Availability measures how much of your planned production time the line was actually running. It's the most visible OEE loss and typically the largest — but also the most improvable. This guide covers everything: the formula, all seven downtime categories, MTTR, MTBF, changeover analysis, CIP, start-up losses and how to build a downtime Pareto that drives real action.
Availability is the ratio of actual run time to planned production time. It captures every minute the line was stopped when it should have been running — whether planned or unplanned, equipment-related or operator-related.
Where: Actual Run Time = Planned Production Time − All Downtime
| Planned production time | 480 min (8 hrs) |
| Unplanned breakdown (conveyor jam) | − 35 min |
| Changeover (SKU A to SKU B) | − 28 min |
| Material shortage (film roll changeover wait) | − 12 min |
| Start-up stabilisation after changeover | − 8 min |
| Other / unclassified stop | − 7 min |
| Total downtime | − 90 min |
| Actual run time | 390 min |
| Availability (390 ÷ 480) | 81.25% |
Note: five separate downtime categories contributed to the 90 minutes lost. Without reason codes on each event, you'd never know where to focus improvement effort.
Every downtime event in a manufacturing environment falls into one of these seven categories. Capturing reason codes against each event is what turns raw downtime data into actionable insight. Without categorisation, you have minutes lost — with it, you have a prioritised improvement agenda.
Unplanned breakdowns are failures that occur during planned production time without warning — mechanical failures, electrical faults, sensor failures, component wear and structural failures. They are the most disruptive downtime category because they cannot be prepared for in the same way as planned events.
Common examples in food manufacturing:
How to track it: Capture start time, end time, equipment reference, fault description, action taken and technician. The combination of MTBF (frequency) and MTTR (repair duration) tells you whether the problem is recurring failures or slow repair response.
Mean Time Between Failures tells you how often a piece of equipment breaks down. A declining MTBF on a specific asset signals deterioration and should trigger a maintenance review before the failure becomes critical.
Planned maintenance includes scheduled preventive maintenance tasks, lubrication rounds, inspections, filter changes and calibrations that occur during planned production time. These are known events that the production plan should accommodate — but when they occur during planned run time rather than during scheduled downtime, they become Availability losses.
Common examples:
Improvement lever: Planned maintenance should ideally be scheduled during changeovers, CIP cycles or planned breaks rather than cutting into run time. Review your PM schedule against your production plan quarterly to minimise overlap.
A well-managed maintenance function should run at least 70–80% planned (preventive) vs 20–30% reactive. If your downtime Pareto shows more unplanned than planned maintenance time, your preventive programme is not keeping pace with equipment condition.
Changeovers are typically the largest single Availability loss on multi-SKU food manufacturing lines. A changeover is any activity that stops the line to transition from one product, format, pack size or allergen group to another. The time lost includes the physical changeover, cleaning, rinsing, reassembly, first-off checks and approval to run.
Types of changeover in food manufacturing:
SMED (Single Minute Exchange of Die) separates changeover tasks into two types:
Can only be done with the line stopped
Can be done while the line is still running
Converting internal tasks to external — and streamlining the remaining internal tasks — typically reduces changeover time by 30–50% without any capital investment. On a line doing 3 changeovers per day at 30 minutes each, a 40% SMED improvement recovers 36 minutes of Availability per day — around 130 hours per year.
Start-up losses are the time between the line being mechanically ready and the line achieving stable, conforming production output. This period is often underrecorded because the line is technically running — but it is producing rejects, running slowly or requiring frequent operator adjustment. For OEE purposes, the time spent stabilising after a start or changeover should be recorded as an Availability loss, separate from Performance or Quality losses during steady-state running.
What drives start-up losses in food manufacturing:
How to reduce start-up losses: Standardise start-up sequences with documented step-by-step checklists. Pre-set machine parameters to the correct product standard before starting. Track start-up duration as a separate KPI by line and product. Target progressive reduction — a 2-minute improvement across 3 daily starts recovers 30 minutes of Availability per week.
The rejects produced during start-up are Quality losses in OEE terms — but the time spent stabilising before accepting the first good unit is an Availability loss. Separating these two is important for directing improvement effort correctly. If you only see high start-up reject rates (Quality), you may miss the fact that start-up duration (Availability) is equally costly.
In food manufacturing, CIP (Clean In Place) and manual cleaning cycles are a mandatory and significant Availability loss. Unlike most downtime categories, cleaning cannot be eliminated — but it can be optimised, scheduled more effectively and reduced in duration through process improvement and chemical optimisation.
