data source within its appropriate collection. Provide the connection details, assign a Managed Identity or service principal for authentication, and document the source system owner. Prioritise your highest-business-value sources first — typically the platforms feeding your analytics and reporting workloads.
Scan Configuration
Configure scan rule sets to control what gets scanned, how deeply, and how frequently. Define incremental scan schedules to keep metadata fresh without overwhelming the source system. For large data lakes, use partition sampling to balance coverage with performance. Always run a test scan on a small subset before scheduling full scans.
Classification & Sensitivity Labels
Enable Purview's built-in system classification rules (covering PII, financial data, health records, and more) alongside any custom classifiers tailored to your business. Integrate Microsoft Information Protection sensitivity labels to ensure data assets are consistently labelled across the Microsoft 365 and Azure ecosystem. Review classification results and tune rules to reduce false positives.
Best Practice: Onboard data sources in waves, starting with the platforms your analysts use most frequently. This ensures the catalogue is useful from the earliest stages and encourages organic adoption before formal training is even delivered.
Phase 3 — Enrichment
Business Glossary, Data Owners, Descriptions, and Lineage Mapping
Automated scanning produces technical metadata — schemas, data types, classification labels. But it is the human-added context of Phase 3 that transforms raw metadata into trusted, decision-ready knowledge. Enrichment is where the catalogue becomes truly valuable, and it is the phase that requires the most cross-functional collaboration between data engineers, stewards, and business subject matter experts.
Business Glossary Development
A well-governed business glossary is the semantic backbone of your catalogue. It ensures that a term like "Customer" or "Revenue" has a single, agreed definition that all teams use consistently — eliminating the ambiguity that causes errors in reporting and analytics.
- Start with a core set of 50–100 enterprise-wide terms before expanding to domain-specific vocabularies
- Assign a Steward (accountable owner) and Expert (subject matter authority) to each term
- Define approval workflows for new term creation and modification to maintain quality over time
- Link glossary terms to catalogue assets, schema columns, and classification labels to create navigable semantic connections
Asset Enrichment Priorities
- Data Owner Assignment — Every critical asset should have a named data owner who is accountable for its accuracy, access governance, and lifecycle. This is a people process as much as a technical one.
- Asset Descriptions — Encourage data owners and curators to write plain-language descriptions for all high-value assets. Even two to three sentences dramatically improve discoverability.
- Custom Attributes — Use Purview's custom attribute groups to capture organisation-specific metadata such as data product tier, business criticality, regulatory scope, or SLA classification.
- Contacts & Experts — Link assets to named contacts so that data consumers can reach the right person when they have questions about fitness for purpose.
Data Lineage Mapping
Lineage is one of Purview's most powerful differentiators. For sources that support native lineage (Azure Data Factory, Synapse Pipelines, dbt, Apache Spark), lineage is captured automatically during scanning. For custom pipelines, use the Apache Atlas Lineage API to push lineage programmatically. Prioritise lineage mapping for your highest-risk reporting pipelines first — those feeding regulatory, financial, or board-level dashboards — where the ability to trace data back to its source has the greatest compliance and trust value.
Phase 4 — Activation
Enabling Search, Browse, and Self-Service Discovery for End Users
Activation is the moment the catalogue goes live for its intended audience — the data consumers, analysts, scientists, and business users who will benefit most from it. This phase is not just a technical go-live; it is a change management exercise that requires structured communication, training, and feedback collection to drive genuine adoption.
Configure Search Experience
Ensure assets are fully indexed by verifying scan completion and enrichment coverage. Configure browse experiences by domain or collection so users can explore the catalogue intuitively even without a specific search query. Test search relevance with real user queries collected during the readiness phase.
Onboard Data Reader Roles
Provision Data Reader access for all intended catalogue consumers. Consider a phased user rollout — starting with data champions in each domain who can provide early feedback and act as internal advocates. Integrate Purview catalogue access into your existing data portal or intranet where possible to reduce friction.
Deliver Targeted Training
Run role-specific training sessions: a brief orientation for business users focused on search and browse, and a deeper session for data stewards covering curation workflows, glossary contribution, and asset certification. Provide short video walkthroughs and reference guides for async consumption. Avoid one-size-fits-all training.
Launch Communication Campaign
Announce the catalogue launch through executive communications, data team newsletters, and team-level briefings. Highlight specific high-value use cases — "You can now find all certified customer datasets in one place" — rather than leading with features. Make the value proposition concrete and relevant to each audience segment.
