OEE in Make-to-Order vs. Make-to-Stock Production: Guide to Optimization Strategies

OEE for make-to-order vs. make-to-stock manufacturing is a topic that most manufacturers are unaware of. A factory that produces to stock and a factory that produces to order may use the same machines and the same ERP system. But the way OEE impacts their performance, production planning, and profitability is fundamentally different.

Yet the majority of manufacturers apply the same production targets and improvement methods regardless of their model. This is a mistake.

An OEE of 75% on a high-variety, low-volume made-to-order line does not mean the same thing as 75% on a dedicated line producing the same item 24 hours a day.

This guide details what OEE actually measures in each context, where losses are hidden, and how to adapt your monitoring strategy to your production reality. Understanding this distinction is essential for any credible continuous improvement initiative.

Why the manufacturing model changes planning and production capacity

In make-to-stock production, manufacturing is driven by forecasts and the production schedule. Runs are long, changes are infrequent, and the main objective is to maximize throughput. OEE functions as a pure performance indicator: each point gained translates into additional volume to feed warehouses and ensure inventory management.

In make-to-order production, manufacturing responds to specific production orders. Runs are short, changes are frequent, and flexibility takes precedence over gross volume.

OEE is therefore an indicator of production capacity to meet deadlines rather than an indicator of volume.

The direct consequence is that the levers for improving OEE are not the same. Optimizing the pace on a made-to-order line without reducing changeover times means improving a figure without improving actual performance. The monitoring system must reflect this difference.

The operational complexity of a made-to-order environment is structurally greater than that of a made-to-stock environment. More references, more adjustments, more risk of error. Ignoring this reality when setting OEE parameters is like comparing apples and oranges.

OEE and resource availability: optimizing production from stock

Critical losses, yield rates, and quality in long runs

In production on stock, availability losses are the number one enemy. Every minute of unplanned downtime represents a number of parts not produced that should have been added to stock. Machine breakdowns, raw material delays, and preventive maintenance failures are the main causes.

Actual operating time is often less than what manual reports indicate. The gap between field perception and measured reality can reach 15 to 20 OEE points. This is a blind spot that distorts all downstream planning.

Performance losses come in second place. On long runs, micro-stops of 3 to 5 seconds that are repeated 80 times per shift go unnoticed but represent several hours lost per week. Actual cycle time deviates from theoretical cycle time, and no one notices this without automated measurement.

Quality losses have a multiplier effect: a batch of 10,000 parts with 3% scrap means 300 parts to be reproduced and a delayed schedule. The overall yield rate suffers directly as a result. The cost of non-quality is not limited to scrap: it includes machine time consumed, wasted material, and delays in subsequent production orders.

OEE strategy and inventory management: stability before absolute performance

The goal is to achieve a consistently high OEE. An average OEE of 78% that fluctuates between 60% and 90% is more problematic than a stable OEE of 72%. Why? Because planning is based on predictability.

If the logistics department cannot predict actual production to within 5%, safety stocks will skyrocket. Inventory turnover will collapse and cash tied up in inventory will become a financial drain.

Every percentage point of OEE variability translates into additional inventory costs.

The key indicators to monitor are the mean time between failures, the gap between actual and theoretical cycle times, and the batch rejection rate. Real-time monitoring is critical to detecting deviations before they impact inventory levels.

A stable OEE also makes upstream supplies more reliable. When actual capacity is known and predictable, supplier orders are more accurate, emergencies are less frequent, and express shipping surcharges disappear.

OEE in make-to-order production: controlling efficiency and delays on production lines

Critical losses and bottlenecks in short runs

In make-to-order production, the primary destroyer of OEE is changeover time. A line that changes reference 8 to 12 times per shift can lose 20 to 35% of its available time in adjustments. This is a structural bottleneck on production lines that will not disappear by purchasing a faster machine.

Start-up losses are the second critical source. After each changeover, the first few parts are often non-compliant. On small runs, these losses can represent 10 to 15% of total production. Actual manufacturing time is extended and delays accumulate order after order.

Traditional performance losses also exist, but they are more difficult to isolate because the theoretical cycle time changes with each reference. Without an automated system, the performance component of OEE is often inaccurate in make-to-order production.

