Moving Legacy SQL Data Warehouses to Microsoft Fabric Using Azure Databricks: The Complete Migration Guide

Understanding the Migration Landscape

Migrating legacy SQL data warehouses represents one of the most significant infrastructure transformations an enterprise can undertake. Organisations across Australia and the Asia-Pacific region are increasingly recognising that traditional on-premises or legacy cloud data warehouse solutions no longer meet the demands of modern analytics, artificial intelligence, and real-time business intelligence requirements.

The shift from monolithic SQL data warehouses to cloud-native platforms is not merely a technology upgrade; it represents a fundamental rethinking of how organisations collect, process, analyse, and derive insights from their data assets. As highlighted in recent industry analysis, the shift from SQL data warehouses to Microsoft Fabric reflects broader enterprise trends toward unified analytics platforms that seamlessly integrate data engineering, analytics, and AI capabilities.

For CIOs, Heads of Data and Analytics, and technology leaders in mid-to-large Australian enterprises, understanding the migration pathway is essential. The journey from legacy systems to Microsoft Fabric, enhanced by Azure Databricks’ advanced analytics capabilities, requires careful planning, technical expertise, and a clear understanding of both the benefits and challenges involved.

This comprehensive guide provides Australian enterprises with a roadmap for successfully transitioning from legacy SQL data warehouses to the modern Microsoft data and AI platform. Whether you’re operating on-premises SQL Server data warehouses, legacy cloud solutions, or hybrid architectures, this guide addresses the strategic, technical, and operational considerations that will determine your migration success.

Why Migrate from Legacy SQL Data Warehouses

Legacy SQL data warehouses, whilst proven and stable, increasingly constrain modern organisations in several critical ways. Understanding these limitations is the first step in building a compelling business case for migration.

Operational and Financial Constraints

Traditional SQL data warehouses typically operate on fixed capacity models. Your infrastructure must be provisioned for peak load, meaning substantial compute resources sit idle during off-peak periods. This leads to inefficient cost structures where you pay for capacity you don’t consistently use. Microsoft Fabric addresses this through elastic scaling, where you provision resources based on actual demand, paying only for what you consume.

Maintenance overhead represents another significant cost driver. Legacy systems require dedicated database administrators, regular patching cycles, infrastructure management, and capacity planning efforts. These operational expenses accumulate rapidly, particularly in Australian organisations where skilled technical talent commands premium salaries.

Analytical and AI Limitations

Legacy SQL data warehouses were designed primarily for structured, batch-oriented analytics. They struggle with the demands of modern workloads including real-time streaming data, unstructured content analysis, machine learning model integration, and generative AI applications. As enterprises increasingly seek to leverage enterprise data migration trends and the rise of Microsoft Fabric, the competitive advantage goes to organisations that can rapidly integrate new data sources and analytical methods.

Azure Databricks, integrated with Microsoft Fabric, provides native support for complex analytics, machine learning pipelines, and AI model deployment. This capability becomes increasingly critical as organisations compete on their ability to extract intelligence from data.

Data Governance and Compliance Challenges

Australian enterprises operate under stringent regulatory requirements including the Privacy Act 1988 and increasingly, industry-specific regulations. Legacy systems often lack the sophisticated governance, lineage tracking, and access control capabilities required for modern compliance frameworks. Microsoft Purview, integrated within the Microsoft Fabric ecosystem, provides enterprise-grade data governance that helps organisations maintain compliance whilst enabling data democratisation.

Microsoft Fabric and Azure Databricks: The Modern Data Stack

Before embarking on migration, it’s essential to understand what Microsoft Fabric and Azure Databricks offer and how they complement each other in a unified analytics platform.

What is Microsoft Fabric?

Microsoft Fabric is a unified analytics platform that brings together data engineering, data warehouse, analytics, real-time analytics, and business intelligence capabilities into a single, integrated experience. Rather than managing disparate tools and platforms, Fabric provides a cohesive workspace where data professionals and business analysts collaborate within a single semantic model and governance framework.

