Blog / OEE
OEE · Guide

OEE by Product: What a SKU Split Reveals

OL By OEE Lab Editorial |Updated 2 November 2026 |6 min read
Short answer

For each product, OEE = good units × ideal cycle time ÷ the scheduled time spent on that product, including its changeover. We charge each changeover to the incoming product, though practice varies. Rank products by lost minutes as well as by OEE, because a product with a good percentage can still be your largest loss.

What the line average hides

A line at 70% tells you little about what to do. It may be 70% on everything, or it may be one product at 80% carrying three at 40%.

Those two lines need different plans. A split by product shows you which one you have.

How to split it

You need one thing that many logs do not have: every count and stop tied to the product running at the time. An order or run number on each record is enough.

1
Tag each run
Every stop, count and scrap record carries the order or product number.
2
Set the time boundaries
A product's time starts when the changeover to it starts, and ends when the next changeover starts.
3
Convert to minutes
Good minutes = good units x that product's ideal cycle time.
4
Check the sum
The products' scheduled minutes must add up to the line's. Their weighted OEE must match the line's OEE.
The split in four moves.

OEE method says a changeover is an availability loss for the line, as how to count changeovers in OEE explains. It does not say which product carries it, so this rule is a choice and practice varies.

We charge the incoming product, because the changeover happens so that product can run. Other plants charge the outgoing product, or show changeovers as their own row beside the products.

Each option adds back up to the same line total. Pick one and keep it, because switching rules moves product scores with no change on the floor.

Keep your ideal rates honest too, using the rule in Overall Equipment Effectiveness (R. C. Hansen, 2001):

  • One accredited rate per family of the same size and format.
  • A product that is hard to run keeps that rate and shows as low OEE, which is the point of the split.

A week on one line, by SKU

Example: one line, 2,400 scheduled minutes, four SKUs.

SKUScheduled minGood minOEELost min
A1,20096080.0%240
B72050470.0%216
C30015050.0%150
D1806335.0%117
Line2,4001,67769.9%723
Line OEE = 1,677 ÷ 2,400. The simple average of the four, 58.8%, is wrong because it ignores scheduled time.

The combining rule is the one in how to average OEE across machines. If your line total does not match, a record is missing a product tag.

Reveal 1: the worst score is not the worst loss

D has the lowest OEE, so it gets the blame. Now rank by lost minutes.

A, OEE 80%240 min
B, OEE 70%216 min
C, OEE 50%150 min
D, OEE 35%117 min
Bar length is each SKU's share of the 723 lost minutes. The best product is the largest loss.

A, the best product, loses twice as many minutes as D. A small gain on a large product can beat a heroic fix on a small one.

Each loss is worth a look, but start from the minutes. A downtime Pareto by product makes this ranking easy to read each week.

Reveal 2: each product loses in its own way

Split each product's lost minutes into four buckets, as you would in an OEE loss tree.

SKUChangeoverStopsSpeedScrapTotal
A301107030240
B40966020216
C70401525150
D60151230117
All200261157105723
Lost minutes per SKU. The largest loss in each row is in bold.

Now the actions differ by product.

  • A and B: stops and speed lead. These are machine and line issues, so start with unplanned downtime and speed losses.
  • C: changeover takes nearly half its losses. This is a SMED or sequencing question.
  • D: changeover and scrap are 90 of its 117 lost minutes. It runs so short that the first bad units count heavily. The cause lies in order size and start-up, and little of it is in running speed.

Reveal 3: a low score can still pay

OEE says how well line time was used. It does not say what that time earned, and this is where product OEE becomes a business tool.

C, OEE 50%15 per min
A, OEE 80%12 per min
B, OEE 70%10 per min
D, OEE 35%5 per min
Hypothetical contribution per scheduled minute. Bars are scaled to 15 euros per minute.

C has a poor score and earns the most per minute of line time. Dropping it because of its OEE would be a costly mistake, so shorten its changeovers.

D is the one to question. It has a low score and earns little, so it needs a minimum order, a higher price or a place at a natural break in the sequence; the changeover cost calculator puts a figure on it.

Make it a monthly habit

A product split does not need a daily meeting. Once a month, take the top three products by lost minutes and the top three by lowest OEE.

Bring production and planning into the room together. Stops belong to the line, while order size and sequence belong to planning, and the split shows which is which.

Each loss then gets one owner and one action. Next month, look at the same table and see whether those minutes moved.

How much data you need

A SKU that ran once tells you about that run. Our advice: judge a product only after several runs, and show the number of runs next to its OEE.

With hundreds of SKUs, group by family first and drill down only where a family stands out. Short, rare runs will always look noisy, so read them in groups.

Where a product split goes wrong

  • A softer ideal rate for hard products. That hides the very cost you are looking for. See ideal cycle time.
  • Using OEE to rank the crew by product. Products are not assigned at random, so this compares schedules.
  • Leaving out changeovers. Short runs then look better than they are.
  • Mixing the sequence effect in. A changeover from C to A may differ from one from A to C, so track the pairs that matter.
  • Acting on OEE alone. Check the money too, starting with scrap in the cost of poor quality.

Need the line-level sum first? OEE with mixed products shows how to add different products into one honest figure.

See it live in Fabrico

A product split is only as good as the stop and count records behind it. Fabrico takes stops, speed and counts live 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. See how Fabrico tracks stops.

What people ask next

How do I calculate OEE per product on a shared line?

Add up the scheduled time for each product, including its changeover, and convert its good units to minutes with its ideal cycle time. Then divide. To sum the products back up, use scheduled-time weighting.

Who gets the changeover time, the old product or the new one?

Practice varies, and no OEE rule settles it. We suggest the incoming product, and what matters is one rule used every time. How to count changeovers in OEE covers the line-level rule.

Why is my product OEE average different from line OEE?

A simple average ignores how long each product ran, so it will differ. The scheduled-time weighted figure should match the line exactly, and if it does not, a record is missing a tag.

Should each SKU have its own ideal cycle time?

Each family of the same size and format gets one rate. SKUs that differ only in a label or recipe usually share one, as ideal cycle time explains.

Should I drop a product with low OEE?

Not on OEE alone. Check what it earns per minute of line time, then try shorter changeovers, a minimum order or a better place in the sequence. The six big losses calculator helps you size each option.

See OEE per line and shift on your own machines

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.

Book a Fabrico demo

← More OEE Lab guides · OEE with mixed products →

OEE essentials: OEE calculator · What is OEE · Six big losses · Availability · TEEP · Cost of downtime · OEE benchmark · Free OEE tracker · Free OEE software · How to track downtime · How to implement OEE · Light MES vs full MES