What a Production Loss Analysis Is
Every manufacturing production line has a theoretical maximum output.
Given its design speed, its scheduled production time, and perfect execution with zero quality losses, zero unplanned stoppages, and zero speed reductions, it would produce a specific number of units in a specific period.
It never achieves that theoretical maximum in practice.
The gap between the theoretical maximum and the actual output is the production loss.
A production loss analysis is the structured process of understanding that gap in detail.
Not just that the gap exists.
Not just that OEE is 74% rather than 85%.
But specifically: of the 26 OEE points separating actual from target, how many came from unplanned machine failures, how many from changeover overruns, how many from micro-stops that operators could not log manually, how many from running below design speed on specific products, how many from quality defects at startup, and how many from quality defects during steady-state production?
Each of these categories has a different cause. Each has a different appropriate intervention. Each represents a different financial recovery opportunity.
Production loss analysis converts the headline OEE number into a specific, ordered list of opportunities that maintenance engineers, process engineers, and operations managers can act on with targeted interventions rather than general improvement effort.
Why Production Loss Analysis Matters
The most common manufacturing improvement failure is not lack of effort.
It is effort directed at the wrong losses.
A production team that identifies "machine downtime" as its primary OEE problem and launches a maintenance improvement program may be addressing the right category or the wrong one.
If the Six Big Losses Pareto for that line shows that unplanned equipment failures account for 45% of total loss and changeover losses account for only 15%, the maintenance improvement program is well-targeted.
If the same Pareto shows that changeover losses account for 45% of total loss and unplanned equipment failures account for only 15%, the maintenance improvement program is addressing the third most important loss category while the largest category receives no targeted attention.
A production loss analysis prevents this misallocation of improvement resources by making the relative magnitude of each loss category visible before the improvement program is designed.
The financial stakes of this visibility are significant.
For a production line generating 5 million euros annually at target OEE, a 10 OEE point gap represents approximately 500,000 euros of recoverable production value.
If improvement resources are directed at the loss category that accounts for 15% of that gap rather than the one that accounts for 45%, the improvement program is working on a 75,000 euro opportunity while a 225,000 euro opportunity goes unaddressed.
Production loss analysis ensures that improvement investment is proportionate to recovery opportunity.
The Six Big Losses Framework
The Six Big Losses framework is the standard categorization structure for production loss analysis in manufacturing.
Developed within the Total Productive Maintenance methodology and refined through decades of industrial application, it divides the total production loss into six categories that collectively account for every gap between theoretical maximum output and actual output.
The six losses are organized under the three OEE components.
Under Availability:
Unplanned equipment failures (also called breakdowns or equipment downtime losses). These are the unscheduled production stops caused by mechanical, electrical, or process failures that were not anticipated in the production plan. This is the loss category most commonly assumed to be dominant but often not the largest in practice.
Planned downtime losses (also called setup and adjustment losses). These cover the production time consumed by planned stops such as changeovers, product line changeovers, planned cleaning cycles, and planned maintenance windows. This category is frequently misunderstood as "already accounted for" because the stops are planned. In OEE analysis, planned downtime is typically excluded from the OEE denominator. Setup and adjustment losses within a planned window specifically cover the time between production stopping and the first good unit of the next run being produced.
Under Performance:
Minor stoppages and idling (also called small stops). These are brief production interruptions that are too short to be logged as formal downtime events but collectively represent significant lost production. A filling line that experiences 30 micro-jams per shift averaging 45 seconds each accumulates 22.5 minutes of Performance loss that does not appear in any downtime record if operators do not log events below a two-minute threshold.
Reduced speed losses. These are the losses incurred when a production line runs below its design speed or below its target speed for the specific product being produced. A packaging line running at 180 packages per minute on a product designed for 220 packages per minute is producing a 18% speed loss that does not appear as a downtime event and is invisible to operator logging systems unless speed is being monitored continuously from machine signals.
Under Quality:
Startup quality losses (also called startup rejects or yield losses at startup). These are the units produced between production start and the point at which the process achieves stable, on-specification output. Every production run, every shift start, and every changeover has a period during which the process parameters have not yet stabilized and the output does not meet quality specification.
Production quality losses (also called steady-state defects or in-process quality losses). These are the units produced during stable production that do not meet quality specification and must be scrapped or reworked. Unlike startup quality losses, these occur throughout the production run rather than only at the beginning.
How to Conduct a Production Loss Analysis
A production loss analysis follows a seven-step process that converts raw production data into a prioritized improvement agenda.
Step 1: Define the analysis period and scope
Select the production period to analyze. For a first production loss analysis, three to six months of data provides enough events for statistically meaningful Pareto analysis while remaining recent enough to reflect current operating conditions.