Key capabilities include:

• Fabric Data Warehouse: A cloud-native data warehouse built on open standards, supporting SQL semantics whilst offering elastic scaling and near-infinite storage
• Lakehouse Architecture: A hybrid approach combining the structure of data warehouses with the flexibility of data lakes
• OneLake: A unified data lake that eliminates data silos and enables single-source-of-truth analytics
• Real-time Analytics: Streaming data ingestion and analysis capabilities for time-sensitive workloads
• Power BI Integration: Native integration with Power BI for seamless analytics and reporting

Azure Databricks Within the Fabric Ecosystem

Azure Databricks complements Fabric by providing advanced analytics, machine learning, and data science capabilities. Whilst Fabric excels at structured analytics and business intelligence, Azure Databricks specialises in:

• Complex data transformations and feature engineering
• Machine learning model development and deployment
• Advanced analytics using Apache Spark
• Python, R, and SQL-based analytical workloads
• Generative AI and LLM integration through OpenAI

The combination of Fabric and Azure Databricks creates a comprehensive platform where data flows seamlessly between systems, governance policies apply consistently, and teams can leverage the best tool for each analytical challenge.

Integration Benefits

When properly architected, Fabric and Azure Databricks work synergistically. Data ingested into Fabric can be accessed by Databricks for advanced analytics. Models and insights developed in Databricks can be published back to Fabric for broader consumption. This bidirectional flow, combined with unified governance through Microsoft Purview, creates an enterprise analytics platform that is both powerful and manageable.

At Agile Insights, we’ve designed Microsoft-certified accelerators that help organisations realise this integration efficiently, reducing implementation time and risk.

Pre-Migration Assessment and Planning

Successful migration begins long before any data moves. A thorough assessment phase establishes the foundation for all subsequent work.

Inventory and Documentation

Your first step is comprehensive documentation of your current state. This includes:

• Schema Documentation: Complete documentation of all tables, views, stored procedures, functions, and relationships
• Data Volume Analysis: Understanding the size of each table, growth rates, and projected future volumes
• Query Analysis: Identifying the most critical and frequently executed queries that will become your performance benchmarks
• Dependency Mapping: Understanding which applications, reports, and processes depend on your data warehouse
• Performance Baselines: Establishing current performance metrics for critical queries and processes

This inventory becomes your migration scope definition. It identifies what must move, what can be retired, and what might be refactored during the migration.

Business Requirements and Objectives

Migration is not purely a technical exercise. Clarify business objectives:

• What cost reductions do you expect to achieve?
• What new analytical capabilities will the migration enable?
• What is your timeline and tolerance for parallel running of old and new systems?
• Which business units and analytics use cases are highest priority?
• What governance and compliance requirements must the new system satisfy?

These business requirements inform technical decisions throughout the migration process. A migration driven by cost reduction may prioritise different architectural choices than one focused on enabling advanced analytics.

Technical Assessment

Conduct a detailed technical assessment of your current environment:

• Database Engine Version and Configuration: Identify any custom configurations, extensions, or version-specific features you rely upon
• Complexity Analysis: Some SQL constructs (particularly legacy T-SQL features) may require refactoring for Fabric compatibility
• Integration Landscape: Map all ETL tools, reporting tools, and applications that connect to your data warehouse
• Data Quality Assessment: Understand data quality issues in your current system that must be addressed or prevented in the new system
• Security and Access Control: Document current authentication, authorisation, and row-level security implementations

As noted in best practices for migrating legacy SQL databases to Azure, this technical assessment should identify potential blockers early, allowing you to plan mitigation strategies.

Estimating Effort and Timeline

Based on your inventory and assessment, develop realistic estimates:

• Data Migration Effort: Volume of data, complexity of transformations, and validation requirements
• Application Refactoring: Effort required to update applications that connect to the warehouse
• Testing Timeline: Comprehensive testing is essential; plan for adequate time
• Parallel Running Period: How long will you run old and new systems simultaneously?
• Contingency: Build in buffer time for unexpected challenges

For most mid-to-large enterprises, comprehensive migrations span 6-12 months, depending on complexity and scope. Underestimating timeline is a common pitfall that creates pressure to cut corners on testing and validation.

Data Architecture and Design Considerations

Migrating to Microsoft Fabric and Azure Databricks is an opportunity to rethink your data architecture, not simply replicate your legacy design in the cloud.

Dimensional Modelling vs. Lakehouse Approaches

Traditional SQL data warehouses typically employ dimensional modelling (star schemas, snowflake schemas). Fabric supports this approach but also enables more flexible lakehouse architectures that combine structured and unstructured data.

