May 2025 Summaries

Reverse ETL is the process of moving enriched, business-ready data from analytical platforms back into operational systems, allowing teams to act on insights directly in their day-to-day tools. This approach closes the loop between analytics and operations, making data truly actionable by syncing it into operational systems such as CRMs and marketing platforms. Automated pipelines are essential for reverse ETL, addressing challenges like schema drift, performance, governance, and extensibility through off-the-shelf platforms that require minimal engineering effort. By unifying data from various sources and pushing insights back into business tools, reverse ETL enables end-to-end data activation, as seen in the success of Canva, which unified data from marketing, sales, and engagement platforms, enriched it, and pushed it into tools like Braze, resulting in improved email open rates, platform engagement, cost savings, and faster marketing segmentation.

With the rise of modern data stacks, companies can now build a single source of truth in their data warehouses. However, getting data out of the warehouse and into operational tools is historically challenging, causing business teams to hit a barrier where important data is stuck in the warehouse. Reverse ETL (Extract, Transform, Load) flips the traditional data pipeline by moving data from the warehouse into operational tools that business teams rely on. By doing so, insights can drive real-time actions across the business, transforming data from something referenced after the fact into something that actively informs decisions in the moment. This approach enables organizations to close the gap between analytics and execution, making data a real-time asset. Reverse ETL is no longer a nice-to-have but a strategic imperative, requiring speed, precision, and governance to execute at scale. The combination of Fivetran and Census provides a complete, enterprise-ready data movement platform that simplifies ingestion and activation with accurate, secure, and always-in-sync data.

Fivetran's SAP ERP on HANA on Hybrid Deployment for SAP offers a secure and controlled way to integrate sensitive data pipelines, allowing organizations to maintain maximum control over their data while leveraging Fivetran's easy-to-use platform. The solution enables businesses to centralize SAP data without sacrificing performance or visibility, and provides flexibility in terms of infrastructure options and scalability. With this deployment, companies can protect their most valuable data assets while ensuring they're available where and when needed, eliminating the need to choose between governance and innovation.

Global enterprises struggle with complexity, particularly when managing data across legacy systems, third-party apps, manual processes, and siloed data. A modernized data infrastructure that automates data integration, centralizes data, and builds the foundation for real-time insights is essential to overcome these challenges. Companies like HubSpot, Saks, DocuSign, and Oldcastle Infrastructure are leveraging Fivetran and Snowflake to simplify their data operations, improve decision-making, and unlock new opportunities for growth. These organizations have achieved significant results, including improved predictive metrics accuracy, reduced pipeline development time, increased ROI, and faster decision-making, by centralizing their data with Fivetran and Snowflake, automating data workflows, and gaining access to real-time insights critical for forecasting and planning. By adopting a modernized data infrastructure, enterprises can position themselves for competitive advantage and sustainable growth in today's data-driven economy.

Data teams grapple with unstructured data, making it crucial to enrich and transform information before business use, a task traditionally reliant on complex ETL processes or custom code. Fivetran Activations AI Columns for Databricks addresses this challenge by allowing teams to apply LLM-based transformations directly to Databricks tables, enhancing data quality and insights without complex coding. AI Columns integrates with AI models like ChatGPT, Claude, or Gemini, enabling users to create structured insights, enrich customer data, and automate data classification. This tool simplifies processes such as extracting themes from support tickets, personalizing marketing engagements, and normalizing job titles, which can then be synced to platforms like Salesforce or Google Ads. Setting up AI Columns involves connecting Fivetran Activations to Databricks, selecting data sources, and crafting prompts, offering an efficient way to manage and utilize data lakes for marketing, sales, and customer engagement.

Enterprises are struggling with AI implementation due to delays, underperformance, and failures, despite pouring billions into data centralization. The issue lies in the gap between strategy and execution, as AI requires fully centralized, governed, and ready data to deliver results. A new report highlights poor data readiness as a leading factor in undermining enterprise AI implementation, emphasizing the long-term need for AI-ready data. The consequences of poor AI execution extend beyond IT, negatively impacting business growth, operational costs, and customer satisfaction, with 38% of enterprises citing increased operational costs due to AI project failures. To achieve greater AI success, organizations must eliminate data silos, embrace automation, and invest in modern data integration tools that automate pipeline management.

