Overview: A TPM consultant improves OEE by first fixing how it is measured, then removing the losses behind it. That means real-time data capture, honest classification of availability, performance and quality losses, and disciplined loss analysis. Real gains come from eliminating the six big losses, not from softening standards or hiding minor stops.
A TPM consultant is usually called in to raise Overall Equipment Effectiveness (OEE), and here lies a quiet risk. The quickest way to move that number is not to fix the machine but to edit how the machine is measured. When a plant chases the score instead of the losses behind it, OEE stops describing equipment and starts flattering a report. Good consultants work the opposite way. They make the number harder to fake, then help the team earn each point honestly.
What is OEE, and why does it get gamed?
OEE measures how much of the planned production time a machine spends making good parts at rated speed. The formula is simple: Availability times Performance times Quality. Each factor sits between 0 and 100 percent, and the three multiply, so losses compound. Three factors at 90 percent each do not give 90; they give roughly 73.
OEE gets gamed because every factor can be moved on paper without touching the shop floor:
- Availability is the easiest. Changeovers, cleaning, minor stoppages and waiting time get quietly reclassified as planned and pushed out of the denominator. Real availability then looks far healthier than it is.
- Performance bends when the ideal cycle time is edited down to the rate the line actually achieves. Set the standard to the achieved speed and performance always reads near 95 percent.
- Quality improves when rework is counted as good output, or when startup scrap is left out.
None of this changes a single part shipped. It only changes the number on the board. A metric maintained this way becomes a vanity score, and TPM built on a vanity score improves nothing.
What does a responsible TPM consultant do differently?
A responsible consultant treats OEE as a diagnostic tool, not a target to be hit. Before asking a plant to raise OEE, they make sure the measurement is honest and consistent. That order matters. You cannot improve what you are measuring wrongly.
The uncomfortable early step is that OEE almost always falls when it is measured properly for the first time. Once minor stops, short interruptions and slow cycles become visible, the number drops. A consultant frames this openly. The lower figure is not bad news; it is the true baseline, and the only honest place to start improving from.
How does accurate data collection change the picture?
Accurate data collection removes guesswork and human bias from the number. Manual logs and end-of-shift estimates miss the small losses that matter most, especially minor stops and reduced speed, which operators rarely record. A good consultant will usually:
- Capture stop reasons in real time, from the machine where possible, rather than reconstructing them later from memory.
- Fix the ideal cycle time to a genuine engineering standard, not to yesterday’s average.
- Define, in writing, what counts as planned versus unplanned time, so the same event is always classified the same way.
- Keep quality data tied to actual inspection results, with rework recorded as a loss.
When the rules are written down and the data is captured at the source, two shifts measuring the same machine produce the same OEE. That consistency is what makes the metric trustworthy.
Loss analysis: turning OEE into a to-do list
OEE on its own only tells you that a machine is underperforming. Loss analysis tells you why, and that is where a TPM consultant earns their keep. The standard frame is the six big losses, grouped under the three OEE factors.
| OEE factor | The six big losses |
|---|---|
| Availability | Breakdowns and equipment failure; setup and changeover |
| Performance | Idling and minor stops; reduced running speed |
| Quality | Process defects and rework; startup and yield losses |
The method is to measure each loss, rank them by the time or cost they consume, and attack the largest few first. A consultant maps the top three or four losses to the relevant TPM pillars, such as autonomous maintenance for minor stops and cleaning-related defects, or planned maintenance for breakdowns. This keeps improvement focused on real problems rather than on dressing up the score.
How should results be measured responsibly?
OEE should never travel alone. Reported beside supporting measures, it is far harder to game and far more useful. Sensible companions include mean time between failures (MTBF) and mean time to repair (MTTR) for reliability, changeover time for setup losses, first-pass yield or defect rate for quality, and unplanned downtime in hours. If OEE rises while downtime and defects do not fall, the gain is on paper only, and the supporting metrics expose it.
Trend also beats snapshot. A consultant watches the direction of OEE over weeks against a fixed definition, rather than celebrating a single strong shift. Comparing OEE across different machines or plants is usually a mistake, because the definitions and standards behind each number rarely match.
Common pitfalls to avoid
- Setting an OEE target before the measurement method is agreed and stable.
- Excluding changeovers, cleaning or waiting time to protect the number.
- Using nameplate or best-ever speed as the ideal cycle time.
- Treating a high OEE as the goal, rather than as a signal about where losses sit.
- Comparing OEE figures that were calculated under different rules.
Conclusion
The real job of a TPM consultant is not to make OEE look good but to make it tell the truth, then use that truth to remove losses one by one. Honest data, clear definitions and disciplined loss analysis turn OEE from a number people manage into a tool that drives genuine equipment improvement. A slightly lower but honest OEE is worth far more than a high one nobody trusts.
Frequently asked questions
What does a TPM consultant improve first?
Usually the measurement itself. Before raising OEE, a good consultant makes sure data is captured in real time and every loss is classified consistently, so the baseline is honest.
Why did our OEE drop after a consultant arrived?
Because it is finally being measured properly. Minor stops and slow cycles that were invisible before now show up. The lower figure is your true starting point, not a step backwards.
Can OEE be manipulated?
Yes, easily. Reclassifying stoppages as planned time, editing the ideal cycle time down, or counting rework as good output all inflate OEE without improving output.
What is a good OEE score?
It depends on the process, so chasing a single world-class number can mislead. A rising trend under a fixed definition beats a high one-off reading.
How does loss analysis link to OEE?
It breaks OEE down into the six big losses, so teams can see exactly where time and quality are lost and fix the largest causes first.
Which metrics should sit alongside OEE?
MTBF, MTTR, changeover time, first-pass yield, defect rate and unplanned downtime. Together they make OEE harder to game and easier to trust.
