Upgrade to Pro — share decks privately, control downloads, hide ads and more …

FABCON SQLCON Ignite Data Governance Insights t...

Sponsored · SiteGround - Reliable hosting with speed, security, and support you can count on.

FABCON SQLCON Ignite Data Governance Insights to Accelerate Development

Avatar for Karen Lopez

Karen Lopez PRO

March 19, 2026

More Decks by Karen Lopez

Other Decks in Technology

Transcript

  1. ATLANTA26 JOIN THE CONVERSATION #FABCONSQLCON26 Ship faster • Fewer errors

    • Higher data quality Karen Lopez, InfoAdvisors Microsoft MVP, Data Platforms Ignite Governance Insights Across Databases, Lakes, and Models
  2. Karen Lopez ADGP, Microsoft MVP - Data Platform Microsoft Certified

    Trainer Data management expert, space enthusiast, and #TeamData evangelist www.datamodel.com @datachick.bksy.social
  3. Who? Governance for people who ship: fewer surprises, less rework,

    and faster delivery, because you can find it fast, change it safely, access it cleanly, and trust it by default.” © InfoAdvisors 2026
  4. What Do You Hear About Data Governance? Governance = more

    meetings It’s built by people who don’t deliver Definitions are stupid; everyone disagrees Governance creates gatekeepers Governance is for compliance, not builders I’m a developer. I don’t have time for this Governance is a business thing, not an IT thing We are agile so we don’t do governance I’m a DBA. I don’t have time for this It’s documentation
  5. Data Governance is the primary input into AI Governance (c)

    InfoAdvisors 2026. All rights reserved
  6. Surprise changes Shadow tables (including spreadsheets) Lack of trust Constant

    fixes for bad data Unknown impact of changes Your Enemy Isn’t Tickets: It’s Rework
  7. The trials of data quality DATA DATA WAREHOUSE ANALYTICS DATA

    SCIENCE AI (c) InfoAdvisors 2026. All rights reserved
  8. Adventuroso Everything is working and running like your dashboards were

    having a good hair day (c) InfoAdvisors 2026. All rights reserved
  9. Adventuroso And then, the dashboard goes from amazing to broken

    (c) InfoAdvisors 2026. All rights reserved
  10. Adventuroso Because hope is not a strategy, but rerunning feels

    productive (c) InfoAdvisors 2026. All rights reserved
  11. Adventuroso It can’t be us, so let’s look at the

    data (c) InfoAdvisors 2026. All rights reserved
  12. Adventuroso Lost time, costly resources on both the data engineering

    and the business side (c) InfoAdvisors 2026. All rights reserved
  13. Can I find the right data fast? Can I change

    stuff without breaking things? Can I access it without drama? Can I trust it without a meeting? Can I use it without ROI? Is the data good enough? Where do I get help? (c) InfoAdvisors 2026. All rights reserved
  14. What data exists Where it lives How it is created

    and updated Who owns it and who uses it How it connects to other data (c) InfoAdvisors 2026. All rights reserved Data Catalog
  15. A catalog turns business and tech knowledge into shared knowledge

    (c) InfoAdvisors 2026. All rights reserved
  16. Glossaries for Good Stop guesswork Reduce rework Lineage debugging faster

    Frees up time they don’t want to spend explaining things Enables scale-up of teams Increased collaboration Faster requirements and implementations Drill down for the win (c) InfoAdvisors 2026. All rights reserved
  17. Bugs Requirements defects ~20–25% Heavily under-reported Design defects ~25–30% Requirements,

    architecture, design, code, documents, and bad fixes Coding defects ~30–35% Classic coding errors Test-creation defects ~6% Test cases
  18. Data Related Bugs Database Defects ~10% Data Quality Defects ~20-35%

    Invalid formats Inconsistent schemas Missing or outdated data Duplicates Mislabeled data Integration Defects Sources Destinations Transformation bugs
  19. ATLANTA26 JOIN THE CONVERSATION #FABCONSQLCON26 Data Inventories Wrong table Wrong

    column Recreating data Conflicting definitions Obsolete data Data consent Data masking Encryption
  20. ATLANTA26 JOIN THE CONVERSATION #FABCONSQLCON26 Data Classification Is it PII?

