Keeping Knowledge Current: Taxonomy Governance and Enrichment for the New York Power Authority.

About the Client

NYPA, the largest state public power organization in the United States, operated 16 electricity generation facilities and more than 1,400 circuit-miles of transmission lines at the time of this engagement, managing content across two primary systems — OpenText Content Server and Microsoft SharePoint. Iknow’s prior engagement had built NYPA’s first enterprise taxonomy and unified search experience across both systems using Smartlogic’s Semaphore platform; this follow-on project picked up where that work left off.

Industry Context

Enterprise taxonomies are a known perishable asset: the initial build typically receives significant investment and attention, but without an active, recurring governance process, vocabularies drift out of step with evolving business language, and the precision and recall of autoclassification degrade as a result. A defined update cadence — commonly semiannual or annual — paired with formal subject matter expert review is widely regarded as best practice for sustaining enterprise taxonomies over time, precisely because it forces organizations to treat taxonomy management as an ongoing operational discipline rather than a one-time project deliverable.

Current Situation

NYPA retained Iknow specifically to review its initial enterprise taxonomy in light of a year of accumulated feedback, install a governance model for future taxonomy management, and carry out the first round of semiannual updates to the ontology and rules within Semaphore — continuing the taxonomy management work begun under Iknow’s earlier engagement.

• Taxonomy drift after initial deployment. A year of real-world use surfaced feedback and new requirements — particularly detailed technical vocabulary for expertise and lessons learned in the generating plant environment — that the original roughly 1,100-term taxonomy had not been built to capture.
• No formal, repeatable governance process. The original engagement had built the taxonomy itself, but NYPA had no standing process for reviewing, approving, and periodically updating it going forward, risking either governance paralysis or uncontrolled ad hoc changes.
• Risk of introducing new inconsistencies while expanding the vocabulary. Adding new, more detailed terms without careful testing risked systematic over-tagging or under-tagging that could quietly erode the search precision and recall NYPA had just gained.
• Enterprise-wide re-classification at scale. Any meaningful taxonomy update needed to be propagated across NYPA’s full existing content base, not just applied to new content going forward.

• Review the current enterprise taxonomy against a year of user feedback and new KM team requirements.
• Update and expand the taxonomy, particularly deeper technical vocabulary for generating-plant expertise and lessons learned.
• Validate the updated ontology and autoclassification rules through a real trial before wider rollout.
• Install a durable, SME-driven governance model for future semiannual taxonomy management.
• Execute NYPA’s first full round of semiannual updates, including bulk re-classification of existing enterprise content.

How Iknow Structured the Work

Iknow structured the engagement around four phases that moved from taxonomy review, to controlled testing, to durable governance, to full-scale re-classification — deliberately building on the governance foundation and procedures established during the original 2017 engagement rather than starting from scratch.

Key Activities & Decisions

• Initial Taxonomy Review. Iknow reviewed user feedback on the current enterprise taxonomy alongside new requirements the KM team had developed since the close of the earlier project, with particular focus on more detailed technical vocabulary for describing expertise and lessons learned in the generating plant environment. Iknow performed a draft manual classification of new content against the proposed vocabulary, then implemented draft taxonomy updates and additions ready for review by NYPA subject matter experts.
• Autoclassification Trial of New Content. To test the new ontology and rules before wider rollout, Iknow ran a trial autoclassification using best practices, lessons learned, and expertise-related content, using the results to refine the ontology — especially alternative labels and other evidence values — and the classification rules as needed.
• Initiation of the New Ontology Governance Process. Iknow assisted NYPA in preparing and executing the new governance model, particularly the review sessions with SMEs from each NYPA department. These sessions were used to discuss and approve recommended taxonomy changes and to review autoclassification results for any systematic over-tagging or under-tagging, developing correction strategies where needed.
• Bulk Re-Classification of Enterprise Content. As in the previous engagement, Iknow assisted with planning and initiating the bulk re-classification jobs that applied the updated taxonomy and rules set across NYPA’s existing enterprise content, then conducted a final review of the results.

Stakeholders & Collaboration

Iknow worked closely with NYPA’s KM team throughout the engagement, along with subject matter experts from each NYPA department who reviewed and approved taxonomy changes and helped validate autoclassification results against real operational knowledge.

Expanding a Live, Already-Adopted Taxonomy Without Disrupting Users

NYPA employees had already adapted to the original taxonomy structure, and a poorly tested expansion risked confusing users or degrading search quality they had come to rely on. Iknow addressed this with a draft-then-validate approach: manually classifying new content against proposed additions first, then running a trial autoclassification specifically to test how the new terms and rules behaved on real content before touching production — catching over-tagging and under-tagging issues while still in a safe testing environment.

Building a Governance Process NYPA Could Run Independently

A governance model that only Iknow could operate would simply recreate the dependency the engagement was meant to resolve. Iknow addressed this by structuring the governance sessions explicitly around NYPA’s own departmental SMEs reviewing and approving changes, and by reusing the same procedures established in the original engagement — candidate term list compilation, group reviews, development-environment updates, and testing — giving NYPA a repeatable playbook rather than another one-time deliverable.

Operational Outcomes

• Updated more than 1,000 of the original roughly 1,100 taxonomy terms, including hierarchy and grouping changes and synonym and acronym updates.
• Added 380 new terms for more specific concepts shown to be useful for searching, developed through multiple SME and KM team reviews, search simulations, and classification tests.
• Completed NYPA’s first full semiannual update cycle and bulk re-classification of enterprise content under the newly installed governance model.

Strategic and Organizational Outcomes

Beyond the taxonomy enrichment itself, the engagement gave NYPA a proven, SME-driven governance process it can use for every future update cycle without depending on outside consulting support for routine maintenance. The enriched taxonomy is expected to further improve the precision and recall of NYPA’s enterprise search tools, along with enhanced browsing and content discovery for users across the organization.

Timeline to Impact

Iknow completed all four phases within the 12-month engagement, carrying NYPA’s first full semiannual governance and update cycle through to completion before the engagement closed — meaning the new governance model was proven in practice, not just designed on paper.

Core Skills

• Taxonomy & Ontology Management
• Enterprise Content Governance Design
• Autoclassification Tuning & Quality Assurance
• Search Precision & Recall Optimization

Methods & Frameworks

• SME-driven governance review methodology
• Draft-then-validate taxonomy expansion approach
• Trial autoclassification testing
• Bulk re-classification planning and execution

Technologies & Tools

• Smartlogic Semaphore (ontology management and autoclassification)
• SharePoint 2013 search tools
• OpenText Content Server