The core production KPIs are OEE (with availability, performance and quality), scrap rate, throughput, cycle time, plan attainment, downtime % and MTBF/MTTR. They all come from a few inputs: scheduled time, stops, ideal rate, units and good units. One example shift below shows all twelve.
The short list
Production KPIs are the numbers that tell you if the plant makes what it should, at the pace it should, with the quality it should. There are dozens. You need a few, each tied to a question.
Below are twelve that cover most plants, with the formula and the question each one answers. After the list, one shift of data shows how they all come from the same few inputs.
12 production KPIs with formulas
| KPI | Formula | Question it answers |
|---|---|---|
| OEE | (good units / ideal rate) / scheduled time | How well did we run when scheduled? |
| Availability | run time / scheduled time | How much time did we lose to stops? |
| Performance | units x ideal cycle time / run time | How fast did we run while running? |
| Quality | good units / total units | How much of it was good? |
| First pass yield | good first time / units entering | How much needed no rework? Details |
| Scrap rate | scrap units / total units | How much did we throw away? Details |
| Throughput | good units / time | How much good product per hour? Details |
| Actual cycle time | run time / units made | How long did one unit really take? Details |
| Plan attainment | good units / planned units | Did we make what we planned? |
| Downtime % | stop time / scheduled time | How much of the shift was stopped? |
| Changeover share | changeover time / scheduled time | How much does switching cost? Details |
| MTBF and MTTR | run time / failures, repair time / failures | How often do we fail and how fast do we recover? Details |
Outside the shift, three more help: manufacturing lead time, cost of poor quality and overall labor effectiveness.
Worked example: one shift, all the KPIs
Good KPIs come from a few raw numbers. Example: one shift gives these inputs.
Run time = 480 - 62 = 418 min. Ideal output in the run time = 418 x 50 = 20,900. Rework counts as a defect for OEE, so defects = 380 + 190 = 570, and good units = 18,810 - 570 = 18,240.
| KPI | Calculation | Value |
|---|---|---|
| Availability | 418 / 480 | 87.1% |
| Performance | 18,810 / 20,900 | 90.0% |
| Quality | 18,240 / 18,810 | 97.0% |
| OEE | (18,240 / 50) / 480 | 76.0% |
| Scrap rate | 380 / 18,810 | 2.0% |
| First pass yield | 18,240 / 18,810 | 97.0% |
| Throughput per scheduled hour | 18,240 / 8 h | 2,280 / h |
| Throughput per run hour | 18,240 / 6.97 h | 2,618 / h |
| Actual cycle time (ideal is 1.20 s) | 418 x 60 / 18,810 | 1.33 s |
| Plan attainment | 18,240 / 20,000 | 91.2% |
| Downtime % | 62 / 480 | 12.9% |
| Changeover share | 30 / 480 | 6.3% |
| MTBF (3 breakdowns) | 418 / 3 | 139.3 min |
| MTTR (27 min in 3 breakdowns) | 27 / 3 | 9.0 min |
Check: 0.871 x 0.900 x 0.970 = 0.760. The factors and the good-units method agree.
Quality and first pass yield match here at 97.0%, because neither counts the 190 reworked units as good. Those units still ship, so also track rework on its own, or it hides.
The losses add up to 100%
In the method of Overall Equipment Effectiveness (R. C. Hansen, 2001), OEE and all the loss percentages must add to 100% of scheduled time. If they do not, a loss is missing.
Here: stops 62 / 480 = 12.9%. Slow running = (418 - 18,810 / 50) / 480 = 41.8 / 480 = 8.7%. Rejects = (570 / 50) / 480 = 11.4 / 480 = 2.4%. Together with 76.0% OEE: 76.0 + 12.9 + 8.7 + 2.4 = 100%.
The biggest loss is stops, then slow running. That points the team to the stop log first. See the six big losses.
Which KPI for whom
Not everyone needs every number. Match the KPI to the person who can act on it.
| Who | What they need | How often |
|---|---|---|
| Operator | Units against plan this hour, current stop and its reason | Live, at the line |
| Supervisor | Shift OEE with its three factors, top three stops, scrap, plan attainment | Every shift |
| Plant manager | Weekly OEE trend by line, loss breakdown, scrap cost, MTBF and MTTR | Weekly |
| Improvement team | Pareto of stops by cause, changeover time, first pass yield by step | Per project |
Some KPIs warn early and some report late. Stops per hour and current OEE are early: you can still act. Scrap cost and plan attainment for the month are late: they confirm what already happened. Keep a mix of both on every board.
How to pick your KPIs
- Start from questions, not from a list. What do you need to decide this week?
- Keep it to five to seven. More than that, and nobody reads them.
- Pair a time view with a money view. Hansen warns that units per hour alone do not show business results, because margins differ by product.
- Use one definition per KPI. Write it on the dashboard.
- Show the trend, not only today. A KPI without a trend is a number, not a signal.
A good starter set: OEE with its three factors, scrap rate, plan attainment and downtime by cause. For a first dashboard, see free production tracking software. For setting up the data, see how to implement OEE.
Targets and trends
Give each KPI a target only after you have a month of clean data. A target set without a baseline is a guess, and people stop trusting it.
Set the target as a step from the baseline, for example "cut slow running from 8.7% to 6% of scheduled time". Review it every quarter. For the OEE target itself, see how to set OEE targets.
Common mistakes
- Too many KPIs. The key ones get lost.
- Only lagging KPIs. Scrap rate shows what happened. Stops per hour warn you early, while you can still act.
- Counting rejects as output. Use good units.
- Comparing lines with different ideal rates or definitions. Align first.
- Measuring everything and acting on nothing. Each KPI needs an owner and a review slot, such as the daily shift meeting.
A KPI set only works if the inputs are right: stops, speed and counts for every shift. Fabrico is a light MES for OEE that reads them from PLCs or IoT sensors, with add-on AI cameras counting output where a machine has no usable signal, and shows OEE per line and shift. See how Fabrico shows OEE per line and shift.
What people ask next
What are the most important production KPIs?
For most plants: OEE with availability, performance and quality, scrap rate, plan attainment and downtime by cause. Add throughput and cycle time if you balance lines. Pick what fits your questions.
What is the difference between a KPI and a metric?
A metric is any number you measure. A KPI is a metric you have chosen because it tracks a key goal and someone owns it. All KPIs are metrics, but most metrics are not KPIs.
How often should production KPIs be reviewed?
Shift-level KPIs should be seen at the shift change or the daily meeting. Weekly trends go to the weekly review. When you roll up, weight each shift by its scheduled time instead of taking a plain average.
Is OEE the best production KPI?
OEE is the best single number for how well equipment runs when scheduled. It does not show cost, delivery or safety. Use it with others. See what is OEE.
Which KPIs belong on a shop floor dashboard?
Live OEE for the shift, the current stop reason, units made against plan, and scrap. Keep it to what operators can act on. See OEE dashboard examples.
Fabrico is a light MES for OEE. It takes stops, speed and counts from PLCs or IoT sensors, with add-on AI cameras counting output where a machine has no usable signal, and operators tag the cause at the line.
Related guides: Lean Metrics That Matter on a Line, With Formulas · Plant Manager OEE KPIs: What to See Each Week · Production Efficiency vs OEE: How They Differ · MES Dashboard: What to Show and Who Acts on It · Manufacturing KPI Formulas: A Full Cheat Sheet