A finance director at a mid-sized industrial manufacturer asked her planning team a simple question in the first week of October: how much of our cash is currently on a ship? Nobody could answer it. The ERP knew what had been ordered and what had been received. The forwarder's portal knew where the boxes were. Between those two facts sat about six weeks of goods and nobody owned the number.
She was asking because her transit times had quietly stretched. Port congestion is now tying up roughly 12 percent of global container vessel capacity, the highest level in four years, with the pressure concentrated in East Asia, Northern Europe and the Indian subcontinent (C.H. Robinson's October market update). Scheduled capacity has not collapsed. Usable capacity has. The practical result is that the same lane, booked the same way, with the same carrier, now takes longer than it did in June, and every extra day is a day that cash sits on the water instead of in the business.
The number that falls between two teams
In-transit inventory is goods you have paid for, or committed to pay for, that have left the supplier and have not yet been received into a warehouse you control. It is inventory on the balance sheet in every meaningful sense. It carries the same capital cost as stock sitting on a rack, and in a long ocean lane there is frequently more of it than there is in the destination warehouse.
It falls between two teams because neither system is built to hold it. Finance works from the ERP, which records a purchase order and then a goods receipt, with a gap in between that it treats as a transaction in flight rather than an asset in motion. Operations works from carrier and forwarder tracking, which knows exactly where container MSKU4471882 is but attaches no value to it. Both halves exist. Nobody joins them, so the number gets estimated once a year during a working capital review and then forgotten.
That was tolerable when transit times were stable. It stops being tolerable when the days are moving, because the days are the multiplier.
Three inputs, and only one of them is hard
The calculation itself is not complicated. In-transit inventory value equals the annual cost of goods moving on a lane, divided by 365, multiplied by the average days in transit. Carrying cost is that value multiplied by your annual carrying rate. Most finance teams land somewhere between 18 and 25 percent once they add capital cost, insurance, obsolescence risk and the cost of the safety stock they hold to cover variability.
Two of those three inputs you already have. Annual cost of goods per lane comes out of the ERP with a little work on supplier and ship-from grouping. The carrying rate is a finance decision and is usually already set for warehouse inventory, so you are reusing a number rather than inventing one.
The hard input is days. Not booked days, which is what most teams reach for because it is sitting in the routing guide. Actual door-to-door days, measured from the date the goods left the supplier to the date they were received, averaged across enough shipments to be meaningful. The gap between the two is where this whole exercise earns its keep. We cover how to build that measurement properly in how to measure transit time variability across your lanes, and the short version is that you want the distribution, not just the mean.
Why your days number is probably wrong
Booked transit is a plan. Actual transit is a record of everything that happened to the plan. The two diverge for reasons that are entirely ordinary right now: a vessel waits five days for a berth at Shanghai, a carrier omits a port call to claw back schedule and your box gets transshipped through a hub nobody planned for, a typhoon shelters a fleet for four days and the whole queue comes back at once. None of those events touch the booking. All of them touch the cash.
The event that matters most is the one that happens after the goods are already moving, because that is the one you cannot plan around. A blank sailing announced before you book can be absorbed into a booking plan. An omitted port call on a vessel already carrying your freight cannot, and it is the harder case to even detect. That detection problem is the subject of blank sailings and port omissions, and the same bunching of arrivals that stretches your transit times is what then lands weeks of cargo in days at the other end, which we worked through in how to manage an inbound container surge.
This is where the data problem becomes a capital problem. The departure date usually lives with the supplier or the forwarder, the milestone events live with the carrier, and the receipt date lives in the WMS. A command center like Orkestra exists to hold those three in one record, normalizing carrier and forwarder events against the purchase order so that actual door-to-door days is a field you can read rather than a reconstruction somebody does in a spreadsheet each quarter. That is the whole point of treating in-transit inventory as a tracked position rather than a gap between two transactions. Bridgestone used exactly that normalization to put data accuracy behind production planning decisions, and found $11.6K in weekly savings across 22 shipments.
A worked example on one lane
Take a single Asia to North Europe lane carrying $9M of goods a year. That is $24,658 of value crossing the lane every day.
At an average 38 days door to door, the in-transit inventory on that lane is about $937,000. At a 20 percent carrying rate, it costs roughly $187,000 a year to hold.
Now stretch it to 42 days, which is a modest four-day slip against what congestion is currently doing to East Asia departures. In-transit inventory rises to about $1,036,000. That is $99,000 of additional cash locked up permanently for as long as the condition persists, and about $20,000 a year in extra carrying cost, on one lane.
The permanent lock-up is the part finance teams react to. The $20,000 is an expense line. The $99,000 is working capital that has left the business and will not come back until the transit time does. Multiply across thirty lanes and the number stops being a rounding error, which is precisely why it is worth measuring per lane rather than as a single network average. Averages hide the three lanes doing the damage.
What the number actually changes
Once you can see in-transit capital by lane, three decisions get easier and all three are ones operators already make, just with better evidence behind them.
Lane and mode choices stop being judged on freight rate alone. A faster service that costs more per container can be cheaper overall once you price the capital it releases, and until you have the number that argument is unwinnable in a room with procurement. It also sharpens the one-off version of the same call, whether to expedite a shipment that is already late, where the capital sitting in the box is one input and the cost of the delay is the other.
Safety stock gets sized against measured variability rather than against the worst month anyone remembers. If a lane's transit time has widened by six days at the ninetieth percentile, that is the buffer you are really funding, and you can see whether it is cheaper to hold it or to fix the lane.
And the conversation with finance changes shape. Instead of reporting that freight is up, you can show which lanes moved, by how many days, and what that did to cash. Cost intelligence of that kind is what Orkestra's analytics layer is built to surface, and it is the same modelling discipline behind cost to serve, which answers the adjacent question of which customers and orders are actually profitable once freight inputs move.
Start with one lane this week
Pick your highest-value ocean lane. Pull twelve months of receipts, calculate actual door-to-door days for every shipment on it, and compare the average against what your routing guide claims. Multiply the gap by the daily value of goods on that lane. That single number, produced in an afternoon, is usually enough to get the rest of the exercise funded.
The teams who come out of this period in good shape will not be the ones who predicted the congestion. They will be the ones who could see, week by week, what it was costing them, and could point at the lanes where it was costing the most. If you want to see how that looks when the order, shipment and receipt data sit in one place rather than three, book a demo.