An additional difficulty in make-to-order production is traceability. Each order has its own specifications, tolerances, and requirements. OEE tracking must be able to link performance data to each production order to identify problematic items.

OEE improvement strategy and cost reduction through SMED

In make-to-order production, the goal is not to maximize gross OEE but to maximize the time available to produce value. Reducing changeover times using the SMED method is the number one lever and a direct factor in cost reduction. Every minute saved on a changeover is an extra minute of production.

The key indicators are average changeover time, first-pass yield to reduce start-up losses, and the ratio of value-added production time to total time.

Overall OEE remains relevant as a trend indicator, but it is the detailed analysis of the causes of downtime that generates gains.

Rigorous control of manufacturing processes after each changeover reduces start-up scrap. The best results are achieved when operators have a validation checklist integrated into the tracking system.

A critical point: the measurement system must manage reference changes automatically. If each change requires manual intervention, operators will abandon the system. Technology must adapt to the field, not the other way around.

A classic mistake: applying production targets that are unsuited to the production process

Many manufacturers set a universal OEE target of 85%, inspired by benchmarks of excellence. This figure makes sense in stock production. It makes no sense in make-to-order production, where changes structurally consume 20 to 30% of the available time.

For example, an aerospace line with 10 changes per shift will never be able to reach 85%, no matter how efficient its operators are. Setting this target discourages teams and discredits the indicator.

The result: teams working to order are perceived as underperforming. Operators become demotivated when faced with an unattainable target.

The factory abandons OEE monitoring precisely where it would be most useful.

The right approach: a production plan with OEE targets tailored to the model. In make-to-order production, an OEE of 60% with a 2-point improvement per month is an excellent result. The important thing is the trajectory, not the absolute value.

Management must also review how OEE is communicated to teams. An achievable and contextualized target motivates. A target that is disconnected from the field demotivates. Understanding the context by all stakeholders is a prerequisite for success.

Impact of OEE on production planning and customer lead times

OEE and reliability of production planning

Production planning is based on a capacity assumption. If this assumption is wrong, the entire production plan collapses. In make-to-stock production, an overestimated OEE leads to shortages. In make-to-order production, it leads to contractual delays.

Integrating actual OEE into the planning process is a paradigm shift. Instead of planning based on 85% theoretical capacity, planning is based on 65% measured capacity. The result: fewer broken promises, fewer express shipments, and fewer penalties.

Companies that connect their OEE tracking to their ERP or manufacturing execution system see a 15-25% improvement in their on-time delivery rate within the first three months. This is a rapid and measurable return on investment.

OEE and customer satisfaction

Customer satisfaction in industrial B2B is mainly measured by the on-time and complete delivery rate. This rate is directly influenced by actual production capacity, i.e., OEE.

A factory that plans for a theoretical capacity of 85% when its OEE fluctuates between 55% and 65% will inevitably accumulate delays. Unmeasured production losses are the main factor in broken promises.

Demonstrating to a customer that you have improved your OEE is more powerful than any sales pitch. It is a measurable competitive advantage that builds trust and loyalty among customers.

Mixed environments: when make-to-order and make-to-stock production coexist

Many factories operate in mixed mode. Some lines produce long runs for stock, while others fulfill specific orders. Sometimes, the same line alternates between the two models depending on the period and the projects in progress.

Comparing the OEE of a stock line and a made-to-order line on the same dashboard without contextualization leads to absurd decisions.

The 62% made-to-order line that delivers 98% of its orders on time is more efficient than the 80% stock line that generates excess inventory.

Customer satisfaction is the final judge, not the raw OEE figure. OEE is a diagnostic tool, not a competition between lines.

Multi-site groups must integrate the production model into their OEE reference system using a shared database.

Internal comparisons only make sense if you are comparing like with like. Two sites producing to order can be compared with each other. Comparing a site producing to order with a site producing to stock is misleading.

Establishing a common reference system requires standardizing definitions: what is a planned shutdown? How should a change of series be classified? These questions seem simple, but their answers vary from site to site and distort any comparison.

The role of real-time monitoring: a performance indicator for each model

In production to stock, real-time monitoring allows performance deviations to be detected before they impact stock levels. A 5% drop in speed that goes undetected for a week on a high-speed line represents thousands of missing parts.