Define the scope: which production lines, which production areas, and which shifts will be included. Starting with a single production line or a single shift makes the first analysis more manageable and the learnings more specific.
Step 2: Collect and validate the production data
Gather the production data for the defined scope and period.
The data required includes total scheduled production time, actual production time, planned stop events with their categorization and duration, unplanned stop events with their fault code and duration, actual production rate relative to target rate by time period, and quality data including scrapped and reworked units.
Before proceeding to analysis, validate the data quality. Data from manual operator logs is likely to contain systematic gaps in micro-stop recording and speed loss capture. Data from machine-connected OEE monitoring is more complete but may contain sensor errors or miscategorized fault codes that need cleaning before analysis.
Data quality validation is not optional. A production loss Pareto built from incomplete data produces incorrect priority rankings that direct improvement effort toward the losses that were logged rather than the losses that occurred.
Step 3: Calculate the total OEE loss by Six Big Losses category
For each loss event in the dataset, assign it to one of the six categories and calculate the total production time equivalent of each category's losses.
Total production time lost to unplanned equipment failures: sum of all unplanned stop event durations.
Total production time lost to planned stops within production windows: sum of all changeover and setup adjustment durations.
Total production time lost to minor stoppages: this requires machine-connected data. Without continuous cycle time monitoring, minor stoppages below the operator logging threshold are invisible and their contribution to total loss is unmeasurable from manual records.
Total production time equivalent of speed losses: calculated from the difference between actual production rate and target production rate, integrated over time. A line running at 85% of target speed for 200 hours loses the equivalent of 30 hours of production at full speed.
Total production time lost to startup quality losses: calculated from the number of nonconforming units produced during startup periods multiplied by the target cycle time.
Total production time lost to steady-state quality losses: calculated from the total number of nonconforming units produced during stable production multiplied by the target cycle time.
Step 4: Build the Six Big Losses Pareto
With the total production time equivalent of each loss category calculated, build the Pareto chart that orders the six categories from largest to smallest.
The Pareto is the core output of the production loss analysis. It reveals the relative contribution of each loss category to the total OEE gap and identifies which category or categories account for the majority of recoverable production value.
In most manufacturing environments, two or three of the six categories account for 70 to 80% of total production loss. The Pareto makes this concentration visible in a format that directs improvement energy toward the dominant categories without requiring stakeholders to hold multiple data tables in their heads simultaneously.
Step 5: Drill into the dominant loss categories
For each loss category in the top 70 to 80% of the Pareto, conduct a more granular analysis that identifies the specific assets, products, shifts, or fault codes that account for the majority of loss within that category.
Within unplanned equipment failures: which specific assets generate the most downtime? What specific fault codes are most frequent on those assets?
Within speed losses: on which products does the line most frequently run below target? On which shifts is speed loss most common?
Within startup quality losses: which changeover types produce the longest startup loss period? Which products take longest to stabilize after a changeover?
This drill-down converts the category-level Pareto into asset-level and cause-level specificity that improvement interventions can be designed around.
Step 6: Quantify the financial value of each loss
Convert each loss category's production time equivalent into a financial recovery opportunity by multiplying by the fully-loaded production cost per hour or the contribution margin per unit not produced.
A production line with a fully-loaded cost of 400 euros per hour that loses 120 hours per month to unplanned equipment failures has a 48,000 euro per month recovery opportunity in that category.
A 50% reduction in that loss category through targeted reliability improvement would recover 24,000 euros per month, or 288,000 euros per year.
This financial quantification converts the production loss analysis from an operational report into a business case for targeted investment.
Step 7: Build the prioritized improvement agenda
Order the improvement opportunities by financial recovery potential per unit of improvement effort.
The highest-priority targets are the loss subcategories with the largest financial impact per improvement point that are also most responsive to known interventions.
An unplanned equipment failure loss that is concentrated on three specific bad actor assets is highly responsive to targeted reliability improvement because the scope is specific and the intervention is well-understood.
A speed loss that is distributed randomly across all products, all shifts, and all assets is less responsive to targeted intervention because the root cause is diffuse and the appropriate intervention is unclear.
The prioritized improvement agenda is the output that the production loss analysis delivers to the operations and maintenance teams as an actionable improvement plan.
Machine-Connected Data Versus Operator-Reported Data in Production Loss Analysis
The accuracy of a production loss analysis depends entirely on the accuracy and completeness of the underlying production data.
This makes the data collection method the most consequential decision in the entire analysis process.
