Quantifying performance via OEE 06 February 2024

Overall equipment effectiveness

Overall equipment effectiveness (OEE) can be a powerful way of looking at – and quantifying – the performance of any process or industry. Toby Clark explains how it works

OEE (overall equipment effectiveness) has been described as a measure of how well a system (a machine, a set of equipment or a production line, for instance) is utilised relative to its full potential. Fundamentally it is a simple statistic – a percentage score. A 100% OEE score cannot be improved, but it is also unlikely to be achieved.

OEE arose from the philosophy of total productive maintenance (TPM) developed by Japanese firms from the 1950s onward, and is not just a key performance indicator (KPI): it can be used as a ‘heuristic’, a problem-solving tool to pinpoint where issues have arisen. Improvements in OEE are likely to result in direct improvements in profits and return on capital employed (ROCE). Also, a system with a high OEE tends to work predictably, which can reduce costs in inbound and outbound logistics and other areas.

However, it is easy to get lost in acronyms, jargon and vague definitions. Here we sift through the varying descriptions of OEE, to get usable metrics and how they can be applied.

OEE is very simple to calculate, as it is just the product of three factors: availability, performance and quality (A, P, Q). Each factor is itself scored as a percentage, or an index with a maximum value of 1. To get a 100% OEE, a production line must be only producing good parts (100% Quality) as quickly as possible (100% Performance) and with no significant downtime (100% Availability).

If you have more realistic results like an availability of 90%, performance of 85% and quality of 95%, then the OEE is the product of the percentages converted to decimals, and then back to a percentage: 73%.

It’s notable that three apparently excellent factors (none below 85%) result in what seems to be a good – but obviously improvable – OEE score. This is one of the things that makes OEE such a useful metric. While there are many standards for OEE, a ‘world class’ result for a manufacturing plant might be 85%; if you are getting much more, you might be measuring it incorrectly.


Availability indicates the proportion of time that a system is available to be used. For this you need to know the total operating time; any time lost to scheduled stops such as setup, adjustment and maintenance – in other words, planned downtime; time lost to breakdowns (or power outages, or material shortages – any sort of unplanned loss of function) or any other reason.


Availability = (Total operating time - total time lost)/Total operating time

Say you have an injection-moulding machine which operates for 16 hours (960 minutes) a day, but it requires 15 minutes of setup time each day, and 20 minutes of clean-up. Then say that you have monitored it for a month, and that it experiences an average breakdown time of eight minutes a day:

Availability A = (960 - 15 - 20 - 8)/960 = 917/960 = 0.955 = 96%


This is the measure of operating pace (speed, production rate, etc) compared with the theoretical maximum. If our injection-moulding machine can theoretically produce 1,400 items an hour, but worn bearings and hydraulics mean that it cannot average more than 1,250 items an hour, then:

Performance P = 1,250 / 1,400 = 0.893 = 89%


Q relates to the proportion of product (or products, or other output) that is satisfactory. In the case of our injection moulding machine, sticking valve and ejector pins might result in rejected mouldings when the machine starts up, while poor temperature control or a worn mould might lead to rejected products throughout the day.

Quality Q = (Total Production - Startup rejects - Production rejects) /Total Production

Say our machine produces 19,100 mouldings a day, but the first 100 are rejects and there are typically 300 other rejects each day:

Quality Q = (19,100-100-300)/19,100 = 18,700/19,100 = 0.979 = 98%

So the Overall Equipment Efficiency of our injection moulding machine is:

OEE = A x P x Q = 96% x 89% x 98% = 84%

Strictly speaking, OEE is only measured in relation to the system’s hours of operation. An alternative is to measure performance against ‘calendar hours’ – to assume that ideally the system would operate 24 hours a day, 365 days a year. This measurement is known as total effective equipment performance (TEEP). Effectively, TEEP is equal to OEE multiplied by another factor: the proportion of operational hours relative to calendar hours. This factor is known as loading. So if your production line runs sixteen hours a day, five days a week, its loading is (16/24) x (5/7) = 48%.

A simple snapshot of how a system is performing will not give the best inputs for OEE: rather, data should be be measured over a long enough period that trends become clear. Having said that, it is important to recognise when periodic inefficiencies – such as annual holidays or more-expensive night shifts – change the balance of inputs.

Obviously, there are different ways to define losses, which can affect the management of the system. For example: if you count setup time as part of planned downtime, it may seem inevitable, or less likely to be tackled than if it is seen as unplanned. But whether it is ‘planned’ or not, set-up time still reduces availability and lowers OEE.

Similarly, performance is judged relative to maximum operating speed: is this maximum defined by the theoretical speed of the machine quoted by the manufacturer? Or the maximum speed ever attained in that system? These may be very different numbers, and will directly affect the OEE score – so if you benchmark OEE between systems, it is important to know how it is defined.

Quality can also be defined in very different ways: say our injection moulding machine produces some parts which are not complete rejects, but can be reworked in another part of the plant. Best practice is to measure ‘first pass yield’, where items which need rework are counted as rejects. Otherwise, they might not affect the Q score for that machine, and we might say that the more parts go through the rework station, the better its performance – a positively unhelpful KPI.

So defining the boundaries of OEE measurements is important – and setting targets that relate to what management wants to achieve, rather than just giving an inflated OEE score.


The basic calculations for OEE are straightforward, and a simple Excel spreadsheet may tell you all you need to know. However, numerous companies provide software solutions to help with evaluating OEE, most of them using automated monitoring or even IoT (Internet of Things) interfaces. Some provide mobile apps for real-time monitoring, or even physical ‘scoreboard’ displays for shopfloor use.

Among the firms with a UK presence are:

  • PlantRun (plantrun.co.uk)
  • FourJaw Manufacturing Analytics (fourjaw.com)
  • Seiki Systems (seikisystems.co.uk)
  • OEEsystems (oeesystems.com)
  • ABB (new.abb.com/industrial-software)


    OEE often uses the concept of ‘The Six Big Losses’ to classify production losses.

    Availability-related: Breakdowns: machine malfunctions and unplanned downtime

    Setup, changeover time & adjustments (or planned downtime)

    Performance issues: Minor stops & idle time, speed loss

    Quality issues: rejects on startup, production rejects (and rework)

    Toby Clark

    Related Companies
    Seiki Systems

    This material is protected by MA Business copyright
    See Terms and Conditions.
    One-off usage is permitted but bulk copying is not.
    For multiple copies contact the sales team.