A Pareto chart of the causes of downtime allows corrective actions to be prioritized based on the most costly operating losses. This simple tool transforms raw data into a concrete action plan.

In make-to-order production, monitoring is even more critical because deadlines are contractual. If a changeover takes 45 minutes instead of 20, the impact is immediate. The operator who sees the delay in real time can alert the planning department.

Integration with a manufacturing execution system automates this reporting and adjusts resource availability in real time. The planning department no longer has to wait for the end-of-shift report to know that there is a problem.

In both cases, operators improve performance when they see the truth in real time. The return on investment for such a system can be measured in weeks, not months. The difference between manual tracking reconstructed after the fact and automated real-time tracking is the difference between reacting and anticipating.

Case studies: adapting OEE to the manufacturing process in the field

Production to stock: the Hutchinson case in the automotive industry

Hutchinson, an automotive equipment manufacturer producing to stock, increased OEE at one site from 42% to 75% by identifying downtime that was not detected by manual monitoring. The strategy was clearly focused on long runs: reducing downtime, stabilizing the pace of production, and ensuring reliable supply to manufacturers.

The gains were achieved in a matter of weeks thanks to the deployment of automated monitoring. The teams discovered that micro-stops, invisible in manual reports, were the primary source of loss. Without automated measurement, these losses would have remained hidden.

The impact on the supply chain was immediate: fewer delays, less express transport, and increased confidence on the part of the automotive manufacturers who place the orders.

Make-to-order production: aerospace and short runs

In aerospace, subcontractors operate on a build-to-order basis with high value-added parts. A gross OEE of 45 to 55% is normal. The challenge is to reduce changeover times and start-up scrap, not to chase an unrealistic figure.

The most effective improvement projects in this sector target the standardization of series change procedures. Every minute saved on a change translates directly into additional capacity to fulfill orders.

Mixed environment: Nutriset and the dual humanitarian challenge

Nutriset illustrates the dual challenge of mixed environments: reliability in production from stock for humanitarian reserves, responsiveness to orders for urgent needs in crisis areas. In the humanitarian context, every day of delay has direct consequences on the ground.

These examples show that the key is not the OEE figure itself, but understanding the losses it reveals and the ability of teams to act on the right causes.

How to configure your OEE monitoring organization according to your model

Step 1: Identify the production model for each line. A single factory may have both stock lines and order-to-order lines. The OEE monitoring settings must reflect this reality. This diagnostic step is fundamental and must not be rushed.

Step 2: Define the theoretical cycle times by reference for made-to-order lines. Without this basis, the performance component of OEE is meaningless. Modern IoT solutions automatically manage reference changes.

Step 3: Separate changeover time from downtime causes. In make-to-order production, changeovers are not anomalies, they are an operational reality. Measuring them accurately is a prerequisite for improving them through SMED.

Step 4: Adapt dashboards. Stock lines display overall OEE and throughput trends. Make-to-order lines display average changeover time and on-time delivery. Each model has its own priority indicators.

Step 5: Deploy ready-to-use automated monitoring that captures data directly from the machines in two hours, without modifying the infrastructure. Ease of deployment is a key factor in adoption by field teams.

Step 6: Train teams to read OEE dashboards. A tracking tool is only valuable if operators and line managers know how to interpret the data and take corrective action. Training is an investment, not a cost.

FAQ: OEE in make-to-order vs. make-to-stock production

Conclusion: OEE is not a number, it is a diagnosis

OEE is only valuable in the right context. In stock production, it is an indicator of throughput and consistency. In made-to-order production, it is an indicator of flexibility and on-time delivery. Applying a single grid to both environments means missing the real levers.

What matters is understanding where losses occur, why they exist, and how to reduce them. Real-time monitoring provides this visibility.

Field teams that see the truth make the right decisions, regardless of the production model.

Properly interpreted OEE transforms the relationship between production, logistics, and management. It replaces intuition with facts, accusations with diagnoses, and promises with commitments kept.

TEEPTRAK deploys ready-to-use IoT solutions that measure OEE in real time, adapted to made-to-order, made-to-stock, and mixed production environments. Installation takes two hours, requires no infrastructure changes, and includes native management of reference changes. More than 400 factories in 30 countries trust our solutions. Request a demonstration.