Subject line: Use Case Families, combined-list knowledge graphs, and an expanded KPI catalog
This week's big addition: Use Case Families
When you browse mfgusecases.com today, you'll see something new in the filters sidebar: a Family filter, with 17 options ranging from Quality Control & Defect Prevention to Equipment Reliability to Leadership Behavior & Accountability.
A use case has always had a Department — its organizational owner. Family is different. Family is the conceptual home the use case shares with related use cases owned by other departments. The Quality team's accountability use case, the Plant Manager's accountability use case, the OpEx team's accountability use case, and the Plant Finance accountability use case all live in different departments — but they share a family: Leadership Behavior & Accountability. Click the family filter, and you can see all of them at once.
Why it matters: cross-cutting analysis is what mfgusecases.com is good at. Families make that explicit. Instead of clicking through five different department filters trying to surface the same conceptual thread, you click one. And on every use case page, there's a new family chip next to the department badge — click it to see every other use case sharing that conceptual home.
Some specifics:
- 17 families covering the full corpus, structured around the practitioner concepts manufacturers actually use (Standard Work, Daily Management, Quality Control, Equipment Reliability, Materials & Flow, Planning & Scheduling, Workforce Capability, Leadership & Accountability, Strategy Deployment, Financial Performance, Supplier & Procurement, Safety/Health/Environmental, Data Architecture & Quality, Digital Infrastructure, and others).
- 515 use cases assigned automatically based on a multi-engine consensus classification (Claude + Gemini Borda aggregation).
- The legacy "Associated Functions" filter has been retired in favor of Family. The underlying Business Functions data is still tied to each use case, just no longer in the sidebar — it still appears on the use case page itself.
The full breakdown of how Families fits alongside the other navigation dimensions is in the How This Site Is Organized Guide page, which now documents eight navigation dimensions instead of seven. The Anatomy of a Use Case page also picked up a note about the new family chip in the use case page hero.
Combined knowledge graph for saved lists (Pro)
If you're a Pro user with a saved list of two or more use cases, you'll see a new "Combined Graph" tab on the list page. It builds an aggregated knowledge graph across every use case in the list — KPIs, value leaks, root causes, and enablers all rolled up to the canonical level so you can see the patterns.
The big move: causes and enablers consolidate to the canonical level when they appear in multiple use cases. So if "Inconsistent measurement methods" shows up in seven of the use cases on your list, it's one node, not seven, and the influence-score edges across all seven aggregate. You get the strongest signals at a list level — useful for portfolio prioritization, focus area identification, or just "what's the common thread here?"
It works on shared lists too. If a Pro user shares a list publicly, viewers can access the same combined graph view through the share token.
A threshold slider on the combined graph defaults aggressively (top-5%-of-edges) so wide lists stay readable. Pull the slider down for the full network when you want it.
KPI catalog: 38 new metrics, 8 new strategic goal areas
The KPI catalog grew substantially this week. Some highlights:
The OEE family now has explicit decomposition. OEE was already in the catalog; we've added the three OEE factors (Availability, Performance, Quality) and TEEP (Total Effective Equipment Performance) so use cases can connect to whichever piece of the OEE story they're actually about, not just the composite.
There's a new cluster of metrics for what practitioners call process velocity: Decision Cycle Time, Continuous Improvement Velocity, Problem-Solving Cycle Time, Strategic Initiative Completion Rate, Change Implementation Success Rate. These came from a corpus audit that surfaced 9 distinct themes in the existing instance metrics that didn't have canonical homes. Adding them as canonicals lets the platform treat process velocity as a first-class diagnostic, not just adjacent prose.
Other new clusters this week:
- Daily Management & Lean Practices — Daily Management Routine Compliance, Action Item Closure Rate, Frontline Suggestion Implementation Rate, Leadership Gemba Engagement Index. New L2 strategic goal: Strengthen Daily Management & Lean Practices.
- Finance Operations — Month-End Close Cycle Time, Finance Reactive Work Ratio, Finance Response Time. Distinct from the cost-and-margin financial metrics that already existed; this is finance-as-a-business-partner territory. New L2: Improve Finance Operations Efficiency.
- Cyber & Digital Risk — Cyber Incident Severity, Patch Compliance Rate. New L2: Reduce Cyber & Digital Risk (under the Resilience L1 goal).
- Quality Issue Resolution & Change Discipline — Engineering Change Order Rate, Work Instruction Revision Frequency, Control Plan Change Cycle Time, Uncontrolled Temporary Change Duration. New L2: Manage Quality Issues, CAPA & Change Discipline.
- Workplace Wellbeing — Workers' Compensation Cost per FTE, Employee Grievance Rate, Labor Compliance Violation Risk Score. Populates the previously-empty Workplace Ergonomics & Occupational Health goal.
