The Monument Was Never the Monument: Why Low OEE on the Wrong Equipment Buys the Wrong Capex

Opening Insight

A frozen food plant ran a blast freezer OEE report. Result: 62 percent. The plant engineer presented a capex case for a new blast cell. 3.8 million dollars to add capacity. The math showed the plant could recover about 1.2 million in annual throughput if the freezer ran at 85 percent.

The case went to committee. One question stopped it.

What percent of the freezer's 38 percent downtime is equipment failure versus starvation?

Nobody in the room had the breakdown. They had assumed downtime was failure, because the OEE report calls anything that is not run-time "downtime."

The following week, the data came back. Actual equipment failure: 9 percent. Scheduled maintenance: 3 percent. Starvation: 26 percent. The freezer was idle not because it had broken, but because cooked product had not arrived at the infeed.

The freezer was not the constraint. Whatever was feeding the freezer was the constraint. The freezer was a symptom.

Buying a new freezer for a starvation problem would have produced a second freezer that also ran 26 percent starved. Net throughput gain: zero.

The monument gets the capex. The constraint keeps the problem.

This is the pattern we see across mid-market manufacturers, especially food and CPG. Low OEE points at equipment. The equipment with the low OEE is often the most visible symptom of a constraint living upstream. The capex follows the symptom; the constraint remains.

System Context

OEE points at the equipment with the lowest utilization. That is its job.

A piece of equipment shows low OEE for three reasons: failure (it breaks), starvation (nothing arrives to run), blockage (downstream cannot accept output). Only one is an equipment problem. The other two are coordination problems that surface on the wrong machine.

Plants that use OEE as a capex signal instead of a diagnostic input routinely buy equipment to fix problems that do not live in the equipment.

Mechanism

OEE is a location metric, not a causation metric. It tells you where in the plant utilization is low. It does not identify which of the three mechanisms (failure, starvation, blockage) produced the low number. Capex cases that treat OEE as causation buy the wrong asset.

Three downtime categories behave completely differently.

Failure: the equipment broke. It needs repair or replacement. Classic constraint.

Starvation: the equipment is ready to run, but the upstream process did not deliver product. The constraint is upstream, not at the equipment.

Blockage: the equipment completed its work, but the downstream process cannot accept the output. The constraint is downstream, not at the equipment.

In the frozen food plant above, the freezer's 38 percent downtime broke down as 12 percent failure-related (9 percent failure plus 3 percent planned maintenance) and 26 percent starvation. The 26 percent starvation was driven by cook and chill room scheduling. The cook room ran batches that finished in waves. The chill room handled each wave before sending to the freezer. When a wave was still in the chill room, the freezer sat idle.

The freezer could not fix this. A second freezer could not fix it either. The fix was upstream coordination between cook and chill scheduling.

Equipment utilization is the answer. The question is: produced by what?

System Interaction

The Four Decisions above a piece of equipment control its utilization. Labor decides how many operators are cross-trained for the feeding process. Automation decides whether upstream equipment matches downstream throughput rate. Scheduling decides how work flows to the equipment. Packaging sourcing affects transition times that insert starvation windows.

When OEE is low, the diagnostic question is which of the Four Decisions is producing the starvation or blockage. The equipment is the reporter. The Four Decisions are the reality.

In the frozen plant: scheduling was the binding decision. The cook-chill sequencing did not match the freezer's steady-state intake. Labor was secondary. Automation and sourcing were not factors.

A capex gate that asked "which of the Four Decisions is producing this low OEE?" would have redirected the 3.8 million toward a scheduling redesign that cost 50 to 100 thousand.

Economic Consequence

Approved on clean math. Running on messy reality.

Misidentified-constraint capex is structurally expensive because the capex does not move the constraint.

For a typical 2 to 5 million dollar misidentified-constraint capex:

• The new equipment runs at the same low utilization as the old, because the upstream coordination did not change
• The approved throughput improvement (often modeled at 30 to 50 percent of the starved capacity) does not materialize
• The plant's year-one realized return is usually 5 to 20 percent of the modeled case
• The coordination fix, had it been made instead, typically costs 3 to 8 percent of the capex and would have closed 40 to 70 percent of the constraint gap

Misidentified-constraint capex cannot be corrected by more capex. Every subsequent capital investment inherits the diagnostic error as its baseline assumption.

The second-order cost is that the capital is gone. The plant now has the wrong equipment at the wrong place, and a scheduling problem it never addressed. Next year's capex cycle produces the same starvation showing up on the new equipment, and another capex case is built against the same wrong constraint.

Diagnostic

The test is a 30-minute exercise.

For the most-complained-about equipment in the plant, pull the downtime log. Most plants already track it.

Categorize each downtime event into one of three buckets: failure, starvation, blockage.

If more than 70 percent is failure, the equipment is the constraint. Capex to replace or upgrade is directionally correct.

If 30 to 70 percent is starvation or blockage, the equipment is a shared constraint. Some of the problem lives there. Some lives upstream or downstream. Investigate both before writing capex.

If more than 70 percent is starvation or blockage, the equipment is not the constraint. The constraint is upstream or downstream. Capex to the equipment will not move throughput.

Decision Output

• Decision type: Constraint diagnostic before capex approval
• Trigger: Any equipment capex case justified by low OEE or "the equipment is our bottleneck"
• Action: Require a downtime-category breakdown (failure, starvation, blockage) for the target equipment. If starvation or blockage exceeds 30 percent, the capex case must include an upstream-coordination analysis and a revised throughput model.
• Tradeoff: Adds 2 to 4 weeks to the capex approval cycle. Surfaces 30 to 60 percent of "equipment" capex cases that are actually coordination problems wearing equipment clothing.
• Evidence: Plants that gate equipment capex on downtime categorization approve 40 to 60 percent fewer equipment capex cases. The ones they approve hit modeled throughput within 10 percent.

Framework Connection

The Monument Was Never the Monument. The visible symptom is not the cause. The equipment with low OEE is the reporter of a constraint that lives elsewhere, usually in the coordination logic that feeds or receives from it.

This is the diagnostic layer of the Silo Tax. The first layer is the math. The second is the org design. The third is the diagnosis itself, which problem you think you have.

Plants that read OEE as diagnosis buy the wrong assets. The Four Decisions framework reframes OEE from a verdict to a signal: "something is producing low utilization here. Which of the four decisions is responsible?"

The machine reports. The Four Decisions decide.

Strategic Perspective

Most mid-market manufacturing capex is written against symptoms. The capex case looks solid because the symptom is real. The equipment really does have low OEE. The line really does miss throughput targets.

But solving the symptom and solving the constraint are not the same thing. The plant that treats low OEE as a diagnostic answer is paying capex to solve a problem that its capex cannot solve.

Equipment OEE measures where the symptom shows up. It does not measure what produced the symptom.

For a mid-market CFO, the test to add to every equipment capex review: before we approve this, show me the downtime breakdown into failure, starvation, and blockage. If starvation or blockage is above 30 percent, the capex case is incomplete.

Factories that think diagnose the constraint before they buy the asset.