Optimizing Multi-Site Food Manufacturing: Reducing Downtime and Scrap

Role: VP Operations

Industry: Mid-size packaged food manufacturer in South America

Situation:

Run a multi-site manufacturing organization with older lines, frequent unplanned downtime, and rising scrap. Maintenance works hard but stays reactive because we lack reliable condition data and the root causes repeat across plants. Quality holds create rework and late shipments. See uneven performance between sites because each plant runs its own playbook and metrics definitions differ. IT has an ERP, but the shop floor data is thin and operators still rely on paper logs. Leadership wants a 12-month plan that improves OEE, reduces scrap, and standardizes daily management without a large headcount increase.

Question to Marcus:

“

Which Management Systems and implementation steps should we prioritize to cut downtime and scrap while standardizing operations across plants?

Total Productive Maintenance

In year‑one prioritize TPM as the operating model to shift maintenance from reactive to mostly preventive—start with a 90‑day pilot on the most problematic line in one plant. Deploy the Autonomous Maintenance pillar: train operators to own daily cleaning, lubrication, basic inspections and standardized checks (one‑page checklists).

Capture simple condition checks (vibration/temperature as needed) via handhelds or paper-to-digital photos. Parallelize with quick wins: eliminate easy-to-fix root causes (belt tensions, guards, feed hoppers) and lock in through Standard Work. Use daily shift handoffs to surface equipment health so maintenance schedules become proactive. KPI focus: reduce minor stops and defects (top two drivers of OEE loss) within 3 months, shift reactive work % down by 25–40% in 6–9 months. Keep scope limited to top 10 assets by risk and volume to avoid overtaxing staff. Ensure TPM activities are mapped to food safety controls (GMP/ISO 22000) so improvements support compliance. Communicate wins visibly to reduce resistance and reinforce operator ownership without adding headcount.

Planned Maintenance

Implement a pragmatic planned maintenance program targeted at critical assets across all sites. First 30 days: inventory top 30 assets (per plant) by failure cost, scrap impact and run hours; classify criticality.

Next 60 days: convert high‑criticality reactive jobs into simple PM tasks (inspection, lubrication, fastener checks, consumable replacements) with clear frequencies and durations. Use maintenance routing sheets and a simple work order template; require trades to log causes and parts used to feed continuous improvement. Introduce kill‑switch: any unplanned stop >30 minutes triggers immediate RCA and PM update. Prioritize spare parts for those critical assets—create a shared regional min/max list to reduce stockouts and capital tied in spares. Rebalance technician schedules to protect planned hours (target 60–70% planned work within 9–12 months). Keep tools and templates light so existing maintenance crews can implement without new hires.

Enterprise Asset Management

Replace paper work orders with an accessible EAM/CMMS capability—this is the backbone for consistent maintenance across sites. If ERP modules are weak, deploy a light CMMS that integrates via CSV/APIs and supports mobile work order entry/offline mode for plant floors with poor connectivity.

Phase 1 (0–90 days): build equipment hierarchy, asset criticality, BOM for top assets, and spare parts catalog for critical spares. Phase 2 (90–180 days): roll out mobile work orders, PM scheduling, and simple KPI reporting (planned vs reactive hours, MTBF, MTTR). Standardize failure coding across plants to enable root‑cause analytics. Use the CMMS to enforce PM SOPs, capture labor and parts usage, and automate reorder triggers for critical spares—this reduces downtime waiting for parts. Keep configuration minimal: avoid heavy customization; prioritize ease of use and Spanish/Portuguese UIs. Expect adoption friction—pair with operator/tech champions and short, practical training sessions.

Data & Analytics

Begin with a “minimum viable” shop‑floor data strategy: unify definitions (OEE, downtime, scrap), capture only high‑value signals, and create a single dashboard per plant. Month 0–2: standardize metric definitions and downtime/scrap code taxonomy across sites.

Month 2–6: digitize operator logs using low‑cost tablets, barcode scanning for batch IDs, or simple IIoT retrofits on the most critical machines (vibration, run/stop). Feed standardized data into a light BI tool to show OEE by line, downtime by cause, and scrap by SKU in near real time. Use Pareto analysis to target top 20% of causes that create 80% of losses. Analytics should drive actionable decision rules (e.g., when a line hits X% unplanned stops, pull a maintenance crew). Keep models simple—descriptive and diagnostic analytics first; reserve predictive maintenance pilots for assets with sufficient data and ROI. Ensure regional IT support and data governance so plants trust “one source of truth.”

