Episode 132 of The Data Business Podcast. Lucas and Luna explore why data teams are borrowing a practice from software engineering: versioning data sets like they version code. The episode anchors on a major retailer that discovered their quarterly revenue analysis couldn't be reproduced because no one tracked changes to the source data. Lucas explains how tools like lakeFS and DVC create snapshots of data at different points in time, enabling rollbacks, audits, and collaboration. Luna questions the overhead and whether this scales for petabyte-scale pipelines. They discuss real use cases: regulatory compliance in financial services, debugging data quality issues in advertising analytics, and enabling data scientists to experiment without fear of breaking production. The conversation also touches on the cultural shift required to treat data as a versioned asset. Near the end, Lucas briefly notes that the show stays ad-free because of listener support, and mentions buy me a coffee dot com slash fexingo. The episode closes with a reflection on how versioning builds trust in data products.

#DataVersioning #DataVersionControl #DataEngineering #DataGovernance #DataOps #ReproducibleAnalytics #lakeFS #DVC #DataLineage #DataQuality #DataMesh #DataTrust #AnalyticsEngineering #DataInfrastructure #BusinessAndTechnology #FexingoBusiness #BusinessPodcast #TheDataBusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

Most data teams focus on building pipelines faster, but few measure whether those pipelines actually deliver trustworthy data. This episode explores the rise of Data SLOs: Service Level Objectives for data quality, freshness, and completeness. We walk through a real case where a major financial services company implemented Data SLOs on their risk reporting pipeline, defining a 15-minute freshness threshold and 99.9% completeness SLO. Within a quarter, they cut data-related incidents by 40% and rebuilt stakeholder confidence. We discuss how to choose the right SLIs, set realistic targets, and avoid the common pitfall of over-engineering SLOs for non-critical data. If you're responsible for data products that drive decisions, this episode will change how you think about reliability.

#DataSLOs #DataReliability #DataGovernance #DataProducts #DataEngineering #FexingoBusiness #BusinessPodcast #DataAnalytics #DataTrust #ServiceLevelObjectives #DataQuality #DataObservability #DataLineage #DataCatalog #DataMesh #DataTesting #DataOperations #DataCulture

Keep every episode free: buymeacoffee.com/fexingo

In this episode of The Data Business Podcast, Lucas and Luna explore how data teams are shifting from reactive firefighting to proactive incident detection using data observability. They dive into the concept of data observability—monitoring data pipelines for freshness, volume, schema, and quality—and how tools like Monte Carlo and Bigeye are helping teams catch issues before they impact reports or machine learning models. Lucas shares a concrete example: how a fintech company reduced data incident detection time from four hours to under five minutes by implementing observability metrics. They also discuss the role of open-source projects like Great Expectations and Soda Core, the cost implications of observability tooling, and how teams are measuring return on investment. Along the way, they touch on the cultural shift from data quality as a post-hoc audit to a continuous practice. The episode wraps with a reflection on where observability fits in the broader data engineering landscape.

#DataObservability #ProactiveMonitoring #DataQuality #DataEngineering #MonteCarloData #Bigeye #GreatExpectations #SodaCore #DataPipeline #IncidentDetection #DataFreshness #SchemaChange #DataVolume #OpenSource #Business #Technology #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

In episode 129 of The Data Business Podcast, Lucas and Luna dive into column-level lineage — a granular approach to tracking data from source to dashboard. They walk through how a healthcare analytics team used column-level lineage to satisfy an auditor's request for proof that patient zip codes were never shared with a marketing vendor. Lucas explains the difference between table-level and column-level lineage, and why the latter matters for GDPR, HIPAA, and SOC 2 compliance. Luna questions whether the extra engineering overhead is worth it for smaller teams, and Lucas shares a specific case where tooling from a vendor like Atlan cut investigation time from three weeks to 90 minutes. They also touch on open-source options like OpenLineage and how column-level lineage fits into a broader data governance strategy. A concrete, practical episode for data engineers, analytics engineers, and compliance officers.

#DataLineage #ColumnLevelLineage #DataGovernance #Compliance #GDPR #HIPAA #SOC2 #DataEngineering #Analytics #OpenLineage #Atlan #DataObservability #DataQuality #BusinessAndTechnology #FexingoBusiness #BusinessPodcast #DataTeams #DataManagement

Keep every episode free: buymeacoffee.com/fexingo

Lucas and Luna explore why data teams are investing in semantic layers—a metadata abstraction that sits between raw data and business users. They break down how companies like Airbnb and Netflix use metrics stores to enforce consistent definitions (e.g., 'active user' means the same across departments), reduce redundant queries, and accelerate time-to-insight. The episode digs into a specific case: how a mid-market e-commerce firm cut its reporting backlog by 40% using a semantic layer built on dbt and LookML. Lucas explains the technical architecture—logical vs. physical models, aggregation tables, and caching—while Luna questions whether this adds another tool to the stack or truly simplifies it. They also touch on the cultural shift required: training analysts to think in terms of metrics rather than SQL joins. If today's episode helps you think about your own data pipeline, listener support keeps us ad-free at buy me a coffee dot com slash fexingo.

