OEE (overall equipment effectiveness) is the share of planned production time that a machine spends making good parts at its rated speed. It is availability times performance times quality, so a line that scores 90% on each is 73% overall. Measured across 884 machines at 23 companies, the average was 65%; across 3,500 machines connected to one monitoring vendor, most plants sat at 55 to 60%. The 85% “world class” figure has no published sample behind it. No calculator here: the calculation takes four lines.
In this guide
What is OEE in manufacturing?
What is the OEE formula?
How do you calculate OEE, step by step?
What is a good OEE score, and where do the benchmarks come from?
What are the six big losses?
What is the difference between OEE and TEEP?
How does OEE get inflated, and how do you keep it honest?
What does low OEE cost a plant?
How is OEE data collected, and why do most plants still use spreadsheets?
Which OEE software should a plant compare?
How much does OEE software cost?
Which standards define OEE?
How do plants raise OEE once they can measure it?
Where does Morsa fit with OEE?
What changed at plants running Morsa?
FAQ
Sources, Changelog, Related pages
What is OEE in manufacturing?
OEE is a total productive maintenance measure of how much of a machine’s planned time turns into good output. The Lean Enterprise Institute defines it as “a total productive maintenance (TPM) measure of how effectively equipment is being used,” calculated as “Availability Rate x Performance Rate x Quality Rate.” Seiichi Nakajima introduced the measure as part of TPM in the 1980s; the sources disagree on the exact year, so this page does not state one. What TPM is, and how OEE sits inside it, is in total productive maintenance.
Vorne, the vendor whose oee.com site most people land on, puts it as “the percentage of manufacturing time that is truly productive.” That is the useful reading: OEE is not a machine’s utilization, and it is not a plant’s throughput. It is the fraction of the time you planned to run in which the machine ran, at speed, making parts that passed. A plant with a high OEE on a machine nobody needs has a different problem, which is why the measure belongs on the constraint first.
What is the OEE formula?
OEE equals availability multiplied by performance multiplied by quality, each expressed as a fraction of the one before it.
Factor | The Lean Enterprise Institute’s definition | Formula |
|---|---|---|
Availability | “Measures downtime losses from equipment failures and adjustments as a percentage of scheduled time” | Run time / planned production time |
Performance | “Measures operating speed losses” | (Ideal cycle time x total count) / run time |
Quality | “Expresses losses due to scrap and rework as a percentage of total parts run” | Good count / total count |
OEE | The product of the three | Availability x performance x quality |
The three multiply, which is why OEE scores look low to people who expect an average. Vorne’s arithmetic: “if all three factors are 90%, the resultant OEE will only be 73%.” The Lean Enterprise Institute’s example runs 90% availability, 95% performance and 99% quality to 84.6%.
There is a one-line form that gives the same answer and skips the components: OEE equals good count times ideal cycle time, divided by planned production time. Vorne states it as “(Good Count × Ideal Cycle Time) / Planned Production Time”. Use the three-factor form to find where the loss is and the one-line form to check the arithmetic. If the two disagree, a count or a time base is wrong, which is the subject of the honesty section below.
How do you calculate OEE, step by step?
Take one shift on one machine and write down four numbers: planned production time, downtime, total count and good count, plus the machine’s ideal cycle time. The example below is illustrative.
1. Planned production time
Input: 8-hour shift less a 30-minute planned break
Calculation: 480 min minus 30 min
Result: 450 min
2. Availability
Input: 60 min of unplanned stops (a changeover overran, a jam)
Calculation: Run time 390 min / 450 min
Result: 86.7%
3. Performance
Input: Ideal cycle time 0.5 min per part; total count 660
Calculation: (0.5 x 660) / 390 min = 330 / 390
Result: 84.6%
4. Quality
Input: Good count 620 of 660
Calculation: 620 / 660
Result: 93.9%
OEE
Calculation: 0.867 x 0.846 x 0.939
Result: 68.9%
Check, one-line form
Calculation: (620 x 0.5) / 450 = 310 / 450
Result: 68.9%
ONE SHIFT, FOUR LINES
How to calculate OEE without a calculator
STEP 01
Planned production time
Shift length minus planned breaks and planned maintenance. Do not subtract anything else.
