Manufacturing & custom software · Bengaluru · Whitefish, USA
Manufacturing

The Six Big Losses in OEE: What They Are, How to Measure Them and How to Reduce Them

A practical guide to the six big losses behind OEE — breakdowns, setups and adjustments, small stops, reduced speed, start-up rejects and production rejects — with examples, how to record each on the shop floor, and proven ways to reduce them.

By Corposol Team··6 min read
CNC press brake in a sheet metal fabrication shop
Photo: Kantmaschine LVD PPS 200-30 02 by Asurnipal, CC BY-SA 4.0 (cropped)

OEE tells you how much of your planned production time is truly productive. The six big losses tell you where the rest went. The framework comes from Total Productive Maintenance (TPM) and remains the most practical way to organise improvement on the shop floor. If you need the basics of OEE first, read what is OEE and how to calculate it.

The six losses and the OEE factor they reduce

OEE factorLossTypical examples
Availability1. Breakdowns (equipment failure)Hydraulic failure, motor burnout, tool breakage requiring repair
2. Setup and adjustmentsChangeover, die or mould change, first-off approval, warm-up
Performance3. Small stops (idling and minor stoppages)Jammed feeder, blocked sensor, waiting for operator for a minute or two
4. Reduced speedRunning below ideal cycle time due to wear, conservative settings, material variation
Quality5. Start-up rejectsScrap and rework while the process stabilises after a start or changeover
6. Production rejectsScrap and rework during stable production

Loss 1: Breakdowns

What it is: unplanned stops when equipment fails and needs repair.

How to record it: every stop longer than a threshold (often five or ten minutes) is logged with a breakdown reason and linked to a maintenance work order. Time from stop to restart is the loss; MTTR and MTBF follow from the records — see how to calculate MTTR and MTBF.

How to reduce it:

  • Preventive maintenance on critical machines, with compliance tracked — see TPM.
  • Autonomous maintenance by operators: cleaning, inspection, lubrication.
  • Root cause analysis on repeat failures using 5 Why.
  • Critical spares stocked and identified — see spare parts criticality.
  • Faster response: breakdown alerts to maintenance phones the moment a machine stops.

Loss 2: Setup and adjustments

What it is: time from the last good part of one job to the first good part at full speed of the next — tool and die changes, programme loading, material change, first-off inspection.

How to record it: operators mark setup start and end on a tablet, or the system detects the gap between jobs. Record setup by machine and product pair, because some changeovers are much longer than others.

How to reduce it:

  • SMED: move work done while the machine is stopped to before or after the stop (preparing tools, materials and fixtures in advance).
  • Standardised setup sheets and quick-release clamping.
  • Sequence jobs to reduce changeovers — similar materials, colours or tooling together — using the production schedule.
  • Faster first-off approval with digital inspection records.

Loss 3: Small stops

What it is: brief interruptions, usually under five minutes, often fixed by the operator: a jammed part, a misfeed, a sensor blocked by chips, waiting for a crane.

How to record it: manual recording is unrealistic. Machine signals reveal small stops as short idle periods between cycles. Many plants are surprised by how much time they add up to.

How to reduce it: observe the machine for a full shift, list each small stop and its cause, then fix the top causes — feeders, chutes, sensor positions, chip management, material presentation.

Loss 4: Reduced speed

What it is: running slower than the ideal cycle time. Causes include worn equipment, conservative feeds and speeds, operator inexperience, and material variation.

How to record it: compare actual output against ideal cycle time × run time. This requires an agreed ideal cycle time per product and machine — the best demonstrated rate or the design rate, not the standard used for costing.

How to reduce it: find out why the machine runs slower — sometimes it is a quality workaround that can be solved differently — restore worn equipment, optimise programmes, and train operators.

Loss 5: Start-up rejects

What it is: scrap and rework produced while the process stabilises after a start, changeover or break — moulds warming up, first coils of a new batch, initial adjustments.

How to record it: separate start-up rejects from production rejects in the reject entry, with the job and changeover they followed.

How to reduce it: standard start-up parameters, pre-heating, better setup accuracy, and process recipes stored and recalled rather than adjusted by trial.

