Learning Systems & Content Operations

The Course Catalog Was Fine. The Architecture Underneath It Wasn’t.

Institute of Nuclear Power Operations · Floor & Décor · LMS Administration & Content Governance · 2019–2026

Two platform modernizations, seven years apart, with the same root cause: content that people could not find, could not trust, and could not tell was out of date. The fix was never a better search box — it was governance, ownership, and enrollment logic that ran itself.

30+
Business units served
10,000+
Associates supported
300+
Resources migrated
4
Platforms administered
Context & Audience

I have owned learning platforms in two very different organizations. At Floor & Décor, the audience was 10,000+ retail associates across a fast-growing store network — high turnover, low tolerance for friction, learning consumed on the sales floor between customers. At INPO, the audience was a few hundred technical specialists in a high-consequence regulated environment, where the cost of someone training against a superseded revision is not an inconvenience.

Across both, I administered or migrated content on Docebo, Workday Learning, SuccessFactors, Mindflash, and SharePoint.

The Business Problem

Both organizations described the same symptom in different words: “people can’t find the training.” Both assumed it was a platform problem, and both were wrong.

The actual failure was structural. Content had no named owner, so nothing was ever retired. There was no review cadence, so learners had no way to tell a current document from a stale one. Enrollment was handled by request and manual assignment, which meant the platform team was a bottleneck on every single launch. And there was no shared taxonomy, so the same topic lived under four different names across three systems.

Migrating that to a new platform without fixing it would have reproduced the problem in a nicer interface.

My Role

Managed and Led. Platform administration and configuration, content architecture and taxonomy design, migration planning and execution, enrollment and prerequisite logic, vendor liaison, dashboard and reporting design, and the governance model that kept the catalog healthy after launch.

Approach

Content Audit Before Migration

Inventoried the existing catalog against actual usage data before moving anything. A meaningful share of it did not need to move at all — the cheapest content to govern is content you retire.

Named Ownership & Review Cadence

Every asset got a business owner and a review date. Not a policy document — a field in the system, visible to learners, that made currency a property of the content rather than an assumption.

Automated Enrollment & Prerequisite Gating

Replaced request-based assignment with rules driven by role, business unit, and completion state. Prerequisite gating let a learning path control access to a downstream capability — the same mechanism later used to gate Copilot licenses.

Taxonomy & Findability

One naming convention, one topic structure, applied across LMS and SharePoint so a search in either place returned the same answer. Unglamorous, and the single highest-return decision in the project.

Dashboards Built for the Decision

Completion reporting designed backward from what a business leader actually does with it — who is at risk, what is overdue, what is about to expire — rather than exporting everything the platform could measure.

Vendor as a Managed Relationship

Ran the vendor relationship as an ongoing partnership rather than a support queue: roadmap input, configuration reviews, and escalation paths agreed before they were needed.

Key Artifacts

LMS governance architecture · content lifecycle and review model · enrollment and prerequisite logic map · migration inventory and taxonomy · completion and risk dashboards · administrator runbooks

LMS Governance Architecture →Content Governance Framework →
Diagrams are sanitized recreations. No tenant data, internal URLs, or employee records are shown.
Results & Evidence
  • 300+ resources migrated and restructured into a governed SharePoint architecture serving 10,000+ associates at Floor & Décor
  • 30+ business units supported on a single governed Docebo catalog at INPO
  • Enrollment moved from manual request to rule-driven automation, removing the platform team as a launch bottleneck
  • Prerequisite gating became reusable infrastructure — later the control mechanism for the enterprise Copilot rollout
  • Every published asset carries a named owner and review date, making catalog currency auditable rather than assumed
Reflection

Governance is the least popular thing to propose and the only thing that makes a platform survive its second year. The temptation in every migration is to move fast and clean up later; later never arrives, because by then the mess is load-bearing.

What I would do differently: bring the content owners into the audit earlier. I built the ownership model and then went looking for owners. Doing it the other way around would have surfaced the political questions — who gets to retire whose content — while there was still room to negotiate them.