Consider your use cases:

• Dimensional Modelling: Best for well-defined, stable business processes with clear dimensions and facts. Excellent for traditional business intelligence and executive reporting.
• Lakehouse Architecture: Better suited for exploratory analytics, machine learning, and scenarios where data structures evolve frequently.
• Hybrid Approach: Many organisations use both, with a refined dimensional model for core business intelligence and a more flexible lakehouse for advanced analytics and AI.

Your migration should align with your analytical strategy. If you’re modernising analytics alongside infrastructure, this is the opportunity to adopt architectural patterns that better support your future analytical needs.

Scalability and Performance Design

Fabric and Azure Databricks handle scale differently than traditional warehouses:

• Separation of Compute and Storage: Unlike traditional warehouses, Fabric separates compute from storage. This allows independent scaling and cost optimisation.
• Partitioning Strategy: Effective partitioning becomes even more critical in cloud-native systems. Plan partition keys based on your query patterns.
• Indexing Approach: Cloud-native systems use different indexing strategies than traditional databases. Fabric and Databricks optimise differently; understand these differences.
• Caching and Materialisation: Determine what data should be cached, materialised, or computed on-demand based on access patterns.

Data Integration and Ingestion Design

Plan how data will flow into your new Fabric environment:

• Batch vs. Streaming: Will you maintain batch-oriented ETL, adopt real-time streaming, or use hybrid approaches?
• Data Source Connectors: Which connectors will you use to ingest data from operational systems?
• Transformation Approach: Will transformations occur during ingestion (ELT) or after (ETL)?
• Data Quality Checks: Build validation and quality checks into your ingestion pipelines.

Azure Databricks excels at complex transformations, whilst Fabric provides simpler ingestion and transformation through Power Query and Dataflows. Your architecture should leverage each platform’s strengths.

Step-by-Step Migration Process

With planning complete, the actual migration proceeds through distinct phases.

Phase 1: Detailed Design and Preparation

Before moving any data:

1. Create Detailed Technical Design Documents: Document the target architecture, data flows, transformation logic, and security model.
2. Develop Migration Scripts: Create scripts for schema creation, data movement, and validation.
3. Set Up Target Infrastructure: Provision Fabric workspaces, Azure Databricks clusters, and supporting services.
4. Configure Governance: Establish data governance policies, security roles, and access controls in Microsoft Purview.
5. Prepare Monitoring and Logging: Set up monitoring to track migration progress and identify issues.

Microsoft provides official guidance on using the Migration Assistant for Fabric Data Warehouse which can automate some aspects of schema migration, though complex migrations often require customisation.

Phase 2: Schema and Infrastructure Migration

Begin by establishing the target environment structure:

1. Migrate Schemas: Use the Fabric Migration Assistant or custom scripts to create table structures, views, and objects in Fabric Data Warehouse.
2. Handle Incompatibilities: Address any T-SQL constructs that aren’t compatible with Fabric SQL semantics. As discussed in SQL to Microsoft Fabric migration strategies, this often involves refactoring stored procedures and complex logic.
3. Create Lakehouse Structure: If adopting a lakehouse approach, establish the bronze, silver, and gold layer structure in OneLake.
4. Set Up Data Pipelines: Create initial data movement pipelines, starting with small test datasets.

Phase 3: Pilot Data Migration

Before migrating all data, run pilots:

1. Select Pilot Tables: Choose a representative sample of your tables, including small, medium, and large tables with various data types.
2. Migrate Pilot Data: Move pilot data using your migration approach (full load, incremental, or hybrid).
3. Validate Data Integrity: Compare source and target data, checking row counts, aggregations, and sample data matches.
4. Performance Testing: Execute your critical queries against the pilot data and compare performance with baselines.
5. Adjust and Optimise: Based on pilot results, refine your migration scripts, data models, and performance tuning.

Pilot migrations typically reveal issues that would otherwise emerge during full migration, allowing you to address them systematically.

Phase 4: Full Data Migration

Once pilots succeed, proceed with full migration:

1. Plan Migration Windows: Schedule migrations during low-activity periods to minimise impact.
2. Implement Change Data Capture: For large tables, implement CDC (Change Data Capture) to track changes during migration, enabling incremental updates.
3. Execute Full Load: Move all data from source to target using your optimised migration scripts.
4. Implement Incremental Updates: Set up ongoing processes to capture and apply changes until cutover.
5. Comprehensive Validation: Validate all migrated data comprehensively, checking not just row counts but data distributions, relationships, and business logic.