Types of cleaning downtime in food manufacturing:
How to optimise cleaning downtime: Review CIP chemical concentrations, temperatures and cycle times against actual soiling data — many sites are over-cleaning relative to their actual contamination risk. Allergen clean durations should be reviewed against verification data. Pre-clean preparation (stripping and staging parts before shutdown) reduces clean duration significantly.
Separate your cleaning events into at least three categories: scheduled cleaning (expected, planned), allergen cleaning (triggered by product schedule), and reactive cleaning (triggered by quality event or failure). This lets you see the true controllable vs uncontrollable cleaning burden and focus improvement effort on the right category.
Material shortage downtime occurs when the line is stopped because the materials needed to run — raw materials, packaging, labels, ingredients — are not available at the line when needed. This is distinct from equipment failure and is typically owned by planning, procurement or warehouse rather than engineering or production.
Common material shortage causes in food manufacturing:
Improvement approach: Material shortage downtime is a planning and logistics problem, not an engineering one. The fix is in visual management at the line (trigger points for replenishment before stock out), staging standards (what should be at the line, how much, when) and communication protocols between production and warehouse.
Every downtime tracking system needs a catch-all category for stoppages that don't fit the defined reason codes. "Other" is not a failure of your system — it's an essential safety valve that prevents operators from forcing events into the wrong category just to avoid leaving a field blank. However, the "Other" category needs active management.
What belongs in "Other":
The rule of thumb: If "Other" exceeds 10% of your total recorded downtime, your reason code structure needs a review. Regular monthly review of "Other" entries — looking for patterns and recurring descriptions — should feed back into the reason code library, migrating common "Other" events into defined categories over time.
Free-text comment fields are essential: Every "Other" entry should require a brief free-text description. This is the raw material for creating new reason codes. Over 3–6 months, patterns emerge from the free-text field that reveal systematic gaps in your category structure.
A growing "Other" proportion over time is a warning sign — either your reason code library is outdated, operators are using it as a shortcut, or new failure modes are emerging that your system isn't capturing. Treat it as a leading indicator, not just a residual bucket.
MTTR and MTBF work together to give you a complete picture of equipment reliability. MTBF tells you how often it breaks. MTTR tells you how quickly you fix it. Both are needed to understand Availability loss from breakdowns.
MTBF measures equipment reliability — how long, on average, it runs between failures. A high MTBF means reliable equipment that fails infrequently. A declining MTBF on a specific asset over time signals deterioration and should trigger a maintenance intervention before the situation worsens.
Example: A filler runs for 390 hours over a 4-week period and has 6 failures. MTBF = 390 ÷ 6 = 65 hours between failures. If the previous 4 weeks showed 80 hours between failures, the decline signals deteriorating condition.
MTTR measures maintenance responsiveness — how long it takes to restore equipment to working order after a failure. A low MTTR means fast diagnosis and repair. A high MTTR may indicate parts availability issues, skill gaps or poor fault diagnosis procedures.
Example: The same filler has 6 failures totalling 126 minutes of repair time. MTTR = 126 ÷ 6 = 21 minutes per repair. If 15 of those minutes are spent waiting for a spare part, improving spares availability would cut MTTR by 70%.
Availability from breakdowns can be estimated as: MTBF ÷ (MTBF + MTTR). In the example above: 65 ÷ (65 + 21) = 75.6% availability from this asset alone. Improving MTBF (fewer failures) or MTTR (faster repairs) both improve Availability, but the higher-leverage action depends on which is the bigger driver of lost time. If failures are rare but repairs are slow, focus on MTTR. If repairs are fast but failures are frequent, focus on MTBF through better preventive maintenance.
A downtime Pareto is the single most effective tool for improving Availability. It takes 4 weeks of downtime data and converts it into a prioritised list of where to focus improvement effort. In most factories, the top 3 reasons account for 60–75% of all downtime losses.
Every stop during planned production time needs a reason code, start time, end time, equipment reference and brief free-text description. No event should be left uncoded. The quality of your Pareto is entirely dependent on the quality of the data going in — partial capture produces a misleading picture.
A single week is not enough to identify true patterns — it may reflect unusual events that distort the picture. Four weeks gives you enough data to distinguish systematic losses from one-off occurrences. For lines with infrequent but high-impact failures, 8 weeks is better.
Aggregate all events under each reason code and calculate the total minutes lost per code over the analysis period. Also count the number of occurrences — a reason code with 50 short events tells a different story from one with 2 long events, even if total minutes are the same.
