Enterprise Metadata Management

Understanding Metadata

The most popular description of metadata is that it provides data about other data. And in many instances, this is accurate. Metadata answers the who, what, when, where, why, and how for users of data in traditional IT environments.

In financial applications, metadata is used to define the dimensions, hierarchies, and properties that allow the application to calculate and aggregate data as defined by system administrators. For example, metadata for chart of accounts segments might include accounts, entities, cost centers, products, projects, and employees. 

The Three Pillars of Metadata Management

Without effective enterprise metadata management, your metadata may be inaccessible, inaccurate, and untrustworthy, which can prevent critical data from loading into a system, or even worse, create disparate financial results across business systems. This has a significant impact on business intelligence, financial reporting, process efficiencies, and customer satisfaction.

Here are a few real-world examples of how the three pillars of enterprise metadata management—accessibility, accuracy, and trustworthiness—can impact business systems:

• Accessibility: A new account dimension member was not created in the ERP and/or EPM application before raising an invoice. Without the account number metadata in place, the transaction fails when data is loaded.
• Accuracy: A member was created in the ERP but not in OneStream or the data warehouse. When changes are deployed, there will be a discrepancy between the systems.
• Trustworthiness: An address was added with a slight variation in the ERP and attributed to a different member name in Hyperion. When data is loaded to both applications, a reconciliation error will occur.

The Business Value of Metadata Management

The three pillars shape the business value of metadata management. When one pillar does not align, the business cannot achieve value from the underlying data. 

By ensuring accurate, reliable metadata is in the right place at the right time, the business can be confident that master data is correct and consistent in every application across the organization.

How Data Governance Supports Metadata Management

Data governance creates a support structure for metadata management in three main ways:

1. Design and enforce data governance practices using highly configurable workflows tied into a robust approval process.
2. Require reviews and approvals to load metadata into subscribing applications.
3. Ensure metadata is valid and technical requirements have been met.

With this framework in place, the change request and deployment process is efficient and accurate, ensuring high-quality, consistent metadata is loaded into all subscribing applications.

Types of Data Governance Rules

Every application that subscribes to metadata needs to have governance requirements in place. These requirements define and enforce policies and protocols that ensure the metadata that is being requested has been properly created.

For example, many public enterprises have a SOX requirement for internal audits that states auditors need to know what metadata was requested, who signed off on it, when it was placed in a user acceptance testing environment, and whether it moved to production with the proper approvals in place. Data governance policies and audit reports make this information easy to track and recall. 

There are two main types of governance rules that apply to metadata:

1. Technical governance rules

These rules govern the technical limitations of the application that is going to consume that metadata, such as naming convention, character limit, and numeric format.

Technical rules help ensure the metadata is accepted by the application and is in place before data is loaded. If metadata is rejected, the data transaction will fail, creating a backlog in the system that will affect performance and require manual intervention to resolve.

2. Functional governance rules

Functional governance rules establish the business logic to say that the metadata you are creating is valid for the data structure you are creating change in. For example, if there is a functional naming rule that all cash accounts must start with 1, and you create a new account that starts with 2, you have violated a governance rule, and the transaction will not proceed for approvals.

Q: What is the difference between master data management and metadata management? 

A: From a purely semantic viewpoint, the difference between master data management and metadata management can be defined as:

• Master data management: “A technology-enabled discipline in which business and IT work together to ensure the uniformity, accuracy, stewardship, semantic consistency, and accountability of the enterprise’s official shared master data assets.” [Gartner]
• Metadata management: “The administration of data that describes other data. It involves establishing policies and processes that ensure information can be integrated, accessed, shared, linked, analyzed, and maintained to best effect across the organization.” [TechTarget]

But what that really boils down to is that master data management links all of a business’s critical data to create a single source of truth, while metadata management builds confidence that the master data is accurate and trustworthy.

Q: What are the symptoms of poor metadata management?

A: There are three main signs that your organization needs better metadata management processes:

1. Inefficient change approval process

If it takes a week to get approval on a change request, you need a new metadata management strategy. Delayed approvals put transactions at risk of failure, causing bottlenecks, lost revenue, and unhappy customers. 

2. Manual processes that can’t keep up with business growth

With greater growth comes greater complexity. When your business needs multi-vendor solutions to manage financial, accounting, and business intelligence systems, it’s time to invest in a metadata management solution such as EPMware to centralize change requests, data transformation, and deployment.

3. High rate of transaction failure

When metadata changes aren’t input into applications fast enough, transactions and data loads fail. A metadata management solution with out-of-the-box integration with leading ERP, CPM, and other business systems can deploy changes to all target applications and assure alignment.

