Everyone Did Their Job. The Project Was Still Six Weeks Late.

There is a specific kind of meeting that follows a hardware project slipping its date, and it is one of the least productive hours in corporate life.

Everyone is in the room. The vendor shows the build finished on schedule. The forwarder shows the container was booked and sailed on schedule. The broker shows the entry was filed and the goods released. The installer turned up on the day they were asked to turn up. Every party has a timestamp, every timestamp is clean, and the racks are still empty six weeks after the go-live date.

So the meeting ends the way these meetings always end. Someone says it was a perfect storm. Someone says we need better visibility. Nobody says the true thing, which is that the delay did not happen inside anyone's scope. It happened in the space between scopes, and that space had no owner.

The gap between the purchase order and the rack

A procurement dashboard tracks two events with any confidence: the day the purchase order was signed, and the day the goods arrived. Everything in between runs on other companies' systems. The vendor's factory schedule. The forwarder's booking platform. The customs entry. The last-mile carrier's dispatch board.

The buyer sees none of it. A project five days behind in week two stays invisible until week eight, when the arrival date moves and the slip has already compounded across three handovers. By then it is not a five-day problem any more.

Carra Globe calls this stretch the Silent Middle, and their breakdown of it is the clearest map of the territory I have seen: six stages from purchase order to installed rack, with realistic 2026 durations, the specific failure point in each stage, and an honest column on which ones are actually preventable. If you run infrastructure projects, read it before you plan your next one rather than after.

The handovers are the problem, not the stages

The stages themselves are mostly fine. Manufacturing is predictable, because the vendor commits to a date and generally meets it. Ocean transit takes as long as ocean transit takes. Customs, for a clean pre-filed entry, releases in hours. Installation takes the weeks it takes to cable, configure and validate.

What is not fine is the seams. Freight gets booked, but the pre-filing documents never reach the broker, so the entry starts cold on arrival and a shipment that should have cleared in hours takes a week. Customs releases the goods, but nobody tells the importer to accept them. The goods clear, but the last-mile carrier was never scheduled because the release notification went to an inbox nobody watches.

Every one of those is a handover failure, and every one is invisible on a status call, because each party is honestly reporting that their own bit went fine. And it did.

Three failures that account for most of the damage

Pre-filing that never happened. Entries can be lodged and released before the shipment lands. When that happens, clearance is a formality. When it does not, the broker begins from a standing start at arrival, which typically adds three to seven days to goods that were perfectly compliant. Most importers never ask their forwarder whether pre-filing is standard practice, and a surprising number of forwarders treat it as optional unless someone insists.

The importer of record appointed after arrival. Someone must legally take title in the destination country and authorise onward movement. If the buyer has no entity there, that role falls to a third party. Appointed before the goods ship, it is a same-day step. Discovered at the port, it becomes a week of due diligence and licence checks while the cargo accrues bonded storage charges. This is retro-engagement, and it is the single most expensive avoidable mistake in the middle of the chain.

The install team booked against a date that no longer exists. Engineers are scheduled against the original planned arrival. The shipment slips a fortnight, the engineers are now on another project, and the next window is a month out. A two-week transit delay becomes a six-week delay to go-live, and the cost is not the freight, it is the idle infrastructure.

What actually fixes it

Not more visibility. Not a better dashboard. Not another weekly call where four vendors each confirm their own bit is fine.

What fixes it is making one party accountable across the whole middle, from freight booking through customs, importer of record acceptance and last-mile delivery. It cannot be the supplier, who is finished at the factory gate. It cannot be the forwarder, who is finished at the port. It cannot be the broker, who is finished at release. It cannot be the carrier, who is finished at the dock. It has to be a party whose scope spans all four, because the failures live at the boundaries between them.

The operational moves that follow from single ownership are unglamorous and cheap. Pre-file every entry before arrival. Confirm the importer of record before the goods leave origin. Schedule the last mile against the actual release notification rather than the planned date. Keep the paperwork consistent at every stage, because one inconsistency drops the shipment into the exception queue at the border.

None of that is difficult. It is simply what happens by default when someone owns the middle, and what never happens when four vendors each own a slice of it.

The question to ask at kickoff

Next time you map an infrastructure project, do not just list the stages. List the handovers between them, and write a name next to each one. If any handover has no name, that is where your project will lose its time, and you have just found it four months early.

Carra Globe operates that single-owner role for technology and data centre hardware as importer of record across 175+ countries, holding the filings, the licences and the delivery under one accountable party. For the full six-stage timeline, including realistic durations and the failure point at each handover, read the complete guide: The Silent Middle: what actually happens to your hardware between the purchase order and the rack.