Rank reason codes by total minutes lost, highest first. Calculate each code's percentage of total downtime and running cumulative percentage. The point at which the cumulative line crosses 80% identifies your vital few — the categories that matter most.
The top 3 reason codes get owners, root cause analysis and time-bound improvement targets. Not all 7 categories simultaneously — focus produces results. Once the top 3 are reduced, the next Pareto cycle reveals the new top 3. This is continuous improvement in practice.
Downtime data should be reviewed daily — yesterday's top losses, who owns the actions, what happened. Weekly and monthly Pareto reviews track trends and verify whether improvement actions are working. The data only drives improvement if it's discussed, owned and acted on.
Availability targets vary significantly by sector and operating model. These benchmarks reflect typical ranges for well-run sites — not theoretical maximums.
| Sector | Typical Availability | Classification | Biggest Availability loss |
|---|---|---|---|
| Continuous single-product line | 88–95% | Good–World class | Unplanned breakdowns |
| Beverage / drinks manufacturing | 78–88% | Good | CIP cycles + format changeovers |
| Bakery & ambient food | 72–84% | Typical–Good | Changeover + planned maintenance |
| Multi-SKU chilled food | 65–80% | Typical | Changeover + CIP + allergen cleans |
| Fresh produce / salads | 60–76% | Typical | Allergen cleans + format changes |
| Ready meals (high SKU) | 58–74% | Below average–Typical | Allergen + CIP + recipe changeovers |
A fresh produce line running 20 allergen changeovers per week cannot be benchmarked against a single-product beverage line. Always benchmark within your own operating model first — your trend over 12 months is more useful than any external number. The question to ask is not "are we at 90%?" but "are we better than we were 6 months ago, and do we know why?"
Enter your shift data and get an instant OEE score with full Availability, Performance and Quality breakdown — plus the financial cost of your losses vs the 85% benchmark.
Availability = Actual Run Time ÷ Planned Production Time. Actual Run Time is Planned Production Time minus all downtime events regardless of cause. If a line is planned to run for 8 hours (480 minutes) and loses 90 minutes to downtime, Actual Run Time is 390 minutes and Availability is 390 ÷ 480 = 81.25%.
Planned Production Time is the total time the line is scheduled to run — from planned start to planned finish. It excludes scheduled breaks, meal times, planned non-production periods and time the line was not scheduled to run at all. The key principle is that Planned Production Time represents the commitment to produce — any time lost within that window is an OEE Availability loss.
Yes — if CIP or cleaning occurs during planned production time, it is an Availability loss. The only exception is if CIP is scheduled as a formal break in the production plan (i.e. not planned production time). In food manufacturing, CIP and allergen cleaning are often the largest single Availability loss categories, particularly on multi-SKU or allergen-complex lines.
MTBF (Mean Time Between Failures) measures reliability — how often equipment breaks. MTTR (Mean Time To Repair) measures maintenance responsiveness — how quickly you fix it. Both affect Availability. A declining MTBF indicates deteriorating equipment condition. A high MTTR indicates slow repair response, often driven by parts availability, skill gaps or poor fault diagnosis procedures.
Every downtime event should be assigned to one of the defined reason code categories — unplanned breakdown, planned maintenance, changeover, start-up loss, CIP/cleaning, material shortage, or other. The reason code should reflect the primary cause of the stop. Events that don't fit should go to "Other" with a free-text description — and the "Other" category should be reviewed monthly to identify patterns and create new specific categories.
SMED (Single Minute Exchange of Die) is a methodology for reducing changeover time by converting internal tasks (done with the line stopped) to external tasks (done while the line is running), and streamlining the remaining internal tasks. Applied correctly, SMED typically reduces changeover time by 30–50%. On a food line doing 3 changeovers per day at 30 minutes each, a 40% SMED improvement recovers approximately 130 hours of Availability per year.
Ideally below 10% of total downtime. If "Other" exceeds 10%, your reason code structure needs a review. Common causes of high "Other" proportions: operators using it as a shortcut rather than selecting the correct code, new failure modes not yet in the reason code library, or multi-cause events where a single code doesn't adequately describe the stop. Monthly review of "Other" free-text entries is essential to keep the category under control.
The fastest improvements come from: (1) Building a downtime Pareto over 4 weeks to identify the top 3 stop reasons — these typically account for 60–75% of all downtime. (2) Applying SMED to your highest-frequency changeover. (3) Reviewing your PM schedule to shift planned maintenance from production time to planned downtime windows. (4) Implementing staging standards for materials to eliminate shortage-related stops. Address the top Pareto item first — focused effort produces faster results than spreading improvement across all categories simultaneously.