An OEE improvement action plan has five steps: pick the bottleneck line, measure OEE and the losses in minutes, choose no more than three projects, give each an owner, a date and a planned gain, then review weekly and re-measure monthly. A first cycle takes about 90 days.
What the plan is for
An action plan turns an OEE number into a short list of projects, each with an owner, a date and a planned gain in OEE points.
It works on one line at a time, so the team finishes what it starts before it takes on more.
This page is the plan. For the methods behind each fix (changeover, micro-stops, breakdowns), see how to improve OEE.
Three phases, 90 days
- 1Days 1 to 30: measure. Choose the line. Set the ideal rate. Collect scheduled time, good units and every stop with a cause. Calculate the baseline OEE and the loss split.
- 2Days 31 to 60: fix the top three. Run one project for each of the largest losses. A short weekly review (about 20 minutes): owner, progress, blockers.
- 3Days 61 to 90: hold and check. Standardize what worked. Re-measure OEE the same way. Compare to the baseline, mark the change, and choose the next three.
The 90-day cycle. After day 90 you restart at step 1 on the next loss or the next line.
Step 1: choose where to work
Start with the bottleneck: the step that limits the output of the whole plant. A gain there is a gain for the plant. A gain on a machine with spare capacity is only a nicer number. See bottleneck analysis.
Hansen asks one question: if the effectiveness of this step improves, will the bottom line be significantly affected? If not, pick another step. He also advises against pilot areas that are not key to the plant, because they get few resources and the lessons do not transfer (Overall Equipment Effectiveness (R. C. Hansen, 2001)).
Step 2: measure the baseline
Use the simplest method first: good units ÷ ideal rate ÷ scheduled time. Then split the missing percentage into losses, so that OEE plus all losses add to 100%. If you need the calculation, see how to calculate OEE, and the OEE loss tree for the structure.
Worked example. A line has a baseline OEE of 62%. The 38% that is missing splits like this:
- OEE62%
- Changeovers11 pts
- Minor stops9 pts
- Breakdowns6 pts
- Slow running5 pts
- Scrap and rework4 pts
- Start-up losses3 pts
Step 3: choose no more than three projects
Hansen works on no more than three key projects at a time, and finishes them fast. More than that and each one starves. Choose by size and by how fixable the loss is. Use a downtime Pareto to rank, and go deeper into the top items by cause.
| Project | Loss today | Planned cut | Planned gain | Owner | Due |
|---|---|---|---|---|---|
| 1. Changeover: split internal and external steps | 11 pts | 25% of 11 | +2.75 pts | Line lead | Day 60 |
| 2. Minor stops: fix the top 3 causes (60% of the loss) | 9 pts | Halve 5.4 pts | +2.7 pts | Maintenance tech | Day 60 |
| 3. Breakdowns: remove the top repeat failure (3 pts) | 6 pts | Cut 2/3 of 3 | +2.0 pts | Reliability eng. | Day 75 |
| Total planned | +7.45 pts |
Planned gains are targets set at the start. They are not a promise. Method for project 1: SMED.
Output gain at the same scheduled time = 69.45 ÷ 62 = 1.120, so +12.0%
Good output moves in step with OEE when the scheduled time is the same, so a gain of 7.45 points on a base of 62 is about 12% more good output. The extra output is only worth having if the plant can sell it. Treat the gain as recovered capacity to meet demand from the equipment you already own.
Step 4: write the plan down
One row per project. Keep it on one page, visible at the line.
| Field | What to write |
|---|---|
| Loss | Which loss, in minutes and OEE points |
| Cause | The root cause, or "to be found" with a date |
| Countermeasure | What will change |
| Owner | One name |
| Due date | A real date |
| Expected gain | In OEE points, as in the table above |
| Check | How and when you will measure the result |
| Result | Filled in after the check |
The fields of the plan. The last two rows are the ones most often skipped.
Step 5: review and re-measure
- Weekly, about 20 minutes. Each owner says what moved, what is stuck and what they need.
- Monthly. Re-calculate OEE with the same method and ideal rate. If you change the ideal rate, mark it on the chart as an intervention point, because OEE often dips after a change.
- After each project. Compare the planned gain to the actual gain. If it missed, ask why before you start the next one.
Hansen describes a program with about three months of education, three months of first applications and about 12 months to reach a 10% gain, and notes that the gain keeps compounding after that (Overall Equipment Effectiveness (R. C. Hansen, 2001)). Do not expect a big jump in 90 days. Expect a clear first step and a method you can repeat.
Common mistakes
- Too many projects. Five projects mean five half-finished ones.
- No baseline. You cannot show an improvement you did not measure.
- Changing the definition mid-way. A new ideal rate or new scheduled time makes the before and after incomparable.
- No owner. "The team" is not an owner.
- Skipping the check. If you never compare plan and result, the plan stops being credible.
- Starting on a line that is not the bottleneck. The gain does not reach the plant.
What to do next
Pick the line and set a date for day 1. Use the OEE calculator for the baseline and the six big losses calculator for the split. Agree targets with how to set OEE targets, and bring the daily numbers to the daily production meeting. Ideas from the floor go into kaizen.
A plan needs a trusted baseline and a way to check each project. Fabrico is a light MES for OEE: stops, speed and counts come live from the machine, operators tag the cause at the line, and OEE is shown per line and shift, so you can compare before and after. See how Fabrico's light MES tracks OEE.
What people ask next
How many improvement projects should we run at once?
No more than three for one line or area. Hansen advises two or three, finished quickly, with spare resources supporting them. Start the next only when one is closed and checked.
Which loss should we fix first?
The biggest one you can fix. Rank the losses in minutes on the bottleneck, then check which of the top ones has a known cause or a method, such as SMED for changeovers. Start where size and fixability meet.
How long until OEE improves?
A first gain in 90 days is realistic on a focused loss. Hansen's program plan reaches a 10% gain in about 12 months, and it keeps going after that. Treat any single number as an example, not a promise.
Who should own the plan?
One person for the plan, usually the plant or production manager, and one owner for each project. Projects owned by the people who run the line stick better than projects handed down.
How do we prove the gain is real?
Re-measure with the same ideal rate, the same scheduled time and the same method, and compare the loss in minutes before and after. If the number moved but output did not, check the ideal rate and the count of good units. See OEE calculation mistakes.
Related guides: How to Set a Downtime Target for a Production Line · MTBF Improvement: How to Get Fewer Failures