EPEI (every part every interval) is the frequency with which a process cycles through its full mix of products. If a line changes over in a sequence that makes every part number it runs once every two days, its EPEI is two days. A smaller EPEI means smaller batches, less inventory, and faster response to demand.

EPEI is the number that connects changeover time to batch size. It answers a question every scheduler faces: given the changeovers this line can afford, how often can I run each product? The answer sets your lot sizes, your inventory, and how quickly you can react when an order changes. Cutting changeover time with SMED is the main lever that shrinks EPEI, which is why the two topics are inseparable. This post covers what EPEI means, how to calculate it from available changeover time, how it sets batch size, and what a good target looks like, inside the broader practice of lean manufacturing.

What is EPEI (every part every interval)?

EPEI is the interval it takes a process to produce every part number assigned to it at least once. It is expressed in time: an EPEI of three days means the line rotates through its entire product mix every three days, so each product is made once every three days. Sometimes it is called EPEx, "every part every x," where x is the interval. The term comes from the Lean Enterprise Institute's lexicon and is a standard measure of scheduling flexibility.

The intuition is a rotation. Picture a line that runs products A, B, C, and D in sequence, then starts over at A. How long one full lap takes is the EPEI. A short lap means you make a little of everything often; a long lap means you make a lot of one thing, then do not touch it again for a while. Short laps are better for flow, inventory, and responsiveness, and the thing that lets you run short laps is cheap changeovers.

EPEI as a production rotationOne full lap through the mix = the intervalABCDEPEI = 1 lapShorter lap = smaller batches, less inventory, faster response.
EPEI is how long one full rotation through the product mix takes. The shorter the lap, the smaller the batches and the more responsive the line.

How do you calculate EPEI?

Calculate EPEI by comparing the changeover time needed to cycle the whole mix once against the changeover time available per day. Work through it as six steps. Take a line that runs six products, each with a 30-minute changeover, on a schedule that leaves 90 minutes a day free for changeovers after meeting demand.

  1. List the product numbers and count them. Six products, so one full rotation requires six changeovers.
  2. Measure the changeover time for each. Time from last good piece to first good piece, the same way SMED measures it. Here every changeover is 30 minutes.
  3. Find the available changeover time per day. Take total available production time and subtract the run time needed to meet daily demand. What is left is time you can spend on changeovers. Here, 90 minutes per day.
  4. Sum the changeover time to cycle every product once. Six products × 30 minutes = 180 minutes of changeover per full rotation.
  5. Divide to get EPEI. 180 minutes needed ÷ 90 minutes available per day = 2 days. Every product every 2 days.
  6. Convert EPEI to batch size. Batch size = interval × average daily demand. A product selling 100 units a day, run every 2 days, needs a batch of about 200 units.
The EPEI calculation, plugged inChangeover needed vs. changeover you can afford6 products × 30 min = 180 min90 min available per day=EPEI = 2 daysEvery part gets made once every two days; batch size = 2 days × daily demand.
EPEI is the changeover time needed for one full rotation divided by the changeover time available per day.

What is available changeover time?

Available changeover time is whatever production time is left over after the line has met its demand. If a line has 20 hours of usable time a day and needs 17 hours of that just to run the quantities customers ordered, the remaining 3 hours are what you have for changeovers. That leftover is the real constraint on EPEI, because every changeover you add is time not spent producing. A common planning rule of thumb is to reserve roughly 10 percent of available time for changeovers, then size the interval to fit; if changeovers eat more than that, either the changeover is too slow or the mix is too wide for the capacity.

This is why EPEI is not a free choice. You cannot simply decide to run everything every day if your changeovers do not fit in the time left over. The interval falls out of the arithmetic between demand, capacity, and changeover time, which is also why the fastest way to a shorter interval is usually a faster changeover, not a heroic schedule.

How does EPEI set batch size and inventory?

EPEI sets batch size directly: the batch of each product is the interval multiplied by that product's average demand over the interval. Run a product every two days and you must make two days of it each time, which means you carry, on average, about a day of that product as inventory between runs. Halve the interval to one day and the batch and the average inventory roughly halve with it. This is the mechanism behind a core lean claim, that small batches cut inventory, stated as a number you can compute.

The connection runs the other way too. A short EPEI is what makes heijunka (production leveling) and pull systems like kanban possible, because you cannot level a schedule around products you can only afford to run occasionally. It also shapes how you buffer between processes: when two steps cannot run at the same rhythm, a sized FIFO lane or supermarket holds the difference, and a tighter EPEI keeps those buffers small.

How does SMED shrink your EPEI?

SMED shrinks EPEI because changeover time is the denominator-eating term in the whole calculation. Cut each changeover on the example line from 30 minutes to 10, and the changeover time to cycle all six products drops from 180 minutes to 60. Against the same 90 minutes of available changeover time per day, the interval falls from 2 days to about two-thirds of a day. You could now run every product more than once a day, with roughly a third of the batch size and a third of the between-run inventory, on the exact same equipment. Nothing about demand or capacity changed; only the changeover did.

Faster changeovers shrink the interval and the batchSame line, faster changeover, shorter interval30-min changeoverEPEI = 2.0 daysbatch ≈ 2 days of demand10-min changeoverEPEI ≈ 0.67 daybatch ≈ two-thirds of a daySMED is the lever: shorter changeovers let you rotate the mix more often.
Cutting changeover time is the direct route to a shorter EPEI. Less changeover per rotation means the line can lap the mix more often at smaller batch sizes.

EPEI: the reference

EPEI is a defined lean scheduling term, not a vendor metric:

  • Definition. EPEI is the frequency with which different part numbers are produced in a process; if a machine is sequenced so every part is made every three days, EPEI is three days (Lean Enterprise Institute, Every Product Every Interval).
  • Changeover budget. A common planning convention allows roughly 10 percent of available time for changeovers, then sizes the interval and lot size to fit the time that remains.
  • Smaller is better. The interval should be as small as possible to produce the right mix to meet demand without building excess inventory, which makes reducing changeover time the primary lever for improving it.

What is a good EPEI target?

The best EPEI is the smallest one your changeover time and capacity can support without starving demand. There is no universal number, because it depends entirely on your mix, demand, and changeover times, but the direction is always the same: shorter. A line running EPEI of a week is a candidate for real gains; getting it to a day or a shift usually unlocks large inventory reductions and a much more responsive schedule. Treat EPEI as a metric you drive down over time, not a fixed setting. Each SMED project on the constraint gives the schedule room to shrink the interval, and each shrink shows up as lower inventory and shorter cycle time through the plant.

How does EPEI relate to takt time and cycle time?

EPEI, takt time, and cycle time answer different questions and should not be confused. Takt time is the rhythm of demand: how often a unit must be completed to keep up with orders. Cycle time is how long one unit actually takes at a step. EPEI is neither; it is how often the line rotates through its whole product mix. A line can have a fast cycle time and still have a terrible EPEI if its changeovers are slow, because it can make each unit quickly but can only afford to switch products rarely. That is exactly the situation EPEI is designed to expose, and exactly the one SMED is designed to fix. Measuring all three, and capturing changeover and run times live rather than from end-of-shift paper, is how a plant keeps the schedule honest. That kind of real-time capture is what CLS built with Harmony (see the CLS case study). No rip-and-replace, just the numbers that size your batches arriving fast enough to act on.