As noted in discussions on how migrating a SQL data warehouse into Fabric, many organisations find that dedicated migration tooling combined with custom validation scripts provides the most reliable approach.

Phase 5: Application and Query Migration

Simultaneously with data migration, prepare applications and queries:

1. Update Connection Strings: Modify applications to connect to Fabric Data Warehouse instead of legacy systems.
2. Refactor Queries: Update any queries that relied on legacy SQL features or syntax.
3. Test Application Functionality: Thoroughly test each application against the new Fabric environment.
4. Update Reports and Dashboards: Migrate Power BI reports and dashboards, leveraging Fabric’s native integration.
5. Validate Business Logic: Ensure that all business logic, calculations, and transformations produce equivalent results.

Phase 6: Parallel Running and Validation

Maintain both systems during a validation period:

1. Run Parallel Queries: Execute critical queries against both systems, comparing results.
2. Monitor Application Performance: Ensure applications perform acceptably against Fabric.
3. Identify and Resolve Discrepancies: Address any differences between old and new systems.
4. Gather Stakeholder Feedback: Engage business users to validate that data and insights match their expectations.
5. Plan Cutover: Once confidence is high, schedule the cutover to retire the legacy system.

Phase 7: Cutover and Retirement

Transition to the new system:

1. Final Validation: Perform final checks before cutover.
2. Execute Cutover: Switch applications and users to the new Fabric environment.
3. Monitor Closely: Maintain heightened monitoring immediately post-cutover to quickly identify any issues.
4. Provide Support: Ensure support teams are prepared to assist users with any issues.
5. Decommission Legacy Systems: Once stability is confirmed, retire the legacy data warehouse.

Managing Data Governance During Migration

Data governance cannot be an afterthought; it must be integrated throughout migration.

Establishing Governance Frameworks

Before migration, establish governance policies:

1. Data Ownership: Define who owns each data domain and is responsible for quality and access.
2. Classification Scheme: Develop a data classification framework (public, internal, confidential, restricted).
3. Access Control Policies: Define who can access different data domains and under what conditions.
4. Data Quality Standards: Establish minimum quality standards for data in Fabric.
5. Lineage Requirements: Determine what lineage tracking is required for compliance and transparency.

Microsoft Purview provides the governance backbone, but your policies must be defined first.

Implementing Microsoft Purview

Integrate Purview throughout your migration:

1. Catalog Data Assets: Register all data sources, tables, and key datasets in the Purview catalog.
2. Define Business Glossary: Create a business glossary that maps technical terms to business concepts.
3. Track Data Lineage: Document how data flows from sources through transformation to consumption.
4. Implement Classification: Apply data classifications to sensitive or regulated data.
5. Set Up Access Reviews: Establish processes for periodic review of data access.

Good governance during migration prevents governance debt later. It’s far easier to establish proper governance as data moves than to retrofit it afterwards.

Compliance and Regulatory Considerations

For Australian enterprises, ensure your Fabric implementation meets regulatory requirements:

1. Data Residency: Ensure data resides in Australian regions where required by regulation or policy.
2. Privacy Compliance: Implement controls to support Privacy Act compliance, including data subject access and deletion capabilities.
3. Audit Logging: Enable comprehensive audit logging to support compliance verification.
4. Encryption: Implement encryption at rest and in transit.
5. Access Controls: Implement role-based access control (RBAC) and row-level security (RLS) where needed.

At Agile Insights, we help Australian enterprises design governance frameworks that balance data accessibility with compliance requirements.

Performance Optimization Post-Migration

After data migration, performance optimisation becomes critical.

Query Performance Analysis

Comprehensively analyse query performance:

1. Identify Slow Queries: Use Fabric’s query monitoring to identify queries that perform poorly.
2. Analyse Execution Plans: Examine execution plans to understand where time is spent.
3. Compare with Baselines: Compare post-migration performance with pre-migration baselines.
4. Identify Root Causes: Determine whether performance issues stem from data volume, query design, or infrastructure configuration.

