Metadata Management: Creating Relevant and Useful Data

Introduction

IpX (The Institute for Process Excellence) has the blueprint for true continuous improvement that empowers an ecosystem of sustainable excellence, which in turn results in the perpetual success of your organization. This blog-post discusses an important aspect of that ecosystem.

The Management of Metadata

Many organizations we work with have a large amount of data, but the metadata and attributes describing the data have not been identified nor controlled properly. This lack of control over a common definition leads the organization into a Data Rich Information Poor (DRIP) spiral. When the metadata or attributes are not collected and maintained properly, the organization is paralyzing its performance improvement efforts. According to Forbes, 2.5 million terabytes of data is created daily and will only increase as systems are becoming digitized and connected. In recent years, verifiability or statistically significant data has been obscured due to the influx of data creation. When data is not trusted, inaccurate, missing, or unable to be queried without timely manual manipulation, day to day business operations are impacted. Business decisions are slowed or often based on bad data, resulting in an increase in corrective action often caused by a buried root cause.

As businesses collect more data, organizations must be strategic about what they collect, why they are collecting it, who collects it, defining the source of truth, and where it needs to be shared. As the volume of data grows, organizations must process and organize data from internal and external sources rapidly so that it ultimately becomes relevant and useful.

Metadata management is often an under emphasized consideration when implementing a PLM system. Metadata provides context to your data by labeling data's attributes, structure, and relationships with other data so that it can be meaningful to the user. Metadata management is often viewed narrowly as an IT or purely technical topic. Therefore, organizations tend to neglect the act of defining a common business vocabulary to ensure data traceability across the end-to-end lifecycle.

Many organizations incorporate dozens upon dozens of attributes without understanding the needs and impacts to the enterprise business processes. Choosing metadata for tools such as PLM depends on the specific wants and needs of the organization. Simply because you have available data does not mean that you should necessarily collect and retain it. Metadata should be intentionally defined so that it provides a thorough view of the product lifecycle and helps in making educated decisions. Metadata should be relevant, accurate, and consistent across the organization. Metadata does not bring valuable use to the organization unless it is easily accessible and searchable.

PLM metadata is no longer just about CAD data. Metadata management is a crucial aspect of product data governance. Proper planning and organization can help your data be relevant and useful to your stakeholders without manual manipulation and SME knowledge:

1. Define and document metadata management goals and objectives: Who needs to access what data and why? What processes and tools are in scope? What metadata will drive business metrics?
2. Define and document consistent business terminology for datasets and metadata: A common business vocabulary used consistently across all processes, people, and tools throughout the product lifecycle to ensure alignment and eliminate redundancies.
3. Create metadata standards: Rules, guidelines and formats that help define how metadata should be structured and stored. The standard should define what is required for all data. If further metadata needs to be collected about specific types of objects, such as fasteners, consider using classification of the data to extend the attributes.
4. Define data and metadata ownership: Who are the creators and users of the data and metadata? What system is the single source of truth? Who is responsible for maintenance? What are the risks of these decisions?
5. What metadata needs to be under formal change control? Define a formal process to update metadata aside from the dataset itself. All metadata may not need to follow the same rigor as the change and configuration management process for product datasets.
6. Enable traceability of data: Data traceability is the ability to track the origin and movement of data throughout its lifecycle; inputs, outputs, interdependencies along with history of has touch/change/consume the data and who will eventually touch/change/consume the data.
7. Ensure integrity of data: Data accuracy or quality assurance of data is accomplished through proper security: access, permissions and verification and validation review and approvals. Conduct regular audits of metadata to help resolve missing, inaccurate, or out of spec attributes.
8. Organizational metadata governance and standards: Do you need to organizational decision makers who oversee the metadata landscape across your organization and define the overarching strategy for data and metadata and make decisions at the interfaces and approve changes to such?

Data governance and metadata management must not follow a one-time ¨set it and forget it" approach. Both should evolve and adapt to the changes in the data landscape including volume, type, speed, complexity, and applicable data regulations. Data governance and metadata management must evolve as well to ensure alignment with the broader changing goals and objectives of the organization.

Resistance to change and adaptation can impede the improvement, growth, and sustainability of valuable data assets. Proper metadata management can streamline the process to collect, assess and utilize readily available data for business decision making at all levels, saving valuable resource time, and increasing overall organizational efficiency.

‍