THE EXAMPLE
480 min less a 30 min break = 450 min
STEP 02
Availability
Run time divided by planned production time. Every unplanned stop counts, however short.
THEN
60 min of stops: 390 / 450 = 86.7%
STEP 03
Performance
Ideal cycle time times total count, divided by run time. The ideal is the machine's rated speed, not last month's average.
THEN
0.5 min x 660 parts = 330 min of work in 390 min run time = 84.6%
STEP 04
Quality
Good parts divided by total parts. Rework counts as not good the first time.
THEN
620 / 660 = 93.9%. OEE = 0.867 x 0.846 x 0.939 = 68.9%
Illustrative figures. Check: 620 good parts x 0.5 min / 450 min = 68.9%.
Three things trip plants up. Planned production time should exclude only what was planned; an unplanned stop that gets reclassified as “planned maintenance” after the fact raises availability without changing anything. Ideal cycle time is the machine’s rated speed, and if performance comes out above 100% the ideal is wrong, not the machine fast. And rework is a quality loss the first time through, even if the part ships later. Evocon’s own calculation guide runs a 252-hour example to 53.2% using the same three factors, and Limble’s runs one to 58.3%; the method is the same on every vendor’s page, which is why a calculator adds nothing.
What is a good OEE score, and where do the benchmarks come from?
A good OEE score is one that is measured honestly and going up. The numbers people quote as benchmarks come from very different places, and no page that ranks for this question puts the sample next to the figure. This one does.
Average measured OEE, peer-reviewed
Figure: 65%
Sample and method: Raw data from 23 companies operating 884 machines, from automatic OEE systems rather than self-report
Source, date: Hedman, Subramaniyan and Almström, Procedia CIRP 57, 2016
Where most plants sit
Figure: 55 to 60%
Sample and method: 3,500+ machines connected to Evocon across 50+ countries, May 2023 to June 2024
Source, date: Evocon, July 2024, updated June 2026 (vendor, method stated)
Share of plants at 85% or above
Figure: About 6%
Sample and method: Same 3,500+ machines
Source, date: Evocon, same
“World class”
Figure: 85%
Sample and method: None stated. “It is often thought that a World-Class OEE score is 85%,” traced to Japanese PM Prize winners
Source, date: Vorne, oee.com (vendor, no method)
“Typical”
Figure: About 60%; more plants under 45% than over 85%
Sample and method: None stated
Source, date: Vorne, same
Red flag
Figure: Over 90%
Sample and method: Practitioner judgment: “any score above 90% deserves serious scrutiny”
Source, date: Evocon, May 2025
BENCHMARKS, WITH THEIR SAMPLES
What is a good OEE score, and where do the benchmarks come from?
65%
average OEE measured from raw data across 884 machines at 23 companies
Procedia CIRP, 2016, peer-reviewed
55 to 60%
where most plants sit across 3,500+ connected machines in 50+ countries, May 2023 to June 2024
Evocon, vendor data, method stated
6%
of plants in the same dataset score 85% or above
Evocon, same dataset
85%
the 'world class' figure quoted on most OEE pages, with no published sample behind it
Vorne, oee.com, undated
READ TOGETHER
The measured averages agree with each other, and the famous benchmark is the one nobody measured. Benchmark against your own machine last quarter.
Hedman et al. abstract verified via the OpenAlex record for the DOI. Evocon's per-industry figures exist only as a chart image and are not reproduced here.
The two measured figures agree, and the famous one is unmeasured. Wikipedia’s OEE article carries a section titled “The ‘85% is World Class’ myth” and calls the assumption “in many cases incorrect,” without offering a source for the alternative either. The honest use of a benchmark is to check whether your number is plausible, then to benchmark the machine against itself.
What are the six big losses?
The six big losses are the categories TPM uses to sort every OEE loss into a factor, so a low score points at something a team can fix. Vorne’s list, which follows Nakajima, maps them as follows.