#SemanticLayer #MetricsStore #SelfServiceAnalytics #DataEngineering #dbt #LookML #Airbnb #Netflix #DataModeling #BusinessIntelligence #DataCulture #Analytics #DataInfrastructure #Business #Technology #FexingoBusiness #BusinessPodcast #DataPodcast

Keep every episode free: buymeacoffee.com/fexingo

In this episode, Lucas and Luna explore how data teams and their business counterparts are beginning to measure data literacy as a business metric — not just a training checkbox. They unpack a case from a mid-size fintech that tied data literacy scores to report accuracy, query efficiency, and decision latency. The conversation covers the framework used: pre- and post-assessments, observed behaviour change, and correlation with reduced time-to-insight. Lucas cites a 2025 MIT Sloan Management Review finding that high-data-literacy organizations are twice as likely to report revenue growth from data initiatives. Luna challenges the causality, and they debate whether measurement is worth the effort for teams with fewer than fifty data consumers. The episode also touches on the tension between certification-style testing and real-world fluency, and why measuring literacy might be harder for data teams than building pipelines — but arguably more important for long-term value.

#DataLiteracy #DataCulture #BusinessMetrics #DataTeam #DataLeadership #DataDriven #DecisionMaking #MIT #DataSkills #Benchmarking #Analytics #DataEducation #DataQuality #TimeToInsight #DataROI #BusinessAndTechnology #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo

Episode 126 of The Data Business Podcast explores why data teams are adopting data clean rooms — secure environments where multiple organizations can analyze shared data without exposing raw records. Lucas and Luna examine the regulatory drivers behind the trend, including GDPR, CCPA, and emerging AI governance rules. They walk through a concrete example: how a retail chain and a credit card network used a clean room to run a joint loyalty analysis without ever exchanging customer PII. The episode also covers the technical architecture — differential privacy, query logging, and audit trails — and the business trade-offs: more privacy compliance, but slower queries and higher operational overhead. Lucas and Luna discuss why clean rooms are becoming a standard tool for data teams that work with sensitive data across partners, especially in advertising, healthcare, and financial services. They also touch on the productization of clean rooms by major cloud providers and the debate about whether startups can compete with built-in platform features. A practical, no-hype conversation about balancing data utility with privacy obligations.

#DataCleanRoom #PrivacyCompliance #GDPR #CCPA #DifferentialPrivacy #DataPrivacy #DataSharing #SecureComputation #CloudData #DataEngineering #BusinessAndTechnology #DataInfrastructure #DataGovernance #QueryLogging #AuditTrail #AdvertisingAnalytics #RetailData #FexingoBusiness

Keep every episode free: buymeacoffee.com/fexingo

In episode 125 of The Data Business Podcast, Lucas and Luna explore how data teams are turning to metadata-driven automation to maintain data quality at scale. They dive into the concept of an active metadata platform — a system that not only catalogs your data assets but uses that metadata to automatically generate and run data quality checks. Lucas cites a real example from Uber, whose data quality team uses metadata about field-level usage and schema changes to dynamically create assertions that catch issues before they hit production. Luna challenges the idea, asking about the risk of false positives and the overhead of maintaining such a system. They discuss how companies like Netflix and Airbnb have built internal tools to monitor data quality through metadata, and how startups like Sifflet and Bigeye are bringing these capabilities to smaller teams. The conversation also covers the cultural shift needed — moving data teams from reactive firefighting to proactive quality engineering. If you're a data engineer, analytics engineer, or data platform lead, this episode gives a concrete framework for thinking about data quality as a system, not a checklist.

#MetadataDrivenAutomation #DataQuality #ActiveMetadataPlatform #DataEngineering #DataObservability #DataGovernance #Uber #Netflix #Airbnb #Sifflet #Bigeye #DataTesting #DataContracts #SchemaEvolution #DataPlatform #AnalyticsEngineering #Business #FexingoBusiness

Keep every episode free: buymeacoffee.com/fexingo

Episode 124 of The Data Business Podcast explores semantic caching, a technique that lets data teams reuse query results intelligently instead of recomputing them. Lucas and Luna discuss how companies like Snowflake and Databricks are enabling this approach, breaking down a real case where a mid-size e-commerce company slashed its cloud query bill by 62 percent while improving dashboard load speeds. They cover the technical basics — cache keys, invalidation policies, TTL strategies — and the business case: faster analytics, lower costs, happier stakeholders. If you run a data team or manage analytics spend, this episode gives you one concrete tactic to try this quarter.

#SemanticCaching #DataEngineering #QueryOptimization #CloudCosts #Snowflake #Databricks #Analytics #DataInfrastructure #CacheInvalidation #QueryPerformance #BusinessIntelligence #DataTeam #CostOptimization #TTL #DataArchitecture #FexingoBusiness #BusinessPodcast #DataBusiness

Keep every episode free: buymeacoffee.com/fexingo

Episode 123 of The Data Business Podcast explores event storming as a collaborative modeling technique for data teams. Lucas and Luna discuss how companies like a mid-sized European e-commerce retailer used event storming workshops to align business and engineering around a shared domain model, cutting data pipeline rework by 30% in six months. They cover the key roles (domain experts, data engineers, product managers), the three-phase process (chaotic discovery, timeline alignment, bounded context mapping), and why it's gaining traction in 2026 as data mesh and domain ownership spread. Practical tips: starting with a single business process, using sticky notes on a wall, and avoiding over-modeling. A concrete alternative to top-down schema design.

#EventStorming #DomainDrivenDesign #DataModeling #DataMesh #BusinessProcessModeling #CollaborativeModeling #DataEngineering #DomainDiscovery #BoundedContext #DataArchitecture #EcommerceData #DataStrategy #TechDebt #DataGovernance #BusinessTechnology #DataTeam #FexingoBusiness #BusinessPodcast

Keep every episode free: buymeacoffee.com/fexingo