When OEE drops after a new machine, the cause is usually one of four: the bottleneck moved and the yardstick rose with it, spare hours were added to the schedule, the new machine is still ramping up, or it is short of feed or operators. First check good output on the same product mix. If output rose while OEE fell, the investment is doing its job and the score only needs explaining.
Check output before you worry
A lower OEE after an investment feels like bad news. Often it is only a change in what the percentage is measured against.
So look at one thing first: good units per week, on the same product mix, before and after. Then work through the four causes below.
Cause 1: the bottleneck moved
A new machine usually goes where the constraint was. Once it arrives, some other step becomes the slowest, and line OEE is measured against that step's higher rate.
Example: a line had a filler at 100 per minute as its bottleneck and a packer at 110, and it made 33,600 good units in a 480-minute shift. A new filler at 130 per minute now makes the packer the bottleneck, and the line makes 35,200.
Output rose 4.8% and OEE fell 3.3 points. Both are true, and line OEE vs machine OEE explains the method behind them.
The next step here is plain. The packer's losses now decide the line's result, so move your stop log and your attention to it, with help from bottleneck analysis.
Check the old machines' ideal rates too. A machine that no longer sets the pace should be rated at the speed needed to feed the bottleneck, as in ideal cycle time.
Cause 2: more hours scheduled than orders need
This one catches parallel steps. The new machine comes in, someone adds it to the schedule, and demand has not moved.
Example: two identical machines ran 480 minutes each and made 720 good minutes of output, so OEE was 720 ÷ 960 = 75%. A third joins the schedule, with the same orders to share.
| Set-up | Scheduled min | Good min | OEE |
|---|---|---|---|
| Before: 2 machines | 960 | 720 | 75.0% |
| After: 3 machines, all scheduled | 1,440 | 720 | 50.0% |
| After: 3 machines, 960 min scheduled | 960 | 720 | 75.0% |
The fix is in the plan. Schedule only the machine hours that the orders need, and treat the rest as excluded time, as in OEE with no orders.
The spare capacity then shows up where it should, in TEEP and loading. OEE for parallel machines covers the rest of the rules for a shared step.
Cause 3: the new machine is still ramping up
A new machine rarely runs at its best in the first weeks. Pooled with mature machines, it drags the roll-up down even if nothing else has changed.
Example: the same two old machines hold 75%, and the third is scheduled for a real 480 minutes at 40% during ramp-up.
So report the area both with and without the new machine until ramp-up ends. OEE during commissioning and ramp-up covers when to start counting and when to stop treating it as new.
Cause 4: the new machine shares something scarce
A machine needs people, material, tools and support. If it takes these from the others, OEE can fall across the whole area.
- Operators split. The same crew now covers more machines, so stops wait longer. Look for a rise in waiting-for-operator time.
- Feed runs short. The new capacity is faster than the step before it, so machines sit starved. See blocked and starved.
- Output backs up. The step after cannot take the extra product, so the new machine sits blocked.
- Support stretched. Technicians and setters are busy on the new machine, so the old ones wait.
Here a drop on the old machines is the warning sign. Compare their loss codes before and after, and see which ones grew.
What to measure, in order
Same mix and same hours of orders, before and after. This is the test that matters.
Find the queue, then confirm the line rate you use in OEE.
Are machines scheduled with nothing to run?
Their trend shows whether anything really got worse.
A rising curve is fine. A flat one needs a look.
Starved, blocked and waiting for an operator show flow problems.
Mark the date of the change on your OEE chart. Without that note, in six months nobody will remember why the line stepped down.
How long to wait before judging
Do not judge in the first week, when almost any number is possible. Our advice: compare several weeks before with several weeks after, on a similar mix.
Set the check date before the machine arrives. That keeps the review from being held only when someone is already worried.
How to read what you find
| Good output | OEE | What it likely means |
|---|---|---|
| Up | Down | Yardstick moved; work on the new bottleneck |
| Flat | Down | Spare hours scheduled, or the old limit was elsewhere |
| Down | Down | Real problem: ramp-up, flow or shared resources |
| Up | Up | Good; lock in the new rate and move on |
If output stayed flat, ask whether the new machine was at the real constraint. How to calculate throughput and the production capacity calculator help you test this.
Before the next purchase, try how to increase production capacity. Sometimes the hours you already own are the cheaper answer.
After a change on the line you need before-and-after numbers on the same terms. Fabrico takes stops, speed and counts live from PLCs or IoT sensors, including older machines, and shows OEE per line and shift. See how Fabrico shows OEE per line and shift.
What people ask next
Is it normal for OEE to drop after adding a machine?
It is common, and it does not always mean a problem. Check good output first, then work through the four causes above.
Why did my line OEE fall when output rose?
Most likely the bottleneck moved to a faster step, so the line is measured against a higher rate. Line OEE vs machine OEE shows the sum.
Should I recalculate the ideal rate after adding a machine?
Yes, for the line and for each machine that no longer sets the pace. Date the change and mark it on the chart, as covered in ideal cycle time.
Should the new machine be in the OEE report during ramp-up?
Yes, but show the area with and without it. OEE on a new line covers when to start and stop treating it as new.
How do I prove the new machine paid off?
Compare good output on the same mix, then price the change. The hidden factory calculator and the downtime cost calculator help you put a value on the hours.
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: OEE and Seasonal Demand: Reading the Swings