EPEI and supermarket sizing: reducing WIP without starving the line

Every lean facility eventually faces the same tension: you want to hold as little finished goods inventory as possible — because inventory costs money and hides problems — but you also need enough stock to keep your downstream processes running without interruption. Too little and you starve the line. Too much and you've built a warehouse you didn't need.

EPEI — Every Part Every Interval — is the scheduling tool that navigates this tension. It tells you the minimum interval at which you can produce every part number at least once given your available capacity and changeover constraints. And when EPEI is properly calculated, it gives you the data you need to size your supermarket correctly.

What is EPEI?

EPEI is the shortest production cycle that allows every part number to be run at least once. It's expressed in days (or shifts). An EPEI of 3 days means that on any given day, the parts produced three days ago are now available — and every part number in your schedule gets a production run within every 3-day window.

EPEI answers the question: how often can we cycle through our full range of part numbers? The lower the number, the more frequently you replenish, and the less inventory you need to hold.

Why changeover time is the key constraint

EPEI is fundamentally limited by changeover time. Every time you switch from producing Part A to Part B, you lose productive capacity. That lost time has to come from somewhere — and in a pull system, it comes out of the time available for replenishment.

If you have 420 minutes of available time per day, your total run time across all parts is 350 minutes, and your total changeover time (cycling through all part numbers once) is 90 minutes:

EPEI = 90 ÷ (420 − 350) = 90 ÷ 70 = 1.29 days

You can cycle through every part number approximately every 1.3 days. Use the EPEI calculator to run this for your own part mix — it handles multiple part numbers with individual cycle times, changeover times, and daily demands.

What happens when EPEI is too long

A long EPEI means infrequent production runs, which means large batches, which means large supermarkets. Consider the difference:

The supermarket size scales directly with EPEI. Halve your EPEI and you can halve your finished goods inventory — without any risk of starving the downstream line, because replenishment happens twice as often.

How to size a supermarket using EPEI

A supermarket needs enough stock to cover demand during one full EPEI cycle plus a safety buffer. The formula for each part number is:

Using the example above (EPEI = 1.29 days, Part A daily demand = 100 units, safety factor = 0.5 × EPEI):

Stock = 100 × (1.29 + 0.65) = 100 × 1.94 = 194 units

That's your maximum supermarket stock for Part A. When stock drops to the reorder point (daily demand × EPEI = 129 units), the kanban signal triggers a production run.

The relationship between EPEI and takt time

EPEI and takt time work together. Takt time sets the production rate — how fast you need to produce each unit. EPEI sets the scheduling frequency — how often you cycle through your full product range. Both need to be respected simultaneously.

A common mistake is to calculate an aggressive EPEI target without checking whether the machine has enough capacity to meet demand at that interval. If your total run time already consumes 95% of available time, there's almost no room for changeovers — EPEI will be very long regardless of how efficient your changeovers are.

The right sequence is:

1. Calculate takt time to confirm the machine can meet total demand
2. Calculate current EPEI from actual changeover data
3. Set a target EPEI based on acceptable supermarket size
4. Identify the changeover reduction (SMED) needed to close the gap

Reducing EPEI in practice

Since EPEI is limited by changeover time, the primary lever for reducing it is changeover reduction — specifically SMED (Single-Minute Exchange of Die). SMED separates internal changeover activities (must happen while the machine is stopped) from external ones (can be done while the machine is running), and systematically converts internal time to external.

A 30% reduction in total changeover time typically translates to roughly a 30% reduction in EPEI — which means a 30% reduction in finished goods inventory, all else equal. That's often a very attractive return on a focused SMED workshop.

Track the impact of SMED improvements using the hour-by-hour tracker — monitoring actual output against planned targets during changeover-intensive shifts will show you whether the improvements are sticking in practice.

EPEI in high-mix low-volume environments

EPEI is most powerful — and most challenging — in high-mix low-volume environments where you have many part numbers, each with relatively low daily demand. In these environments, EPEI can easily stretch to weeks if changeover times are long relative to run times.

The practical implication is that you have to be selective. Rather than calculating a single EPEI for all parts, group parts into families with similar setups. Parts within a family share most of their changeover — the changeover between them is short. Parts between families require a full changeover. Run each family as a block, and calculate EPEI at the family level rather than the individual part level.

EPEI is not a one-time calculation. Demand changes, part numbers are added and removed, process improvements change cycle times. Recalculate EPEI whenever any of these inputs change significantly — and review your supermarket sizes at the same time. A well-maintained EPEI is one of the most powerful levers for reducing inventory without service risk.