Q: How do we ensure metadata is compliant and auditable?

A: Implement a unified metadata management and data governance solution that provides real-time validation and a centralized workflow engine. These features ensure metadata is standardized and rationalized before deployment, so all business systems are compliant and audit-ready.

Q: How do I achieve metadata alignment across applications?

A: If a rationalization or harmonization can be done to align metadata, it’s wise to conduct this exercise before implementing any technical solution. With that said, some applications cannot truly rationalize and must require differences in dimensions and use robust business rules in a technical platform to ensure metadata is derived in all subscribing applications while enforcing corporate governance standards. 

With out-of-the-box integrations with supported applications and built-in validations for those supported integrations, EPMware lets business users easily manage metadata changes across multiple business applications from a centralized and web-based platform. 

Key features and functionality that make EPMware a leader in enterprise metadata management include: 

• The ability to author metadata in EPMWare’s centralized hub keeps all subscribing systems and data warehouses in sync.
• Plug-and-play business system adapters create a seamless link to many ERP and EPM applications, eliminating the need to spend months building and maintaining integrations.
• EPMware’s workflow engine migrates changes across multiple environments (e.g., DEV -> TEST -> PROD) from one EPMware production environment.
• Dimension mapping means there is no need to load metadata into individual applications; EPMware derives and maps updates across business systems for you.
• Business logic written into EPMware automatically triggers enforcement of governance rules and policies.
• Metadata can be shared between applications with different data models, which is useful for financial systems.

• Application integration: Merging and optimizing data and workflows across two or more disparate applications 
• Business intelligence: A software-driven process that allows organizations to analyze raw data from multiple sources and use the resulting information to make informed business decisions
• Cloud integration: Connecting multiple cloud-based business systems with each other and with on-premises applications to create a single, cohesive infrastructure
• Code-free integration: Integrations across systems using a visual interface to deploy rather than modifications to the codebase; allows nontechnical users to make changes and create reports without IT assistance
• Corporate performance management (CPM): Includes all of the methodologies, metrics, processes, and systems used to track and manage business performance at the enterprise level
• Data analytics: Extracting insights from one or more data sources that can be used to identify patterns, monitor performance, drive decisions, and shape business outcomes
• Data cleansing: Improving overall data quality by correcting or deleting incorrect, inaccurate, irrelevant, and missing data
• Data governance: The strategies, policies, processes, and technologies used to ensure business data stays in compliance with regulations and adheres to corporate rules
• Data management: Implementation of strategic policies and procedures that allow organizations to control their business data across systems
• Data quality: A measure of the utility of data to serve an intended purpose based on characteristics including accuracy, completeness, consistency, and reliability
• Data repository: Centralized data storage infrastructure used to collect and store data for analysis and reporting
• Data synchronization: The process of creating consistency among data records from source to a target and ensuring harmony of the data over time
• Data transfer: Process of collecting, replicating, and transmitting large datasets from one system to another
• Data transformation: Validation and normalization of business data format, structure, or values to ensure usability in downstream applications and processes
• Data validation: Systematic checks that are built into a system to ensure the data being entered and stored is accurate and has logical consistency
• Data virtualization: Data layer that integrates data from across multiple data sources for analysis and business intelligence
• Data warehouse: Centralized repository that stores aggregated structured data from disparate sources to support reporting and analytics
• Enterprise content management (ECM): Combination of tools, strategies, and processes that support capturing, managing, storing, and delivering data throughout its lifecycle
• Enterprise performance management (EPM) system: Processes and tools that monitor performance across the enterprise that allow stakeholders to analyze, understand, and report on business data
• Enterprise resource planning (ERP): Software designed to improve efficiency through orchestration and coordination of business strategies and operations
• Master data: A consistent and uniform set of identifiers and extended attributes used to describe the core entities of the enterprise, including customers, suppliers, hierarchies, and chart of accounts
• Master data management: Creation of a single source of truth for master data from across the business’s internal and external data sources and applications
• Metadata: Structured reference data that helps sort and identify attributes of information assets and add the context needed to govern systems and data
• Metadata management: Management of policies and processes that ensure metadata can be integrated, accessed, maintained, and analyzed across the organization
• Multi-domain master data management: Integrated management of all domains or data types in a single, centralized platform
• Software as a service (SaaS): Software that is delivered via the internet or the cloud rather than being physically installed on-premises
• Target application: Upstream or downstream application where data changes that receive data updates after changes have been verified, standardized, and rationalized
• Unstructured data: Data that cannot be stored in a traditional relational database because it does not conform to a predefined data model and lacks identifiable structure or architecture

[Sources: Gartner, Informatica, TIBCO]