Migrating to a data lake can be considered for centralizing large volumes of structured and unstructured data, enabling advanced analytics such as AI. A successful migration involves several key considerations including architectural decisions that ensure scalability, cost-effectiveness, and compatibility with the organization's data needs. Key architectural decisions include cloud storage options like AWS S3, Azure Data Lake Storage, or Google Cloud Storage, table formats such as Iceberg or Delta Lake, catalogs like AWS Glue or Unity Catalog, and query engines like Snowflake or Databricks. An interoperable approach is crucial to ensure flexibility in tailoring the architecture to specific needs of the team and use cases. Fivetran's data lake architecture is designed for interoperability, supporting both Iceberg and Delta formats, and allowing integration with third-party catalogs like AWS Glue or Fivetran's Iceberg REST Catalog. Once set up, a data lake can be populated with data using tools like Fivetran Managed Data Lake Service, which provides a straightforward setup guide and supports historical syncs on SaaS and database sources as well as syncing directly from data warehouses. Querying options for existing data stacks include flexible query integration patterns, an Extract, Load, Transform (ELT) approach, and dbt Core-compatible data models that can be used with supported query engines like BigQuery or Databricks. With modern tools and technologies, migrating to a data lake has never been easier, offering automation, flexibility, and interoperability.

Universities face pressure to enhance efficiency, improve student experiences, and make data-driven decisions in admissions and enrollment. Integrating and streamlining these workflows requires strong stakeholder alignment, clear objectives, and centralizing accurate, reliable data. Northeastern University's Director of Analytics Engineering, Ilia Xheblati, emphasizes the need for seamless data integration to eliminate inefficiencies, streamline decision-making, and create a more connected student-centric experience. To achieve this, universities must ask the right questions, such as how to centralize data and what problem they are solving. Without a unified source of truth, frustration builds across departments, leading to conflicting numbers and making it difficult to trust data. Fivetran's automated approach helps scale data movement, reducing setup time to minutes, ensuring scalability, and unlocking efficiency for organizations without dedicated data engineering teams. A pragmatic approach involves starting small with early adopters, clearly communicating benefits and aligning expectations across stakeholders, and articulating how tools will make tasks easier, faster, and more reliable.

Enterprises are rethinking their data architectures to meet new and unforeseen demands for data, as adopting traditional data warehouses has reached its limits. Prioritizing interoperability, ease of use, and scalability is crucial for improving data management and building a resilient enterprise data platform. The emerging data architecture design, the data ecosystem, promises to help enterprises achieve new efficiencies, streamline data access, and optimize analytics workflows by combining metadata analysis, data integration, semantics, knowledge graphs, and machine learning into a powerful framework. A successful data ecosystem should be interoperable, customizable, and scalable, with a governed data lake or lakehouse serving as its centerpiece. Solutions like Fivetran Managed Data Lake Service can support the data fabric by providing automated data movement, open-table formats, and cross-platform interoperability, enabling enterprises to create a dynamic, self-service data environment that is well-governed, optimized for real-time analytics, and adaptable to evolving architectures and multi-cloud strategies. By investing in scalable, flexible architectures, data leaders can future-proof their organizations and unlock the full potential of their data assets.

Fivetran is acquiring Census to expand its platform beyond ingestion and transformation, enabling customers to reliably move governed data into applications where decisions are made. The acquisition brings Fivetran's philosophy of ease and simplicity to the activation layer, while maintaining governance, automation, and reliability. Reverse ETL engine Census offers seamless handling of schema changes and integrates with modern data technologies, supporting single second latency for streaming use cases and near real-time syncs for others. Governance and extensibility are also key features, ensuring data security and compliance across various standards. This acquisition is a major step forward in Fivetran's mission to make access to data as simple, secure, and reliable as electricity.