Optimisation Techniques

Apply optimisation strategies:

1. Partitioning Optimisation: Refine partition keys based on actual query patterns observed post-migration.
2. Indexing Strategy: Implement appropriate indexes for frequently executed queries. In Fabric, this differs from traditional databases.
3. Materialized Views: Create materialized views for complex calculations that are accessed frequently.
4. Query Refactoring: Rewrite queries to leverage Fabric’s optimiser more effectively.
5. Workload Management: Use workload groups to prioritise critical queries.

Azure Databricks provides additional optimisation through query caching, adaptive query execution, and photon acceleration for compute-intensive workloads.

Cost Optimisation

Whilst cloud platforms offer cost benefits, they require active management:

1. Right-Sizing Capacity: Ensure you’ve provisioned appropriate capacity units; too many wastes money, too few causes performance issues.
2. Scheduled Pausing: Pause clusters during non-business hours if appropriate for your use cases.
3. Query Efficiency: Invest in query optimisation; inefficient queries consume capacity unnecessarily.
4. Caching Strategy: Leverage caching to reduce redundant computation.
5. Regular Reviews: Conduct monthly cost reviews to identify optimisation opportunities.

Proper cost optimisation can reduce cloud infrastructure costs by 20-40% compared to unoptimised deployments.

Change Management and Stakeholder Communication

Technical migration success depends on effective change management.

Stakeholder Engagement Strategy

Develop a comprehensive engagement strategy:

1. Executive Sponsorship: Secure visible executive sponsorship for the migration.
2. Steering Committee: Establish a steering committee with representatives from key business units.
3. User Advisory Groups: Form groups of representative users to provide feedback and identify concerns early.
4. Regular Communication: Maintain regular communication about migration progress, timelines, and any changes.
5. Transparency About Risks: Be transparent about risks and mitigation strategies rather than hiding concerns.

Training and Enablement

Ensure users are prepared for the new environment:

1. System Training: Provide training on Fabric and Power BI for analysts and report consumers.
2. Advanced Training: Offer advanced training for data engineers on Databricks and lakehouse architectures.
3. Governance Training: Train data stewards on governance responsibilities in the new environment.
4. Documentation: Create comprehensive documentation including user guides and FAQs.
5. Support Readiness: Prepare support teams with escalation procedures and known issues documentation.

Change Readiness Assessment

Before cutover, assess readiness:

1. Technical Readiness: Are systems stable and performing acceptably?
2. User Readiness: Have users been trained and do they feel confident?
3. Process Readiness: Are new processes (governance, support, operations) in place?
4. Business Readiness: Do business stakeholders understand impacts and have endorsed the migration?
5. Contingency Readiness: Is the team prepared to execute rollback procedures if necessary?

A go/no-go decision should be based on comprehensive readiness assessment, not just technical completion.

Common Challenges and Solutions

Migrations encounter predictable challenges. Understanding these helps you prepare.

Challenge 1: Complex T-SQL Constructs

Legacy SQL systems often use complex T-SQL features, some of which don’t have direct equivalents in Fabric.

Solutions:

• Identify problematic constructs early during assessment
• Refactor stored procedures into Databricks notebooks where needed
• Use Databricks for complex transformations that don’t map cleanly to Fabric SQL
• Consider temporary bridges using Azure Data Factory for complex ETL logic

Challenge 2: Data Quality Issues

Migration often reveals data quality issues hidden in legacy systems.

Solutions:

• Conduct thorough data quality assessment pre-migration
• Implement data quality checks in migration pipelines
• Use the migration as an opportunity to cleanse data
• Establish ongoing data quality monitoring post-migration
• Engage data stewards to resolve quality issues at source

Challenge 3: Performance Regression

Some queries may perform worse in Fabric initially, particularly if migration simply replicates legacy designs.

Solutions:

• Establish performance baselines before migration
• Identify performance regressions early through pilot testing
• Invest in query optimisation and tuning post-migration
• Consider architectural changes (denormalisation, materialised views) that leverage Fabric’s strengths
• Use Azure Databricks for complex analytics that don’t perform well in SQL

As noted in technical challenges of migrating legacy SQL systems to Fabric, performance issues often stem from mismatches between legacy optimisation strategies and cloud-native optimisation approaches.

Challenge 4: Integration Complexity

Legacy systems often have complex integration landscapes with multiple applications and tools.

