Using metadata actively

Metadata as augur & harbinger

Actively using metadata is both a smart way to help de-risk your domain and inevitable for a rewarding digital future based on knowledge infrastructure, the corporate nervous system.

How aware is your business of metadata, and how it can presage both risk & reward?

  • if you grew by acquisition, merger or consolidation, did you harmonise the incoming metadata?
  • when did you last review your metadata?
  • what metadata does the business collect and why?
  • what policies & regulations govern your metadata and how are you implementing that?
  • who manages & monitors metadata standards, accuracy & compliance?
  • who in your organisation enters metadata, how, when & for what reason?
  • are the metadata in emails or back office documents sanitised before despatch?
  • is the metadata in your back office documents inherited or accurate?
  • what metadata do you share in 3rd party systems?
  • have you secured the credentials in your cloud instance metadata API?
  • is your metadata fit for an AI/ML project?
  • how do you update or remove metadata?
  • how do you manage metadata workflow & lifecyle?
  • does your system prevent you, programatically, from having metadata gaps or inconsistencies?

Risk

The regulatory environment spans personally identifiable information (PII), consumer data (GDPR, e-privacy & CCPA), financial & other types of sensitive data. Financial regulations can be multi-jurisdictional, overlapping &  sometimes conflicting – covering financial services, KYC, AML, economic crime, the funding of terrorism (the Patriot Act) or simply good banking practice & risk management (BCBS 239). Severe financial & reputational penalties attach.

Governance and corporate process now intersect in the digital world; in a multi-dimensional dataverse which is actuated by metadata. That is why knowing your metadata is important and why it can alert you to risks such as:

  • data protection or e-privacy non-compliance:
    • through failure to assess or sanitise metadata sources prior to asset acquisition, sharing or disposal;
    • through persistence & dispersion of inaccurate metadata in back office & other systems;
    • through opacity of packaged applications hindering awareness of sensitive, non-consensual or personal data.
  • financial services & AML regulatory non-compliance:
    • through poor governance, reporting & audit due to metadata being inaccurate, incomplete or inappropriate;
    • through (custom) metadata being lost during migration (in certain applications).
  • security risk:
    • user credentials & privileges not secured;
    • cloud instance metadata APIs not secured.

Find out how Safyr  gives visibility of metadata & data structures in opaque packaged applications (CRM, ERP) so you can find & locate (hitherto overlooked) potentially non-compliant data; & quickly reduce your risk profile in packaged applications.

Reward

Your metadata can also be a hacker’s goldmine. Why not make your metadata ‘your’ goldmine instead? Rewards from curated metadata include:

  • value add for AI/ML:
    • curated metadata will enable your data scientists to be scientists, not janitors;
    • historic metadata can provide AI/ML with the baseline for valuable comparative analysis, if its integrity is preserved.
  • integration & interoperability:
    • minimise project delay & cost by being able to access data models & structures in packaged applications;
    • integrate that value into your overall domain.
  • knowledge systems:
    • integrate coherently using metadata, no matter what the data formats, sources or entities are;
    • use your knowledge systems to help with improving data quality;
    • speed your ML model,  use the relationship and entity salience in your knowledge system metadata instead of querying those silos;
    • power integration by bringing hitherto opaque application-centric metadata into your organisational ontology.

Your digital nervous system will rely heavily on active metadata ‘everywhere’ to support the data-centric nature of future operational models so you may as well get ahead of the curve. The longer you fail to address your metadata, the more grievous your eventual pain will be as your data continues to multiply exponentially.