Loss | OEE factor it reduces | What it looks like on the floor |
|---|---|---|
Equipment failure | Availability | Breakdowns, unplanned stops |
Setup and adjustments | Availability | Changeovers, tool changes, warm-up |
Idling and minor stops | Performance | Jams, misfeeds, sensor trips, “typically a minute or two” |
Reduced speed | Performance | Running below rated cycle time |
Process defects | Quality | Scrap and rework during steady-state running |
Reduced yield | Quality | Scrap from startup to stable running |
The measured problem is not the categories but the recording. In the 884-machine study, “Almost half of the recorded OEE losses could not be classified since the loss categories were either lacking or had poor descriptions,” and “90% of the stop time that was classified could be directly related to supporting activities performed by operators and not the automatic process itself”: waiting for material, a tool, a setter, a decision. That second finding moves the cause of most availability loss off the machine and onto coordination, which is the same finding total productive maintenance reaches from the maintenance side.
What is the difference between OEE and TEEP?
OEE measures productive time against planned production time; TEEP (total effective equipment performance) measures it against all time, 24 hours a day, 7 days a week. Vorne’s definition: TEEP equals OEE times utilization, where “Utilization = Planned Production Time / All Time.” OEE answers “how well did we run when we planned to run”; TEEP answers “how much of this asset’s capacity are we using at all.”
Vorne’s worked example: a machine at 65% OEE on two eight-hour shifts, five days a week, is planned for 80 of 168 hours, so utilization is 47.62% and TEEP is 30.95%. A plant manager deciding whether to add a shift or buy a second machine wants TEEP, because it shows that the existing machine has half its hours unplanned. A supervisor trying to get through tonight wants OEE. The number an OEE vendor puts on a dashboard is usually OEE, and a board asking about capital usually means TEEP. OEE is not schedule attainment: a machine can run well on the wrong job.
How does OEE get inflated, and how do you keep it honest?
OEE gets inflated by moving the time base and lowering the ideal, and it stays honest when the inputs come from the machine rather than from the shift report. Evocon, which sells the monitoring, documents the tricks: excluding scheduled maintenance and ignoring breaks can inflate OEE by “5-15% or more, without any real performance gains” (the vendor’s wording); understated speed baselines produce “performance” scores of “110% or more”, which the vendor calls “a physical impossibility”; and once a plant connects machines, “real OEE is often up to 50% lower than what the factory had been reporting.”
MachineMetrics puts the same point in a heading on its product page: “Every plant reports an OEE number. Few can defend how they got it.” Peeter Põder, Mill Manager at Toftan 2, said it from the plant side in Evocon’s case study (a vendor case study): “If you know why a machine is not working, then you can find a solution. If you assume you know why a machine is not working, you are playing a game of hit ‘n’ miss.”
Four rules keep the number defensible: planned production time changes only for things planned before the shift; the ideal cycle time is the rated speed and is written down; every stop over a minute gets a reason code, or the study above repeats itself; and the one-line check is run on every shift’s figures.
What does low OEE cost a plant?
It costs more in lost sales than in repair time, and the reacting is getting slower. NIST’s survey of US discrete manufacturers (AMS 100-34, June 2020, 2016 data) put preventable maintenance losses at “$119.1 billion: $18.1 billion due to downtime, $0.8 billion due to defects, and $100.2 billion due to lost sales from delays and defects,” and found the quarter of plants most reliant on reactive maintenance had “3.3 times more downtime” and “16.0 times more defects” than the least reliant. Its estimate of the upside from more predictive maintenance was $73.8 billion, mostly in sales. Breakdown losses in availability are what predictive maintenance software is built to cut.
Siemens’ True Cost of Downtime 2024, 181 interviews at large plants, found “an average of 25 downtime incidents a month per facility, down from 42 in 2019,” still “27 hours a month,” and that recovery per incident rose from 49 to 81 minutes. Its per-hour costs run from “$36,000 in Fast Moving Consumer Goods” to “$2.3 million in the Automotive sector.” Siemens owns a predictive maintenance vendor and its $1.4 trillion headline is, by its own note, “extrapolated,” so treat the totals as vendor research with a stated sample and the per-plant figures as the useful ones. Plant Engineering cites a rule of thumb from the vendor SmartSights: “for a $1 billion company, every 1% improvement in OEE is worth approximately $7 million annually.” Scale it: for a $50 million plant, a point of OEE is roughly $350,000 a year, before anyone argues about the assumptions.
How is OEE data collected, and why do most plants still use spreadsheets?