- Diversity & Inclusion — Diversity Representation Rate. New L2: Improve Workforce Diversity & Inclusion.
- Material Waste vs. Cost of Poor Quality — Conversion Waste, Material Yield Rate, Material Variance %. The L2 goal split into two: "Reduce Material Waste" (yield/conversion focus, lever is process design) and "Reduce Cost of Poor Quality" (defect-driven failures, lever is process control). Sharper boundary, less arbitrary classification.
- Throughput & Capacity — Throughput Rate, Production Volume, Takt Time, Capacity Utilization. Previously the throughput L2 only had Revenue Growth — a top-line outcome that didn't really fit. Revenue Growth has now moved to a new top-line financial L2 (Grow Revenue & Top-Line Performance), and the throughput L2 has actual throughput metrics.
- OEE Asset Effectiveness — split out as its own L2 (Improve Asset Effectiveness & Utilization), separate from the financial Asset ROI goal where OEE used to live alongside ROA, ROI, and TCO.
Net effect: 44 strategic Level-2 goals (up from 36) under 8 Level-1 goals. 172 canonical metrics (up from 125). Eight new L2 areas, three renames, six metrics relocated to better-fitting L2 homes. The KPI library at /kpis reflects all of it, and the contextual graph views update accordingly.
Smaller but worth knowing
- Metrics now visible to anonymous users. Key Metrics and Financial Metrics moved from the Analysis tab (registered-only) to the Overview tab (visible to everyone). If you're considering an account, you can now see what metric coverage looks like on each use case before deciding.
- "Most Popular" is the new default browse sort. Use cases with the highest 30-day view counts surface first. Powered by a daily refresh job that builds a
view_countcache from the activity log. Sort options: Most Popular / Newest / A–Z / By Department. - Branded OpenGraph image. Share a mfgusecases.com link in Slack, LinkedIn, or Teams and you'll now see a branded preview card with the platform's hourglass-topology illustration instead of a plain URL. Works for the homepage and use case pages.
- Process area as primary label, everywhere. The site has been migrating from "title" to "process_area" as the primary label on use cases. This week we tied off the loose ends across saved lists (owner + shared), AI Finder results, admin analytics, the weekly newsletter, PDF exports, user activity summaries, team usage stats, and Q&A views — sixteen-plus places now consistently show the conceptual scope first and the full title as subtitle.
Behind the scenes
The bigger story underneath this week's user-visible changes is a lot of taxonomy validation and cleanup that doesn't show up in any single screen but improves the foundation across the platform.
We surveyed five external KPI frameworks — ISO 22400, the MESA Metrics Guidebook, SCOR, the OEE/TEEP family, and the operational excellence literature — plus a multi-customer practitioner roster from years of plant assessment work. The result: a documented gap analysis against the platform's KPI catalog, which is what drove the 38-canonical expansion above.
We also added embedding columns and HNSW indexes to canonical_causes, root_causes, canonical_enablers, enablers, metrics, and key_metrics — about 28,000 embeddings populated this week. That powers a new clustering review process: UMAP dimensionality reduction and HDBSCAN clustering on the embeddings surface candidates for taxonomy consolidation (canonicals that should merge, canonicals that should split, dense clusters with no canonical home that suggest a new concept worth naming). The clustering output for causes, enablers, and metrics is documented; reconciliation actions are queued for a focused future sprint.
A reclassification pass against the new canonicals improved the linkage rate on existing instance metrics from 98.0% to 99.5%, and pulled previously-misclassified instances into more accurate canonical homes (especially for the corpus-driven additions like Decision Cycle Time, Daily Management Routine Compliance, Throughput Rate, and the OEE factor decomposition).
Coming soon
- Filter audit. Industry, Competitive Advantage, and Stakeholder Group filters are getting an evaluation pass — some of these dimensions may be high-cardinality with low click-through, and we'd rather prune than carry. Decision next sprint.
- Acting on the clustering candidates. The merge/split/missing-canonical lists from this week's clustering work need domain review and then a reconciliation sprint to consolidate redundant canonicals and add genuine new ones. Several enabler canonicals in particular look ready for consolidation (the AR family, the AI/analytics family, multiple "Standards" canonicals).
- Topics dimension. Sprint 25's family classification also captured second- and third-rank votes per use case, which become "topics" — cross-cutting practice tags that can appear on any use case regardless of family. Designed but not yet built.
- Refinements to the KPI graph. With 44 L2 goals and 172 canonical metrics now in the model, the strategic-goals layer of the knowledge graph is denser than it used to be. Some visual polish queued.
As always, the full use case library is free to browse. Pro access unlocks root cause analysis, enabler details, the deeper layers of each use case, and now the combined-list knowledge graph.
Questions, feedback, or want to chat about whether this fits your workflow? Reply to this email — we read everything.
— Tim Stuart, Visual Decisions