Solutions:

• Map all integrations thoroughly during assessment
• Prioritise integrations by business criticality
• Develop integration testing plans
• Consider using data virtualisation or staging areas during transition
• Leverage Microsoft’s integration tools (Data Factory, Synapse) where appropriate

Challenge 5: Resistance to Change

Users accustomed to legacy systems may resist change.

Solutions:

• Communicate benefits clearly and early
• Involve users in design and testing
• Provide comprehensive training
• Maintain legacy system access during transition period if possible
• Celebrate successes and address concerns promptly

Challenge 6: Budget and Timeline Overruns

Migrations frequently exceed initial estimates.

Solutions:

• Build contingency into estimates (typically 20-30%)
• Establish clear scope management processes
• Track progress against plan weekly
• Identify risks early and develop mitigation strategies
• Consider phased approaches to manage scope and risk

Next Steps and Ongoing Support

Migration is not the end; it’s a transition point.

Post-Migration Optimisation

After cutover, continue optimisation work:

1. Continuous Monitoring: Monitor performance, cost, and usage metrics continuously.
2. Iterative Improvements: Use observed patterns to refine queries, indexes, and workload distribution.
3. Feature Adoption: As new Fabric features become available, evaluate adoption where beneficial.
4. Advanced Analytics: Begin implementing advanced analytics and AI capabilities that weren’t feasible in legacy systems.
5. Governance Evolution: Refine governance policies based on operational experience.

Scaling and Future Roadmap

Plan for future evolution:

1. Additional Data Sources: Identify additional data sources to integrate into Fabric.
2. New Analytical Capabilities: Plan adoption of real-time analytics, machine learning, and AI.
3. Organisational Expansion: Plan how additional business units will adopt Fabric.
4. Integration Deepening: Expand integration between Fabric and Azure Databricks.
5. Competitive Advantage: Identify how Fabric capabilities can drive competitive advantage through new insights and capabilities.

Ongoing Support and Training

Continue supporting your team:

1. Operational Support: Maintain support for day-to-day issues and questions.
2. Advanced Training: Provide ongoing training as users mature and take on new responsibilities.
3. Community Engagement: Engage with Microsoft communities and user groups for best practices and solutions.
4. Skills Development: Invest in developing team skills in Fabric, Databricks, and modern analytics practices.
5. Vendor Partnership: Maintain partnership with experienced vendors like Agile Insights for guidance on advanced implementations.

Leveraging Managed Services

Consider engaging managed services for ongoing support:

1. Operational Management: Outsource routine monitoring, maintenance, and support.
2. Performance Optimisation: Engage specialists for ongoing performance tuning.
3. Advanced Analytics: Partner with specialists to implement machine learning and AI capabilities.
4. Governance Management: Maintain governance frameworks with external expertise.
5. Strategic Planning: Work with advisors to plan evolution and adoption of new capabilities.

At Agile Insights, we provide comprehensive managed services to support organisations post-migration, ensuring they realise the full value of their Fabric investment.

Conclusion

Migrating legacy SQL data warehouses to Microsoft Fabric using Azure Databricks represents a significant undertaking, but one that delivers substantial benefits in terms of cost, capability, and competitive advantage. The journey requires careful planning, technical expertise, and strong change management, but the destination is a modern, scalable, intelligent analytics platform.

As enterprise data migration trends show the rise of Microsoft Fabric, Australian enterprises that successfully navigate this transition position themselves to compete effectively in an increasingly data-driven business environment. The combination of Fabric’s unified analytics capabilities with Azure Databricks’ advanced analytics power creates a platform that supports everything from traditional business intelligence to cutting-edge AI applications.

Success requires approaching migration as a strategic initiative, not merely a technical project. It demands investment in planning, governance, change management, and ongoing optimisation. But for organisations that execute well, the result is a modern data platform that accelerates insights, enables innovation, and drives business value.

Your migration journey is unique, shaped by your current environment, business objectives, and organisational constraints. Engaging experienced partners who understand both the technical and organisational dimensions of migration significantly increases success probability. Whether you’re just beginning to assess migration feasibility or ready to execute, the guidance in this comprehensive guide provides a roadmap for success.

The future of analytics is cloud-native, unified, and intelligent. Microsoft Fabric and Azure Databricks represent the modern platform for that future. The question is not whether to migrate, but how quickly you can execute the migration to begin realising the benefits.