Estimating berth productivity from a daily container count

A row of ship-to-shore container cranes silhouetted against a hazy orange sky above a quay

If you run capacity planning for a terminal or a port authority, you already know the lag. The monthly TEU release tells you what happened six weeks ago. AIS tells you a ship arrived and left. Neither one tells you how fast the gangs were working the vessel this week. A daily box count, read correctly, gets you most of the way there without waiting on anybody's official stats.

A daily box count gives you an estimate built from what you can actually count: boxes stacked on the quay, boxes added to or pulled from specific blocks and bays, day over day. That estimate is usable, and it lands days before the port publishes anything.

What a daily box count tells you

A box count per block and bay, taken daily, gives you two things: a snapshot of what's on the ground right now, and a delta against yesterday's snapshot. The delta is what you plug into the berth productivity formula below.

If Block 4 had 340 TEU yesterday and 410 TEU this morning, and you know a vessel called at that block overnight, you've got a rough discharge or load volume for the call. Pair that with the vessel's known berth window (from the terminal schedule or AIS loitering pattern) and you can back into average moves per hour for the call.

The arithmetic is simple:

(Box count delta attributable to the vessel) ÷ (berth hours worked) = estimated moves per hour

The division is easy. Getting a clean delta is the hard part: yard housekeeping, other vessels working adjacent bays, and rail and truck gate activity all move boxes in and out of the same block on the same day, and any one of them can throw off your number. A bay-level count, not just a block total, helps here, because a single vessel call usually concentrates in a predictable run of bays rather than spreading evenly across the whole block.

Working backward to moves per hour, and where it gets messy

Three adjustments separate a rough estimate from a number you'd put in a planning memo.

First, separate restows from net moves. A box count change captures net boxes added or removed. If the crane restowed forty boxes to get at cargo underneath, those moves happened and burned crane hours, but they don't show up in your delta at all. You won't catch every restow from imagery alone. Flag vessels known for heavy restow patterns (feeder services calling multiple ports on a rotation) and treat their productivity estimates as a floor, not a true figure.

Second, confirm your berth hours against something better than the published schedule. Scheduled berth windows are optimistic. If you have any visibility into actual first-line-ashore and last-line-off times, even an approximate one from AIS, use that instead. Dividing a real box count by an aspirational time window understates productivity every time.

Third, decide what you're measuring: gross berth productivity across the whole call, or gang-level crane moves per hour during active working time. Idle time for shift changes, weather holds, and equipment breakdowns gets averaged into the first number and stripped out of the second. Most terminal planning teams care about the second, but a box count alone can't tell you when the cranes stopped moving. You'll need a separate source, usually crane operations logs if you have access to them, or at minimum knowledge of the shift pattern.

None of this makes the estimate worthless. Run the same method consistently, call after call, and the errors from restows and schedule slop tend to wash out across a quarter even if any single call is off by a bay or two.

The math is simple. What's hard is getting a reliable box count per bay often enough, and early enough, to run the calculation before the vessel is three ports further down the string. Container Counting turns very-high-res satellite and aerial imagery into a daily box count per block and bay, exported as CSV or a map layer, so you're not waiting on a terminal operator to share throughput data or guessing from AIS alone.

If you're building berth productivity estimates on a schedule tighter than the official stats allow, a daily count is worth adding to the stack.

Start a project

← Back to the blog