Four ways, and most plants still use the fourth. Machine signals come from the PLC over OPC UA or a vendor driver; CNC machines expose MTConnect or Fanuc Focas, which is how Datanomix connects DMG Mori, Haas, Okuma, Mazak and others with “No Operator Input Required”; a current or vibration sensor clamped onto a machine gives run/stop without touching the control, which is Guidewheel’s and Amper’s approach; and an operator keys counts and downtime reasons into a tablet or a sheet.
The sheet is the incumbent. IoT Analytics reported in December 2025 that “54% of small- and medium-sized plants use some combination of pen & paper or spreadsheets as their manufacturing execution system” and “just 8% of plants use a commercial MES today.” Rockwell Automation’s July 2026 survey of 1,560 manufacturers in 17 countries found that 93% have an MES but only 23% have it fully integrated, and that “43% acknowledge they are not effectively using their collected data.” Lorenzo Veronesi, Associate Research Director at IDC, said of that survey: “organizations risk leaving significant value on the table if disconnected systems and underutilized data go unaddressed.” Evocon, which gives away a free OEE spreadsheet, says of it that Excel “often leads to delays, inaccuracies, and significant time loss.” Which system holds the floor record, and where OEE lives in it, is covered in what is an MES system. The tools that produce an OEE number, with prices, are compared in shop floor management software.
Which OEE software should a plant compare?
Compare by how the data gets in, then by price. Every entry below carries a fact quoted from the vendor’s own page on 20 September 2026 and the price the vendor publishes; “no public price” means the vendor publishes none. For machine monitoring specifically, with a 10-machine cost table, see best machine monitoring software.
Evocon
Vendor: Evocon
How data gets in: IIoT device on the machine, PLC signals, ERP and CMMS integrations
What the vendor says: License includes “unlimited users,” device firmware updates, implementation and training support
Published price: Per machine per month: Basic $219 (1-year) or $189 (3-year); Professional $289 or $249; Enterprise $379 or $319; IIoT device $24 or $19
Vorne XL
Vendor: Vorne Industries
How data gets in: Sensor inputs plus network; “no software to install, no changes to your PLC code”
What the vendor says: “No recurring costs”; all models run identical software
Published price: One-time: XL Touch $4,990; XL HD+ $4,690; XL810-1 $4,490; 90-day trial
FourJaw
Vendor: FourJaw
How data gets in: MachineLink IoT device per machine, optional 4G gateway
What the vendor says: Minimum deployment 5 machines; hardware included on Pro
Published price: Per machine per month: Standard £90 monthly or £72 annual; Pro £180 or £144; MachineLink £200 one-time on Standard
Worximity
Vendor: Worximity
How data gets in: Per-machine install, “usually takes 1 hour per machine”
What the vendor says: 30-day trial
Published price: $300 per machine across Starter, Pro and Enterprise; monthly, 1-year and 3-year terms
Tulip
Vendor: Tulip Interfaces
How data gets in: Edge devices, GPIO; Machine Monitoring is an add-on that maps tags to attributes
What the vendor says: “Connect machines to Tulip, map tags to attributes, and determine machine state”
Published price: Per interface per month, billed annually, 10-interface minimum: Essentials $100; Professional $250; Enterprise custom
Ignition (with an OEE module)
Vendor: Inductive Automation
How data gets in: OPC UA and PLC drivers; sold per server
What the vendor says: “Standard Ignition and Ignition Edge licenses have a one-time cost, as they are both perpetual licenses”
Published price: Perpetual per server: Application Building Suite $13,500; Industrial Historian $3,500; support 16% to 24% of retail a year; OEE module priced separately
MaintainX
Vendor: MaintainX
How data gets in: CMMS; IoT sensor integrations on Enterprise; meter-based analytics on Premium
What the vendor says: A CMMS, not an OEE product; “Asset Health Insights” on Enterprise
Published price: Per user per month: Basic $0; Essential $20 annual or $25 monthly; Premium $65 or $75; Enterprise custom
MachineMetrics
Vendor: MachineMetrics
How data gets in: Machine connectivity and APIs
What the vendor says: “Pricing is volume based so the more machines you connect, the less it costs per machine”
Published price: No public price
Datanomix
Vendor: Datanomix
How data gets in: MTConnect and Fanuc Focas; direct integrations with DMG Mori, Haas, Okuma, Mazak, Makino and others