Establish Feedback Loops
Create channels for users to report missing assets, incorrect metadata, or confusing descriptions. Establish a regular cadence — monthly at minimum — for the governance team to review feedback and action improvements. User feedback is the primary driver of catalogue quality improvement post-launch.
Governance and Stewardship
Roles, Responsibilities, and Ongoing Curation Workflows
A catalogue without active governance quickly becomes stale, inaccurate, and ultimately abandoned. Sustainable governance requires clearly defined roles, documented responsibilities, and lightweight but consistent curation workflows that fit into people's existing working rhythms rather than adding excessive overhead.
Key Governance Roles
- Chief Data Officer / Data Governance Lead — Sets strategy, resolves escalations, sponsors the programme, and ensures executive accountability for data governance outcomes
- Data Domain Owners — Business leaders accountable for the quality, access governance, and lifecycle of data within their domain. Not necessarily technical.
- Data Stewards — Hands-on curators who manage day-to-day metadata quality: writing descriptions, managing glossary terms, assigning ownership, and reviewing scan results
- Data Source Admins — Technical owners responsible for source registration, scan configuration, and integration runtime management
- Data Consumers — Analysts, scientists, and business users who use the catalogue. Their feedback drives continuous improvement.
Ongoing Curation Workflows
Governance workflows should be lightweight, structured, and integrated into existing tools where possible (e.g., Microsoft Teams notifications, Purview workflow approvals).
- Weekly — Stewards review new assets surfaced by automated scans and assign initial ownership and descriptions for critical assets
- Monthly — Domain owners review asset certification status, approve or reject new glossary terms, and review user feedback from the past month
- Quarterly — Governance committee reviews catalogue health metrics, adjusts scan schedules, expands to new data sources, and refines classification rules based on operational experience
- Annually — Full governance policy review, collection hierarchy reassessment, and strategic roadmap update aligned to evolving organisational data strategy
Organisations that embed curation into existing sprint or agile ceremonies — rather than treating it as a separate governance activity — report significantly higher stewardship participation rates.
Common Pitfalls and How to Avoid Them
Lessons from Real-World Deployments
Across numerous Purview Unified Catalog implementations, the same failure patterns emerge repeatedly. Understanding these pitfalls in advance allows your team to design mitigations into the programme from the outset — rather than discovering them painfully mid-rollout.
⚠️ Trying to Catalogue Everything at Once
What happens: Teams attempt to register all data sources simultaneously, resulting in an overwhelming backlog of unclassified, uncurated assets that no one has time to enrich. The catalogue fills with low-quality metadata and users quickly lose faith in it.
Mitigation: Start with three to five high-value data domains. Achieve high-quality coverage there before expanding scope.
⚠️ Neglecting the Collection Hierarchy
What happens: A flat or poorly designed collection structure makes permissions management unworkable at scale, and re-architecture requires re-scanning all assets — a costly and disruptive exercise.
Mitigation: Spend adequate time in Phase 1 designing the hierarchy with a multi-year view. Validate it with all domain stakeholders before provisioning.
⚠️ Treating Governance as a Technology Problem
What happens: Technical teams configure Purview perfectly but no one takes ownership of curation. Descriptions are empty, ownership is unassigned, and the catalogue becomes a technical curiosity rather than a business tool.
Mitigation: Governance roles must be assigned, trained, and held accountable before go-live. Executive sponsorship is non-negotiable.
⚠️ Underestimating Self-Hosted Integration Runtime Complexity
What happens: On-premises and private-network data source connections require self-hosted integration runtimes (SHIRs) that depend on Windows VMs with specific network access. Firewall and proxy configurations are frequently underestimated, causing scan failures and delays.
Mitigation: Involve network and security teams early. Test SHIR connectivity for all target sources during Phase 1, not Phase 2.
⚠️ No User Adoption Strategy
What happens: The catalogue launches without a communication plan or training programme. Users are unaware it exists, or unsure how to use it, resulting in poor adoption and scepticism about the programme's value.
Mitigation: Invest in change management. Identify and empower domain data champions who advocate for the catalogue from within their teams.
Measuring Success
KPIs, Adoption Metrics, and Continuous Improvement Framework
What gets measured gets managed. Defining clear KPIs before launch establishes a shared baseline, enables objective progress tracking, and gives the programme team the evidence needed to demonstrate ROI and secure continued investment. Metrics should span technical coverage, curation quality, and end-user adoption.