Operator-reported data limitations
Manual OEE reporting systems where operators log production counts, downtime events, and reason codes at shift end or at the time of events produce data with systematic gaps.
Minor stoppages below the logging threshold are never captured. On a high-speed line with a two-minute logging threshold, every micro-stop under two minutes is invisible in the dataset. These events may collectively account for 15 to 25% of total production loss while appearing as zero in the analysis.
Speed losses are rarely captured in operator-reported systems. Operators do not continuously monitor production rate relative to target rate. They notice that the line is running slowly but rarely translate the observation into a logged speed loss event with accurate duration and magnitude.
Event durations are rounded and recalled imprecisely. An operator logging a 37-minute stoppage from recall at shift end will typically record 30 or 40 minutes. The 3 to 7 minute rounding error per event, accumulated across 40 events per month, represents meaningful distortion in the Availability loss calculation.
Machine-connected data advantages
Machine-connected OEE monitoring captures every production cycle, every stop event, and every speed deviation from machine signals in real time. The data is complete, accurate, and continuous.
Minor stoppages as brief as five seconds are captured as cycle time anomalies.
Speed deviations are calculated continuously by comparing actual cycle time to target cycle time for the specific product being produced.
Fault codes from PLC stop events provide specific failure mode categorization rather than generic operator-assigned reason codes.
The practical consequence of this difference is that the first machine-connected production loss analysis on a line previously analyzed from operator-reported data almost always reveals a significantly different loss profile than the manual analysis suggested. The losses that were invisible in manual data become visible in machine data. The Pareto rankings shift. The improvement priorities change.
Operating an improvement program calibrated against a manual data loss Pareto means directing improvement resources toward the losses that were visible in manual data rather than toward the losses that actually exist.
The Difference Between Production Loss Analysis and OEE Reporting
These two concepts are closely related but meaningfully distinct.
OEE reporting produces a number or a set of numbers that describes production performance in a defined period. OEE is 73% this week. Availability is 89%, Performance is 87%, Quality is 94%. The trend over the last 12 weeks is shown on a chart.
Production loss analysis uses those numbers as inputs to a deeper investigative process. Given that OEE is 73%, what specifically is causing the 27-point gap? In what proportions do the six loss categories contribute? Which specific assets, products, and fault codes drive each category? What is the financial value of each category's recovery opportunity? Which categories should be targeted first given the financial impact and the availability of effective improvement interventions?
OEE reporting is a measurement activity. Production loss analysis is an improvement targeting activity.
The most common mistake is stopping at OEE reporting and treating the OEE number as an output rather than as an input to the improvement targeting process. The OEE number reports that there is a problem. The production loss analysis explains what the problem is and where to focus the resources to address it.
Frequently Asked Questions
How often should a production loss analysis be conducted?
A formal production loss analysis should be conducted at two cadences.
A monthly operational analysis of the most recent four weeks of data supports the ongoing improvement program by identifying whether the loss profile is changing, whether improvement interventions are shifting the Pareto, and whether new loss sources are emerging that were not present in previous periods.
A quarterly strategic analysis covering the full quarter's data supports longer-term improvement planning and capital investment decisions by identifying trends that are not visible in monthly snapshots.
Can a production loss analysis be conducted without machine-connected OEE data?
Yes, but with significant limitations in accuracy and completeness.
A production loss analysis from operator-reported data can identify the loss categories that are large enough to be consistently logged by operators. It cannot reliably identify minor stoppages, speed losses, or the precise magnitude of any category that relies on accurate duration recording from recall.
For organizations that do not yet have machine-connected OEE monitoring, a production loss analysis from operator-reported data is better than no analysis, as long as the limitations of the data source are explicitly acknowledged and the results are treated as directionally indicative rather than precisely accurate.
What software is needed to conduct a production loss analysis?
The minimum requirement is production data in a structured format that can be analyzed. A spreadsheet application can support a basic production loss analysis from operator-reported data.
For machine-connected production loss analysis, an OEE monitoring platform that captures event-level production data and provides built-in Pareto analysis, Six Big Losses categorization, and drill-down capability from line level to asset level to fault code level significantly reduces the time required to move from raw data to prioritized improvement agenda.
The difference between conducting a production loss analysis manually in a spreadsheet and using a purpose-built OEE platform is primarily a time and refresh frequency difference. The manual approach takes days and produces a point-in-time snapshot. The platform approach produces the same analysis in minutes and can be refreshed at any time.
A production loss analysis does not improve production performance by itself. It identifies precisely where improvement will deliver the most value. The improvement program that follows it is more effective because it is working on the right losses in the right order, with improvement resources proportionate to recovery opportunity rather than to operational visibility.
Login
Blog