What the vendor says: “No Operator Input Required”; “Automated Downtime Insights + OEE”
Published price: No public price
Guidewheel
Vendor: Guidewheel
How data gets in: Non-intrusive sensor; ethernet, WiFi or LTE
What the vendor says: “We price per machine and per site”
Published price: No public price
Amper (now part of ECI Solutions)
Vendor: ECI Solutions
How data gets in: Non-invasive sensors, aggregators, PLC integrations
What the vendor says: “Setup takes minutes and requires no interruption to production”
Published price: No public price
Sight Machine
Vendor: Sight Machine
How data gets in: Connects “every OT and IT system” (controls, historians, MES, ERP “and more”) to a semantic model
What the vendor says: Claims “10%+ output gains” with deployment in “days”
Published price: No public price
Oden Technologies
Vendor: Oden
How data gets in: Not specified on the page
What the vendor says: “+21.4% OEE” at one customer
Published price: No public price
TrakSYS
Vendor: Parsec
How data gets in: PLCs, equipment and enterprise systems
What the vendor says: “Measure and improve OEE and plant productivity with real-time performance tracking”
Published price: No public price
Redzone
Vendor: QAD Redzone
How data gets in: Frontline platform across production, maintenance and quality
What the vendor says: “Get a 26% productivity increase in just 90 days with Redzone” (no method published)
Published price: No public price
Opcenter
Vendor: Siemens
How data gets in: MOM portfolio; Opcenter Intelligence is the analytics module
What the vendor says: “A unified manufacturing operations management (MOM) solutions portfolio”; OEE not named on the overview page
Published price: No public price
Limble
Vendor: Limble
How data gets in: CMMS; no machine-data layer named
What the vendor says: Standard, Premium+ and Enterprise tiers; OEE not named in any tier
Published price: No public price; price calculator only
Excel or paper
Vendor: Any
How data gets in: Manual entry
What the vendor says: The incumbent at 54% of small and mid-size plants
Published price: Free
Not listed: Rockwell Plex and Sepasoft’s OEE module for Ignition, whose pages could not be read on 20 September 2026; OEEsystems, whose domain now redirects to MaintMaster. Amper’s pricing page now redirects to ECI Solutions, which acquired it.
How much does OEE software cost?
From about $2,300 a year for one machine to a one-time $4,500, with most enterprise tools quote-only. Published prices on 20 September 2026:
Model | Range | Examples |
|---|---|---|
Per machine per month | $189 to $379, or £72 to £180 | Evocon, FourJaw; Worximity $300 |
One-time per machine | $4,490 to $4,990, no recurring cost | Vorne XL |
Per interface per month | $100 to $250, 10 minimum | Tulip |
Perpetual per server | $3,500 to $13,500 plus 16% to 24% support | Ignition |
Per user per month | $0 to $75 | MaintainX (a CMMS with OEE-adjacent analytics) |
Quote only | MachineMetrics, Datanomix, Guidewheel, Amper, Sight Machine, Oden, TrakSYS, Redzone, Opcenter, Limble |
Two cost checks before buying. A ten-machine cell on Evocon Basic at $189 is about $22,700 a year plus devices; the same cell on Vorne XL810 units is $44,900 once. And against the SmartSights rule of thumb above, either pays for itself on a fraction of a point of OEE at a $50 million plant, provided someone acts on the reason codes, which is the part no license includes.
Which standards define OEE?
SEMI publishes the two standards that define OEE measurement for equipment, and semiconductor fabs work to both. SEMI E10-0422, the “Specification for Definition and Measurement of Equipment Reliability, Availability, and Maintainability (RAM) and Utilization,” defines “six mutually exclusive, basic equipment states” and costs $286 for members and $380 for non-members. SEMI E79-0422 “defines metrics and calculations for measurement of equipment productivity, including overall equipment efficiency (OEE),” at $148 and $193. Both were revised in April 2022.
For discrete plants outside semiconductor, the working definition is the Lean Enterprise Institute’s, and the practical standard is whichever your customer audits against. What a standard buys a plant is a fixed list of equipment states and a fixed time base, which is exactly what the honesty section says gets moved. If two plants in the same company report OEE with different definitions of planned time, the comparison is fiction; a written definition, adopted once, is worth more than any benchmark.
How do plants raise OEE once they can measure it?
