Why Power BI for Manufacturing Now
Manufacturing analytics has been fragmented for twenty years. Various tools serve different purposes:
- Wonderware InTouch: Used for HMI and dashboards on the shop floor.
- PI System ProcessBook: Used for historian visualization.
- Oracle Hyperion or Cognos: Handle corporate finance.
- Custom Excel: Often used for quality management.
The shift to Power BI is driven by several factors:
- Many manufacturers already run Microsoft 365, Azure and Dynamics 365, so the analytics can sit on the same stack.
- Power BI integrates seamlessly with the historian, MES, and ERP ecosystem.
- Microsoft Fabric offers a modern data platform for real-time shop-floor and enterprise analytics.
EPC Group has delivered Power BI to various industries, including:
- Automotive OEMs and Tier 1-3 suppliers
- Aerospace (commercial and defense)
- Industrial (heavy equipment, electrical apparatus, instrumentation)
- CPG (food, beverage, consumer durables, apparel)
- Process manufacturing (chemicals, plastics, paper, pharma)
High-Value OEE + Plant Performance Dashboards
- OEE = Availability × Performance × Quality — at plant + line + cell + asset level with drill-down to downtime Pareto
- Downtime classification — planned (PM, changeover) vs unplanned (breakdown, micro-stops, starved/blocked). Pareto by reason code + asset
- Scrap + yield + first-pass yield (FPY) — at each process step + cumulative through the production flow
- Takt time vs cycle time — gap analysis identifying bottlenecks
- Andon + escalation — real-time alert tracking + response time analytics
- Hourly + shift + daily + weekly + monthly rollups
Define and reconcile the measures before the dashboard
Plants rarely capture downtime, cycle time and scrap the same way, so an OEE comparison built before the definitions agree compares bookkeeping, not performance. Each component is defined once, owned by a named process owner, and tested against the source systems before any plant is ranked.
| Measure | Decision to record | Where plants usually differ | Reconciliation test |
|---|---|---|---|
| Planned production time | Which stops are planned (breaks, changeovers, planned maintenance) and excluded | Changeovers booked as planned at one plant and as downtime at another | For a sample shift, scheduled time minus planned stops equals the planned production time the dashboard shows |
| Availability | Run time divided by planned production time, and the threshold below which a stop counts as a minor stop | Different minor-stop thresholds by line or plant | The same shift recomputed from the MES or historian export matches the dashboard within an agreed tolerance |
| Performance | (Ideal cycle time × total count) divided by run time, and whether ideal cycle time is the engineering standard or the best demonstrated rate | Ideal cycle times kept in local spreadsheets | One versioned ideal-cycle-time table, owned by process engineering, used by every plant |
| Quality | Good count divided by total count, and whether rework counts as good | Scrap recorded at the line in one plant and at final inspection in another | Good, scrap and rework counts reconcile to the ERP production-order confirmations |
| OEE | Availability × Performance × Quality from the three agreed components | Plants comparing OEE built on different definitions | Plant comparison is shown only once every plant passes the component tests on the same definition version |
The semantic model that carries the agreed definitions is certified only after the tests pass. In Fabric and Power BI, a Certified badge means an organization-authorized reviewer has confirmed the item meets the organization's quality standards, and only users a Fabric administrator specifies can certify. The sample KPI reconciliation record (CSV) shows two plants on one definition version, with the open differences and their owners. Its rows are synthetic examples, not client data.
Primary sources
- Endorsement overview (Promoted, Certified, Master data) — Microsoft Learn
- Automatic page refresh in Power BI — Microsoft Learn, updated 2025-12-01
- Real-time streaming in Power BI — Microsoft Learn, updated 2025-01-27
Supply Chain, Quality, Cost, Sustainability
Supply chain. Inbound + outbound logistics, in-transit inventory, supplier OTIF, demand vs supply variance, safety stock + reorder point analytics, transportation cost per unit.
Quality. First-pass yield, scrap + rework cost, defect Pareto, supplier quality, customer complaint + warranty cost, FMEA + control plan effectiveness.
Cost. Standard cost vs actual variance, labor + material + overhead absorption, cost per unit, plant-by-plant cost benchmarking.
Sustainability. Energy intensity, water intensity, waste tracking, emissions tracking (Scope 1 + 2), carbon intensity per unit, supplier sustainability data.
Historian Integration Patterns
OSIsoft PI System, AVEVA Wonderware Historian, Aspen InfoPlus.21 — the major manufacturing historians. EPC Group integration patterns:
- Direct connection via Power BI on-premises gateway for real-time visualization. PI Web API for PI System. ODBC for Wonderware + Aspen.
- Bulk extraction via Data Factory into OneLake (Fabric medallion architecture) for analytics workloads
- PI AF hierarchies mapped to Power BI semantic models for asset-context drill-down
- Modern path (new deployments): Microsoft Fabric + Eventstream for high-frequency tag ingestion + Real-Time Intelligence for sub-second updates
Engagement Investment
Foundation (fixed-fee, 10-16 weeks): Single-plant OEE deployment OR enterprise quality + cost dashboards. F64 capacity. 50-200 users.