    Is it confidential? Does it need special controls? Will I know how to figure that out? Are classification policies enforceable? How do I know what to mask, encrypt, or restrict? What if I have data lifecycle requirements?
  21. ATLANTA26 JOIN THE CONVERSATION #FABCONSQLCON26 Fast Access Controls Auditable Understanding

    before access Shifts decisions from tech to business Standardizes the process for faster access More compliant
  22. ATLANTA26 JOIN THE CONVERSATION #FABCONSQLCON26 Data Quality Measures Explicitly defined

    Standardized measures Identify problematic data Shifts data quality ownership Better decision-making Enables feedback loops Transparency
  23. Can I Change This Without Hating Friday? Silent schema changes

    Surprise dashboard failures Fear-driven release delays On-call chaos (c) InfoAdvisors 2026. All rights reserved
  24. Data Lineage Shows how data moves and transforms across systems,

    from source to consumption. (c) InfoAdvisors 2026. All rights reserved
  25. What Would You Do If They Didn’t Check Lineage? (c)

    InfoAdvisors 2026. All rights reserved
  26. (c) InfoAdvisors 2026. All rights reserved Getting there – Making

    it Safer, Faster, and Less Annoying 1. Inventory your data, prioritizing and triaging Current Projects? Most painful ones? Compliance? Critical data? 2 . Set Stewards, Owners, and Experts first Before rules, policies, and glossaries
  27. ATLANTA26 JOIN THE CONVERSATION #FABCONSQLCON26 Getting there – Making it

    Safer, Faster, and Less Annoying 3. Target useful over complete metadata •Owners/Stewards •Detailed scans •Don’t skip glossaries 4. Do analytics on your governing, but start with it just as a baseline 5. Plan to automate much of this going forward, but not now (c) InfoAdvisors 2026. All rights reserved
  28. (c) InfoAdvisors 2026. All rights reserved Getting there – Making

    it Safer, Faster, and Less Annoying 6. Get metadata into the hands, minds, and keyboards of engineering and development. Plus everyone else. Visibility failures damage trust, a lot Reduce guesswork Reduce rework Not just catalogs and glossaries Get metadata to where they work Discovery before creation
  29. ATLANTA26 JOIN THE CONVERSATION #FABCONSQLCON26 Getting there – Making it

    Safer, Faster, and Less Annoying 7. Overcommunicate Data Quality as just another measure, not a shame game Information about data, not blockages Not all data is the same, so there is flexibility on quality pillars Not all quality pillars need to be applied to all data Triple-check your quality rules (c) InfoAdvisors 2026. All rights reserved
  30. ATLANTA26 JOIN THE CONVERSATION #FABCONSQLCON26 Getting there – Making it

    Safer, Faster, and Less Annoying 8. Prioritize lineage as a rework reduction, not just documentation Scan first Integrate/upload Manually build it Will this break something? Is it Friday? What other things depend on this table? (c) InfoAdvisors 2026. All rights reserved
  31. All this is why bad governance gets ignored, and good

    governance gets used (c) InfoAdvisors 2026. All rights reserved
  32. Key Takeaways Governance is an Engineering Function Enterprise Tools for

    Enterprise Actions Managing Data Complexity Mastering Multi- Platform Complexity Governance at Scale Guessing is bad Productivity Managing Risk Importance of Laziness © InfoAdvisors 2026
  33. More Real-life Stories Find me and tell me more (c)

    InfoAdvisors 2026. All rights reserved
  34. Data Governance in AI and ML © InfoAdvisors 2026 (c)

    InfoAdvisors 2026. All rights reserved
  35. Karen Lopez Data Evangelist InfoAdvisors, Inc. @datachick linkedin.com/in/karenlopez www.datamodel.com Thank

    you, ad astra! Credit: Karen Lopez (c) InfoAdvisors 2026. All rights reserved