They raise it by attacking one loss on one machine with a named team and a date, then holding it. The documented cases share that shape. A Kaizen event on a packaging line, reported in Plant Engineering, took OEE from 52.6% to 87.6% averaged across two shifts within three months, against a company target of 70%. Coastal Machine and Supply raised utilization on a DMG MORI five-axis machine by 46% in the first two months of 2026 after connecting it to Datanomix, per Modern Machine Shop. Toftan 2, a sawmill, went from 40% to 60% in four months and 75% in a year after tying a bonus to OEE, in Evocon’s case study (vendor-published).
The measured loss structure says where to look first. If 90% of stop time is operators waiting for material, tools, setters or decisions, the biggest availability gain is usually in coordination rather than in the machine. Kathie Mahoney, President of MassMEP, told Manufacturing Dive in September 2026 that for small and mid-sized manufacturers the question is “how do we implement the technologies in the most efficient way for each facility.” The answer that holds is the same as in root cause analysis: pick the loss with the most minutes, find the condition behind it, and give the fix an owner. The software that holds that investigation is compared in root cause analysis tools.
Where does Morsa fit with OEE?
Morsa, the AI that operates the factory for you, does not measure OEE. It acts on the stops that OEE records, which is the 90% of stop time the study above traces to waiting rather than to the machine.
What it reads. The OEE tool or MES, the CMMS, the ERP and the schedule, and the places people report stops, whatever they are: WhatsApp, Microsoft Teams, email, SMS or whatever the plant runs on. A supervisor’s “Line 2 down, no inserts” is a signal the same way a downtime code is. Machines, PLCs and sensors can be read directly.
What it decides. Which work order, customer order and shift the stop affects, what it blocks downstream, and within approved rules what happens next: route the material request to stores, the tool request to tooling, the breakdown to maintenance; resequence; escalate; or ask for approval.
What it does. Opens the job with an owner and a due time, chases before the machine has waited an hour, escalates when it goes quiet, and closes on proof: the material staged, the tool issued, the work order confirmed. It also keeps the counts, such as the same minor stop on the same press four times this week, that a Pareto chart of losses needs.
What the user sees. The supervisor sees the ask and the deadline in the group they already read. The plant manager sees which machines are waiting on whom. How the loop works is in AI copilot for manufacturing.
What changed at plants running Morsa?
J4S, a 120-person glass plant onboarded in two days, took on-time completion of operational commitments from about 30% to about 75% in the first four weeks, across about 900 commitments, with no new software (how the glass plant did it with no new software). Those commitments include the ones that decide a machine’s availability: material staged before a changeover, a roller change window agreed with the shift, a quality clearance before a run. Plant Head Sunil K Verma: “I used to spend the first hour of every morning reconstructing yesterday. Now the chasing happens in the WhatsApp groups my supervisors already use, whether or not I remember.” At JRG Automotive, running live, Morsa identified 5 of 8 production-stopping material shortages early enough to act, and saved the plant $2 million.
Sources
Lean Enterprise Institute, Overall Equipment Effectiveness, Lean Lexicon, https://www.lean.org/lexicon-terms/overall-equipment-effectiveness/
Vorne Industries, oee.com: home, World Class OEE, TEEP, The Six Big Losses (vendor; no method stated for benchmarks), https://oee.com, https://oee.com/world-class-oee/, https://oee.com/teep/, https://oee.com/oee-six-big-losses/
Hedman, Subramaniyan and Almström, Analysis of Critical Factors for Automatic Measurement of OEE, Procedia CIRP 57, 2016, pp. 128 to 133, abstract verified via the OpenAlex record for the DOI, https://doi.org/10.1016/j.procir.2016.11.023
