OEE for Bottling Lines

OEE for a bottling line is measured at the filler, the machine that sets line rate. Because the upstream depalletizer and rinser and the downstream capper, labeler, and packer all run faster and are decoupled by accumulation, the filler's Availability × Performance × Quality, including the short starve and block stops it suffers, is the honest measure of a bottling line's output.

A high-speed bottling line is a chain of fast machines linked by conveyors and accumulation tables, engineered so no single stop halts everything at once. That design is what makes bottling OEE tricky: the buffers that keep the line running also hide where the time goes. This guide shows why the filler is the measurement point, how accumulation reshapes what you see, and how to keep micro-stops from disappearing. For the base method see the OEE calculation; for the liquid-process side of a beverage plant, see OEE for beverage production.

Why measure bottling OEE at the filler?

Measure at the filler because it is the constraint. A bottling line is deliberately built with the filler as its slowest, most valuable machine, and every other station is sized to run faster so it can catch up after a stop. That makes the filler the pace-setter: line output equals filler output, so the filler's OEE is the line's OEE. This is the same constraint logic as OEE for bottleneck machines and theory of constraints applied to a specific and very common line.

Measuring OEE at the capper or labeler instead tells you mostly whether those machines were starved or blocked by the filler, not their own effectiveness. Keep rigorous OEE at the filler and simple downtime tracking at the other stations, where the useful question is whether they ever starved or blocked the filler.

How does accumulation change what OEE you see?

Accumulation changes OEE by hiding short stops. Back-pressure accumulation tables let bottles slip against a moving belt or pool in a buffer, so a brief stop at the capper does not immediately stop the filler, the filler keeps feeding into the buffer until it fills, then blocks. Sensors near the infeed and discharge slow or speed the upstream machines to keep the buffer in range. The result is smooth flow, but also that many small stops never become clean, loggable filler-downtime events.

Instead, those seconds show up as the filler running below rate, a Performance loss, where they are easy to overlook and impossible to fix by hand. This is the classic minor-stop problem in the six big losses: individually trivial, collectively the largest loss on many bottling lines. The only way to see them is to count from the machine, so the difference between a slow second and a micro-stop is recorded rather than averaged away. This is exactly why machine downtime on a bottling line must be captured automatically.

What are the biggest availability losses on a bottling line?

The biggest Availability losses are changeovers and jams. A size or label change means mechanical adjustment across the filler, capper, and labeler; a product change may add a rinse or clean. These are planned-production time the line did not produce, so they belong in Availability, and they are usually the most improvable loss because SMED can cut setup time sharply without capital. Jams, a fallen bottle, a mis-fed cap, a bridged infeed, are the other big Availability drain, and on a buffered line they only stop the filler once the buffers around them empty or fill.

Because accumulation delays when a jam reaches the filler, the honest accounting ties each filler stop back to the station that actually caused it. A capper jam that eventually blocks the filler is a capper problem showing up as filler downtime, and attribution is what keeps you from “fixing” the wrong machine. Balancing station speeds and buffer sizes so no station chronically starves or blocks the filler is a line balancing job.

How do you handle micro-stops and speed losses?

Handle them by counting from the machine and splitting minor stops out of raw speed loss. A hypothetical worked example makes the stakes clear:

Availability = 390 ÷ 450 = 86.7%; Performance = 140,400 ÷ (390 × 400) = 140,400 ÷ 156,000 = 90.0%; Quality = 137,592 ÷ 140,400 = 98.0%; OEE = 76.5%. The 10-point Performance gap is where the micro-stops hide. If you cannot say how much of that gap is genuine slow-running versus thousands of two-second starve/block events, you cannot fix it, and on bottling lines it is almost always the micro-stops.

What are the quality losses on a bottling line?

Bottling quality losses are fill level, cap application and torque, seal integrity, and label placement. Only bottles right the first time count as good, the same rule as first pass yield: a short-fill caught by the checkweigher, a cocked cap, a crooked label, each is a Quality loss, and most cannot be reworked at line speed. As on other filling lines, watch fill giveaway too: the extra volume dosed above target to stay above the labeled amount consumes product and filler time without ever tripping a reject, so it hides from OEE and needs separate tracking. For food and beverage bottling specifically, tie the OEE view into food manufacturing software and CPG software and into packaging line automation for the downstream case-packing.

How do you stand up bottling-line OEE?

Build it at the filler, in order:

1. Set the filler as the measurement point. One line, one constraint, one OEE. Log downtime at the other stations.
2. Fix the filler nameplate per format. Bottle size and product change the rate; do not carry one number across formats.
3. Count from the machine. Automatic counts are the only way to separate micro-stops from genuine slow-running in Performance.
4. Attribute filler stops to their cause station. A capper jam that blocks the filler is a capper problem, not a filler problem.
5. Attack changeovers with SMED. Setup reduction on the filler and companions is usually the highest-return Availability work.
6. Reject on first pass at the checkweigher and vision. Fill, cap, seal, and label rejects are Quality losses, not rework.
7. Track giveaway beside OEE. It is money the percentage cannot see.

What is a realistic OEE for a bottling line?

A realistic bottling OEE depends on format mix, changeover frequency, and how honestly you count micro-stops. Use these reference points for context only:

• ISO 22400-2:2014 defines OEE and its time-states precisely, so two bottling lines using it reach the same number from the same facts (ISO 22400-2).
• The 85% world-class figure is Nakajima's 1980s TPM reference point, not a certified standard; commonly cited food-and-beverage line averages sit near 45–65%. Treat both as folklore, see what a good OEE score is.
• U.S. manufacturing capacity utilization was 75.7% in May 2026 per the Federal Reserve's G.17 release (Federal Reserve G.17), a reminder that real plants run below theoretical maximums by any measure.

The honest use of bottling OEE is trend and decomposition: is the filler's OEE rising, and did the gain come from changeovers, micro-stops, or rejects? A number you trust beats a flattering one, which is why computing OEE from the filler's own signals, with micro-stops counted and stops attributed to their cause station, matters (see the platform and the CLS case study). Then run your own inputs through the OEE calculator.