Loss 6: Production rejects

What it is: scrap and rework during stable production.

How to record it: rejects by defect, machine, product, shift and operator, captured at inspection points with in-process checks — see in-process quality checks in MES.

How to reduce it: Pareto analysis of defects, root cause and corrective action through CAPA, poka-yoke (error-proofing), and process capability studies. See how to reduce rejection rate.

Building a loss tree for your factory

The six big losses are categories. Under each, define a small number of specific reasons that apply to your machines. For a press shop, setup reasons might include die change, coil change and first-off approval; for an injection moulding shop, mould change, colour change and material drying. Keep the list short — operators must choose the right reason in seconds. See designing downtime reason codes.

Making losses visible every day

  1. Capture — machine signals for time and counts; operators for reasons and rejects.
  2. Show — a daily loss chart by machine and category, ranked by minutes lost.
  3. Review — a 15-minute meeting each morning on yesterday's top three losses.
  4. Act — assign an owner and date to each action.
  5. Verify — check the loss falls in the following weeks.

Software makes the first two steps automatic. OEE software and machine monitoring collect the data and draw the charts; the improvement still comes from the people in the morning meeting.

A worked example

A CNC turning cell runs two shifts of eight hours with a 30-minute break in each: planned production time is 900 minutes. One day, the cell records 60 minutes of breakdown, 90 minutes of setups, 750 minutes of run time, 1,100 parts at an ideal cycle of 36 seconds, and 33 rejects (12 at start-up, 21 during production).

  • Availability = 750 / 900 = 83.3%
  • Performance = (1,100 × 0.6 min) / 750 = 660 / 750 = 88.0%
  • Quality = (1,100 − 33) / 1,100 = 97.0%
  • OEE = 0.833 × 0.880 × 0.970 ≈ 71.1%

The losses in minutes: breakdowns 60, setups 90, small stops and reduced speed together 90 (750 − 660), and rejects about 20 minutes of ideal time (33 × 0.6). Setups are the largest single loss — so SMED on this cell is the first project, not maintenance.

Mapping the six losses to who acts on them

LossUsually owned byTypical tools
BreakdownsMaintenancePM, autonomous maintenance, root cause analysis, spares
Setup and adjustmentsProduction engineering, settersSMED, setup sheets, sequencing
Small stopsProduction, maintenanceObservation, minor fixes, 5S
Reduced speedProcess engineeringProgramme optimisation, equipment restoration
Start-up rejectsProduction, qualityStandard parameters, first-off approval
Production rejectsQuality, productionPareto, CAPA, poka-yoke, SPC

Assigning ownership by loss category prevents the common situation where OEE is everyone's metric and no one's responsibility.

Common mistakes when using the six big losses

  • "Others" as the largest category — the reason list does not match reality; revise it.
  • Recording only long stops — small stops disappear into performance loss without explanation.
  • Using costing standards as ideal cycle times — hides speed loss.
  • Counting planned downtime as a loss — or, conversely, calling unplanned waiting "planned".
  • Working on all losses at once — focus on the biggest loss at the bottleneck.

Frequently asked questions

Which of the six big losses is usually the largest?

It varies by factory, but in many small and mid-sized plants, setups and changeovers, small stops and unrecorded speed losses together are larger than breakdowns. They are just less visible, because no one writes them down.

How do we measure small stops?

Small stops are hard to capture by hand. Automatic machine signals — a cycle count or running signal — reveal them as the gap between ideal and actual output, or as short idle periods between cycles.

Are planned breaks and maintenance part of the six big losses?

Usually not. Planned stops are excluded from planned production time before OEE is calculated. Some companies track them separately as part of TEEP or overall utilisation.

Where should we start reducing losses?

At the bottleneck machine, with the single largest loss category from a few weeks of reliable data. Fixing one loss well builds the habit and the credibility for the next.

  • Six big losses
  • OEE
  • Downtime
  • TPM
  • Lean manufacturing

Let's discuss your factory's software

Tell us about your process or idea. You'll get a free consultation and a clear, fixed quote — no obligation.

  • Free consultation
  • Prototype before you commit
  • Fixed, itemised quote
  • You own the software

Get a free quote

We reply within one working day. Your details are never shared.