Evocon, James Devonshire, World-class OEE: industry benchmarks from more than 50 countries, 23 July 2024, updated 26 June 2026 (vendor data, 3,500+ machines), https://evocon.com/articles/world-class-oee-industry-benchmarks-from-more-than-50-countries/
Evocon, Spiros Vamvakas, How OEE gets manipulated, 29 May 2025, updated 13 October 2025, https://evocon.com/articles/how-oee-gets-manipulated/
Evocon, Toftan creates a bonus system around OEE (vendor case study; Peeter Põder quote), 13 August 2020, updated 15 July 2024, https://evocon.com/articles/case_study/toftan-creates-a-bonus-system-around-oee/
Evocon, OEE Excel template and Pricing, https://evocon.com/oee-excel-template/ and https://evocon.com/pricing/
Wikipedia, Overall equipment effectiveness (secondary; the “85% is World Class” myth section), https://en.wikipedia.org/wiki/Overall_equipment_effectiveness
NIST, Thomas and Weiss, Economics of Manufacturing Machinery Maintenance, AMS 100-34, June 2020, https://nvlpubs.nist.gov/nistpubs/ams/NIST.AMS.100-34.pdf
Siemens, The True Cost of Downtime 2024, 181 interviews, April 2019 to March 2023 (vendor research), https://assets.new.siemens.com/siemens/assets/api/uuid:1b43afb5-2d07-47f7-9eb7-893fe7d0bc59/TCOD-2024_original.pdf
Plant Engineering, How to boost OEE with asset management, 3 September 2025 (SmartSights rule of thumb), https://www.plantengineering.com/how-to-boost-oee-with-asset-management/
Plant Engineering, Mike Beauregard, Kaizen team boosts packaging operation OEE by 66.5%, 20 May 2022, https://www.plantengineering.com/kaizen-team-boosts-packaging-operation-oee-by-66-5/
Modern Machine Shop, Evan Doran, Machine Monitoring Boosts Five-Axis Utilization by 46%, 24 April 2026, https://www.mmsonline.com/articles/machine-monitoring-boosts-five-axis-utilization-by-46
IoT Analytics, Anand Taparia, MES vendors replace pen, paper, and spreadsheets, 15 December 2025, https://iot-analytics.com/mes-vendors-replace-pen-paper-spreadsheets/
Rockwell Automation, 93% of Manufacturers Have MES, But Only 23% Have Fully Integrated It, 14 July 2026, 1,560 respondents in 17 countries, https://www.rockwellautomation.com/en-us/company/news/press-releases/93-of-Manufacturers-Have-MES-But-Only-23-Have-Fully-Integrated-It-New-Rockwell-Automation-Report-Finds.html
Manufacturing Dive, Nathan Owens, Majority of manufacturers struggle to scale MES, 15 July 2026 (Lorenzo Veronesi quote), https://www.manufacturingdive.com/news/majority-manufacturers-struggle-scale-mes-rockwell-automation/825278/
Manufacturing Dive, 5 manufacturing professionals talk AI, technology, 9 September 2026 (Kathie Mahoney quote), https://www.manufacturingdive.com/news/5-manufacturing-professionals-talk-ai-technology/829855/
SEMI, SEMI E10-0422 and SEMI E79-0422, store pages with scope and prices, https://store-us.semi.org/products/e01000-semi-e10-specification-for-definition-and-measurement-of-equipment-reliability-availability-and-maintainability-ram-and-utilization and https://store-us.semi.org/products/e07900-semi-e79-specification-for-definition-and-measurement-of-equipment-productivity
Vendor product and pricing pages, all read 20 September 2026: Vorne https://www.vorne.com/products/models/; FourJaw https://www.fourjaw.com/pricing; Worximity https://www.worximity.com/pricing; Tulip https://tulip.co/pricing/; Inductive Automation https://inductiveautomation.com/pricing/ignition; MaintainX https://www.getmaintainx.com/pricing; MachineMetrics https://www.machinemetrics.com/pricing and https://www.machinemetrics.com/oee-software; Datanomix https://datanomix.io; Guidewheel https://www.guidewheel.com/pricing; Amper https://www.ecisolutions.com/products/amper/; Sight Machine https://sightmachine.com; Oden https://oden.io; Parsec https://www.parsec-corp.com/traksys/; QAD Redzone https://rzsoftware.com; Siemens https://www.siemens.com/en-us/products/opcenter/; Limble https://limble.com/pricing
Morsa’s own figures (J4S, JRG Automotive) are Morsa customer data.
Changelog
21 September 2026: first draft. Formula, worked calculation, a benchmark table with a sample column (the peer-reviewed 65% and Evocon’s 3,500-machine data next to Vorne’s unsourced 85%), the six big losses with the 884-machine loss findings, the inflation section, 18 tools with published prices, and SEMI E10 and E79.