Enterprise (fixed-fee, 20-32 weeks): Multi-plant + Fabric data platform + Center of Excellence + EOM full lifecycle. F128. Mid-size multi-plant manufacturer.
Platform (fixed-fee, 36-60 weeks): Enterprise + Microsoft Cloud for Manufacturing + multi-region + sustainability reporting + CMMC L3 (DIB primes). F256.
Related Pages
FAQ
What manufacturing Power BI dashboards do you build?
OEE (Overall Equipment Effectiveness = Availability × Performance × Quality) at plant + line + cell + asset level, with drill-down to downtime Pareto. Downtime classification (planned vs unplanned, changeover, micro-stops). Scrap + yield + first-pass yield (FPY). Takt time vs cycle time. Supply chain (OTIF, inventory days of supply, supplier scorecard). Quality (defect Pareto, FMEA effectiveness, warranty + recall tracking). Cost (standard vs actual variance, plant-by-plant benchmarking). Sustainability (energy intensity, water intensity, waste, Scope 1 + 2 emissions).
How does Power BI integrate with PI System, Wonderware, Aspen InfoPlus.21?
EPC Group has shipped Power BI integrations against the major manufacturing historians: OSIsoft PI System (PI Web API or direct ODBC), AVEVA Wonderware Historian, Aspen InfoPlus.21. Patterns: (1) Direct connection via Power BI on-premises gateway for real-time visualization; (2) Bulk extraction via Data Factory into OneLake (Fabric medallion architecture) for analytics workloads; (3) PI AF (Asset Framework) hierarchies mapped to Power BI semantic models. For new deployments, Microsoft Fabric + Eventstream is the modern path.
Can Power BI handle real-time OEE monitoring on the shop floor?
Yes, with the right storage mode and capacity. Automatic page refresh re-queries a DirectQuery or Direct Lake report page at a fixed interval as short as one second on a Fabric, Premium or Premium Per User capacity, within the minimum the capacity admin sets (five minutes by default); shared capacity is limited to 30 minutes and Import models are not supported (Microsoft Learn, Automatic page refresh in Power BI). For sub-second sensor streams, EPC Group uses Real-Time Intelligence in Microsoft Fabric with Eventstream from MES or SCADA. Power BI streaming semantic models still work, but new ones can be created only until October 31, 2027, and Microsoft recommends Real-Time Intelligence instead (Microsoft Learn, Real-time streaming in Power BI). Custom visuals cover high-frequency time series where the core visuals fall short.
What about manufacturing-specific compliance — CMMC, ITAR, FDA Quality?
CMMC 2.0 for DIB manufacturers: Power BI deployed in GCC High tenant with sensitivity labels for CUI, RLS for cleared-personnel access. ITAR-controlled technical data: GCC High + Customer Key + DKE. FDA Quality (21 CFR Part 11) for life sciences / pharma / medical device manufacturers: validated tenant configuration, electronic signature integration, audit trail per Part 11.10. EPC Group ships all three patterns. See /industries/manufacturing for broader CMMC + ITAR + FDA scope.
How do you handle multi-plant rollouts?
Multi-plant Power BI rollouts: (1) Standardized OEE + quality + cost templates deployed plant-by-plant with local customization; (2) Plant-level semantic models federated into corporate roll-up; (3) RLS enforces plant-manager scope (plant managers see their plant + corporate aggregates only); (4) Corporate dashboard provides plant-comparison + best-practice identification. EPC Group has shipped multi-plant Power BI at automotive OEMs + Tier 1s, aerospace, industrial, CPG (food + beverage + consumer durables).
What's the right Fabric capacity for a manufacturer?
Plant scope drives capacity choice. Single-plant pilot or small mid-market manufacturer: F64 (~$8,410/mo (PAYG)). Mid-size multi-plant (5-15 plants, 200-1,000 users): F128 ($16,819/mo). Enterprise (15+ plants, 1K-5K users, ML training for predictive maintenance): F256 ($33,638/mo pay-as-you-go, East US). See /power-bi-cost-and-licensing-guide-for-desktop-pro-and-premium for full F-SKU pricing.
Why EPC Group for manufacturing Power BI consulting?
Founder Errin O'Connor is a Microsoft Press Power BI author (Microsoft Power BI Dashboards Step by Step). EPC Group's manufacturing work spans automotive (OEMs and Tier 1 to Tier 3 suppliers), aerospace (commercial and defense), industrial (heavy equipment, electrical apparatus), CPG (food, beverage, consumer durables) and process manufacturing (chemicals, plastics, paper). EPC Group holds all six Microsoft Solutions Partner designations. See /industries/manufacturing for the broader practice.
Schedule Manufacturing Power BI Discovery
KPI definitions reconciled before plants are compared. PI System + Wonderware + Aspen experience. CMMC + Part 11 patterns ready.