Coverage & Quality KPIs
Asset Coverage
Target percentage of business-critical data assets registered and scanned within 6 months of go-live
Description Coverage
Percentage of priority assets with a non-empty, human-reviewed description and assigned data owner
Glossary Linkage
Percentage of high-value assets linked to at least one approved business glossary term
Adoption & Usage Metrics
- Monthly Active Users (MAU) — Track unique users accessing the Purview catalogue portal each month. Target 30% of data consumer population within 3 months, 70% within 12 months.
- Search-to-Discovery Rate — Percentage of searches that result in a user navigating to an asset. Low rates indicate poor metadata quality or relevance issues.
- Asset Certification Rate — Percentage of assets formally certified by a data owner. This is a leading indicator of data trust.
- Steward Activity Rate — Number of metadata edits, glossary updates, and ownership assignments per month. Tracks curation health over time.
- Time-to-Find Reduction — Survey data consumers quarterly on how long it takes to find a trusted data asset. Target a 50% reduction within 6 months.
Establish a quarterly Catalogue Health Review with representation from IT, data governance, and business stakeholders to review these metrics, identify trends, and agree on improvement actions for the next quarter.
Key Takeaways
Your Checklist for a Successful Purview Unified Catalog Rollout
A successful Purview Unified Catalog implementation is as much a programme management and change management exercise as it is a technical one. Use this consolidated checklist to ensure nothing critical is overlooked across your rollout journey.
✅ Foundation First
- Complete a formal readiness assessment before touching configuration
- Design your collection hierarchy for a 3–5 year horizon
- Assign and train governance roles before go-live
- Test SHIR connectivity for all on-premises sources in Phase 1
✅ Scope and Pace
- Start with 3–5 high-value data domains, not your entire estate
- Use a phased, iterative approach with clear exit criteria per phase
- Automate classification but curate manually for business context
- Prioritise lineage for high-risk, regulatory-facing pipelines first
✅ People and Process
- Secure executive sponsorship before programme kick-off
- Define stewardship workflows that integrate into existing practices
- Empower data champions within each domain to drive peer adoption
- Invest in role-specific training and a clear launch communication plan
✅ Continuous Improvement
- Define KPIs and baseline before go-live; review quarterly
- Create feedback channels and act on user input visibly and promptly
- Hold regular Catalogue Health Reviews with cross-functional stakeholders
- Treat the catalogue as a product — maintain a backlog and roadmap
The organisations that get the most value from Purview Unified Catalog are those that treat it not as a one-time implementation project, but as an ongoing data product — continuously curated, regularly governed, and deeply embedded in how their people find and trust data.
The organisations that get the most value from Purview Unified Catalog are those that treat it not as a one-time implementation project, but as an ongoing data product — continuously curated, regularly governed, and deeply embedded in how their people find and trust data.
Next Steps and Resources
Recommended Actions, Microsoft Documentation, and Support Contacts
With the implementation framework in hand, the following recommended actions will help your team move from planning to execution with clarity and confidence. Pair these with the rich resources available from Microsoft and the broader Purview community.
Immediate Next Steps (0–30 Days)
- Schedule and complete your formal data estate readiness assessment with IT and governance stakeholders
- Identify your implementation sponsor and constitute a Purview programme team spanning IT, data engineering, and data governance
- Select your first two to three data domains for Phase 1 and Phase 2 focus
- Provision a Purview development/test account to begin collection hierarchy prototyping and SHIR connectivity testing
- Initiate business glossary discovery workshops with domain subject matter experts
Key Microsoft Resources
- Microsoft Purview Documentation — learn.microsoft.com/purview — Comprehensive technical documentation covering all Purview services, connectors, and configuration options
- Purview Unified Catalog Deployment Guide — Microsoft's official best practice guide for enterprise catalogue deployments, covering collection design patterns and scan optimisation
- Microsoft Cloud Adoption Framework — Data Management — Strategic guidance for aligning Purview with your broader Azure landing zone and data platform architecture
- Purview Community & Tech Community Blog — techcommunity.microsoft.com — Active community with implementation case studies, connector updates, and governance pattern discussions
- Microsoft FastTrack for Azure — Eligible organisations can access FastTrack engineers for guided implementation assistance at no additional cost
Consider engaging a Microsoft Partner with Purview specialisation for your initial phases. Experienced partners significantly reduce time-to-value and help avoid the common pitfalls covered in this guide.
📧 Internal Data Governance Office
Your first point of contact for governance policy questions, domain ownership queries, and catalogue contribution workflows
🔧 Data Platform Engineering Team
For technical questions about source registration, scan configuration, SHIR setup, and lineage integration
🌐 Microsoft Support
Raise support tickets via the Azure Portal for Purview service issues, connector bugs, and feature questions