Why one pillar is not enough
Every plant we assess sits at a different maturity level on four different axes. We have seen plants with world-class 5S and visual management (OpEx 85) that still ship late because their scheduling is chaos (Manufacturing Performance 45). We have seen plants with gorgeous dashboards (Digital Manufacturing 80) whose operators ignore them because the standard work was never updated (OpEx 40). And we have seen plants attempting predictive maintenance (AI 55) on assets whose basic downtime data is still being recorded on paper (Digital 25).
Each of those failure modes is the same mistake: investing in one pillar without the supporting pillars underneath it. The four-pillar framework is designed to prevent that mistake.
Pillar 1 — Operations Excellence (the foundation)
OpEx is the cultural and procedural foundation on which everything else rests. It is Lean, 5S, Kaizen, Value Stream Mapping, standard work, visual management, short interval control, Gemba walks, Hoshin Kanri. Without OpEx, no technology deployment survives the first operator turnover.
Typical interventions:
- 5S audits with digital photo evidence, scored weekly per zone
- Standard work documents in the operator's language, read-and-signed with electronic signature
- Short-interval control huddles at the start of every shift with a 10-minute agenda template
- Kaizen idea funnel with a visible savings tracker (closed events per month, euros/year saved)
- Supervisor coaching cadence — 1 hour per supervisor per week in the first 90 days
Typical outcome in 90 days: 15-30% labor productivity gain, 40-60% paper reduction, measurable cultural shift documented by closed Kaizen events per month.
Pillar 2 — Manufacturing Performance (the measurable output)
Where OpEx is the practice, Manufacturing Performance is the measurement. OEE, SMED, TPM, first-pass yield, on-time delivery, bottleneck management. This pillar turns the cultural work of OpEx into numbers that the CFO can see on the executive dashboard.
The 6 big losses framework (from TPM) is the heartbeat of this pillar: breakdowns, setup & adjustment, idling & minor stops, reduced speed, defects, startup losses. Instrument every line against those 6 categories and you have a map of where to attack first.
- Baseline OEE per line, per shift, per product
- SMED kaizen on the slowest-changeover line (expect 30-50% changeover-time reduction)
- TPM program on the top 3 downtime-cost assets
- Daily production huddle with ONE performance target per line
Typical outcome in 90 days: 8-15 OEE points per line, 10-20% throughput increase on bottleneck lines.
Pillar 3 — Digital Manufacturing (the enabler)
Pillars 1 and 2 can be run on paper. But once a plant tries to scale beyond one line or synchronize across sites, paper collapses. Pillar 3 replaces paper with a mobile-first, operator-friendly digital layer: production counts auto-captured, downtime reasons entered on tablets, quality checks digitized, dashboards reflecting shop-floor reality in real time.
Important: Digital Manufacturing is not about implementing SAP. It is about giving the people on the shop floor tools that are faster and more useful than the paper they currently use. If a tablet-based log takes longer than a clipboard, operators will never adopt it — and the entire investment is lost.
- Operator terminals (tablets, kiosks) at every line with offline-first UX
- Automated connection to ERP for order status, bills of material and inventory
- Shop-floor dashboards visible to supervisors in real time (not weekly reports)
- Shift handovers as structured templates rather than hallway conversations
Typical outcome in 6 months: 70-100% paper elimination, 2-4 hours of supervisor time freed per shift.
Pillar 4 — AI & Intelligent Operations (the multiplier)
Once pillars 1-3 are in place, AI becomes the multiplier. Not before. Predictive maintenance on an asset with no basic downtime data is a hallucination machine. AI quality anomaly detection with no digital quality workflow is a vanity dashboard.
With pillars 1-3 as foundation:
- Predictive maintenance on critical assets — vibration + current signatures, failure prediction with lead time
- AI quality anomaly detection on process data (SPC with pattern recognition) and vision data (defect detection on the line)
- AI-driven scheduling that reroutes jobs when a line goes down
- Downtime root-cause copilot that suggests probable root causes based on asset history and similar past events
- LLM-assisted standard-work drafting — the Copilot drafts the new SOP from a Kaizen event, operators review and sign
Typical outcome: 20-40% reduction in unplanned downtime on predictive-maintained assets, 15-30% reduction in internal scrap.
Sequencing — the 12-month program
The pillars overlap but the starting weight moves predictably through the program:
| Phase | Months | Primary focus | Secondary |
|---|---|---|---|
| Diagnose | 0-1 | Pillar 1 + 2 baseline | Pillar 3 architecture |
| Instrument | 1-3 | Pillar 3 digitization | Pillar 1 standard work |
| Lift | 3-6 | Pillar 2 OEE/SMED/TPM | Pillar 1 Kaizen cadence |
| Multiply | 6-12 | Pillar 4 AI pilots | Scale pillars 1-3 to full plant |
Trying to run pillars 2 and 4 in parallel in month 2 is the single most common failure mode in Industry 4.0 programs. Sequence matters.
Where to start
Our free 10-minute Plant Assessment scores your plant on all four pillars simultaneously, benchmarks you against the industry median, and tells you which pillar to attack first based on your gap. In 10 minutes you leave with a personalized sequencing recommendation.