Overall Equipment Effectiveness

Standardize OEE definitions and governance across the footprint in month 0–30 days: common formulas for availability, performance and quality, and rules for planned downtime exclusions (e.g., sanitation windows). Baseline OEE per line within 45 days and display daily targets; focus first on Availability and Quality because those drive scrap and late shipments.

Implement easy data collection (manual signoffs, simple counters, or a sensor per line) and require a short shift‑level review of OEE deltas during the daily huddle. Use OEE to prioritize improvement projects—target lines where small improvements produce large throughput gains. Set realistic phased targets (e.g., +5–10 pp OEE in 6 months on pilot lines). Tie OEE dashboards into incentive and escalation rules but avoid punitive culture—use OEE as a diagnostic tool to resource maintenance and quality investigations.

Lean Daily Management System

Introduce a Lean Daily Management System (LDMS) to standardize cadence and reduce firefighting. Define a simple meeting rhythm: 15‑minute shift huddles (operators, maintenance, QA), a 30‑minute plant daily review (line KPIs, safety, customer priorities), and a weekly cross‑plant review for site managers.

Create an escalation matrix with clear owners and timelines so problems that can’t be resolved within the shift are escalated to maintenance or supply chain without delaying production. Use LDMS to enforce standard work, capture countermeasures from Kaizen events, and track action closure. Keep meeting templates short and metrics-focused (OEE, scrap, customer delivery). Implement digitally where possible so remote leadership can see status without travel. LDMS reduces duplication of effort and harmonizes priorities across plants without adding headcount.

Standard Work

Develop and deploy Standard Work for operators, setup/changeover, and critical maintenance activities to reduce variability that causes scrap and downtime. Prioritize development for: line startup/shutdown, sanitation checks tied to food safety, changeover tasks for top SKUs, and PM tasks on critical assets.

Use one‑page visual SOPs in Spanish/Portuguese with photos and the exact sequence, takt‑based timing, and quality checkpoints. Validate with operators via Gemba—run a TWI or short shadowing program so the developer transfers knowledge and the operator owns the procedure. Enforce via LDMS audits and incorporate deviations into training and CAPA. Standard Work reduces operator-dependent variability quickly and enables faster onboarding of temps or new hires during peak demand without quality shocks.

Quality Management

Attack scrap by integrating quality management with maintenance and operations. First 60 days: map top scrap types by SKU/line and link to equipment states and changeovers.

Implement standard sampling and quick containment steps at the line to stop bad product from flowing (clear stop/go criteria for operators). Use SPC for high-volume/continuous processes and simple attribute checks for discrete operations. Close the loop: every significant scrap or quality hold requires CAPA that includes maintenance actions (e.g., update PM, retrofit guard). Align quality control plans with FSMS/ISO 22000 requirements common in the region to avoid audit issues. Train QA and ops on shared definitions and escalation—this reduces duplicate holds and late shipments. Small investments in poka‑yoke (jigs, sensors) where ROI is clear can eliminate recurrent defects without adding headcount.

Visual Management

Use shop‑floor visual management to make problems visible and accelerate response. Standardize a visual board template across plants showing daily OEE, downtime top‑3 causes, scrap rates, and status of critical corrective actions.

Boards should be bilingual where needed, updated at shift change, and include a color status (green/amber/red) tied to escalation rules. For mobile diagnostics, display line status at the line (traffic‑light LEDs) and provide a simple digital feed for the plant manager. Visuals reduce the time leadership spends hunting for problems and focus cross‑functional responses. Standard templates make cross‑plant benchmarking easier and support faster roll‑out of best practices.

5 Whys

Standardize a simple RCA protocol—use 5 Whys for every unplanned downtime >30 minutes or scrap event above a defined threshold. Require documented 5‑Why trees with corrective actions, owners, and verification dates entered into the CMMS or quality system within 48 hours.

Train a small group of facilitators (ops + maintenance + QA) to run fast, fact‑based sessions at the line (not in a conference room). Ensure that root causes feed back into PMs, Standard Work, and spare parts lists to prevent recurrence. Track closure and effectiveness (did the countermeasure reduce recurrence?) in LDMS reviews. This keeps investigations lean, timely and linked to operational controls rather than creating paperwork that doesn’t change the line.