Managing Project Warehouses with Dedicated or Temporary Capacity

Most warehouses are permanent. A project warehouse is a building you take because a job needs it, run hard for six to eighteen months, and give back.

That temporariness changes the economics of everything around it. Nobody wants to spend three months implementing a system into a building that closes next spring, which is exactly why project warehouses so often end up run on paper and a spreadsheet — and why the operating discipline restarts from zero on every job.

This is the lifecycle: deciding you need one, standing it up, reading its occupancy curve, and closing it without losing the record.

Three arrangements, and the decision is usually made on the wrong variable — available square footage — when the deciding variable is interference.

A corner of the main facility. Cheapest, and correct for smaller programs. It fails when project freight starts competing with steady-state work for dock time, floor space and people during a burst. The tell is not that you run out of room; it is that your regular outbound slips on container days.

A dedicated project building. Right when the program is large enough to justify a crew, or when the client wants their inventory physically separated. Expensive, and the expense is fixed the day you sign, so it wants a project with a known duration.

Overflow taken on short notice. The reactive one, and the most common. Freight arrived early, or a phase slipped and inventory that should have shipped is still with you. Usually the least suitable space and always the least prepared.

The third case is the one worth planning for in advance, because it happens to almost every project and it is where the operating model tends to break — an unprepared building means unprepared locations, which means inventory in places the system does not know about.

The setup that matters is small, and doing it before the first container is the difference between a controlled project and a recovery effort.

1. Create the facility properly. A real warehouse in the system with its own zones and locations, not a location inside another building. This is what lets you report on it, transfer to and from it, and close it later without polluting the rest.
2. Generate locations from the same scheme you use everywhere else. A temporary building is not a reason to invent a temporary naming language. The people working in it may work in your other buildings next month.
3. Name real floor positions. Project freight lives on the floor, not in rack. Floor bulk areas that are actual named locations behave properly; "over by the north wall" does not.
4. Set per-user warehouse access. Decide who can see and act in this building. On a client-dedicated facility this is often a commercial requirement, not just hygiene.
5. Create the project record before the first receipt. With the client's external reference and the hold-for-install date, so the first container lands against something.
6. Reserve staging zones for the first two waves. Not the whole matrix — the schedule will change — but enough that the first sort has somewhere to go.

That is a day of work, and it is the day that decides whether month four is calm.

A project warehouse fills on a curve that a steady-state building never sees: near-empty, then a long climb as containers land, a plateau while everything waits for site readiness, then a steep drop through the install window.

Three things are worth watching against that curve, and the useful question is always about the future rather than the present.

Can the next container fit? Not "how full are we" but "how full will we be when the freight already in transit arrives". A warehouse heatmap showing occupancy and utilization by location, including which locations are empty, answers the first half; what is still inbound against the project answers the second. Both are needed.

Is the plateau lasting longer than planned? A plateau that extends is a schedule slip expressing itself as your problem. It is also the moment overflow space gets taken reactively, so seeing it early is worth real money.

Is the drop going to be fast enough? The install window compresses a warehouse's whole outbound year into a few weeks. If the release waves are sequenced but the picking capacity is not, the constraint moves from space to labour without anyone noticing until it bites.

One honest limit: a heatmap is an occupancy view. It tells you which locations are used and how heavily. It does not predict whether one specific oversized piece will physically fit one specific rack position — that judgement stays with the person standing in front of it.

Almost every project warehouse operates alongside something else — your main facility, or another project's building, or both.

Three mechanics carry most of that:

Inventory stays per facility. Stock in the project building is stock in the project building. A pooled organization-level number that spans buildings is the fastest way to promise something you cannot pick.

Transfers are recorded moves, not adjustments. Freight going from overflow into the main building is a transfer order with a source, a destination, a quantity and both endpoints on the record. Doing it as two adjustments loses the fact that it was one movement, which is exactly what you need when reconciling later.

Access is scoped per user. Who can see and act in which building. On a client-dedicated facility this is often contractual.

The organization-level view sits on top of those rather than replacing them: you can ask what a project holds across every building, while each building's own numbers stay separate and pickable.

The last ten percent of a project warehouse is the part that gets improvised, and it is where records are most often lost.

What is actually still in the building at the end:

Handled properly, that is a short close-out list. Handled by clearing the floor, it becomes a pallet of unidentified furniture that turns up in the main building six months later with no project attached.

A workable close-out:

1. Clear every remaining hold deliberately, with a disposition. A hold with no decision is an open question, not an ending.
2. Transfer attic stock to whichever facility will keep it — as a transfer, so it arrives with its project association intact.
3. Get a written disposition on surplus from the client before anything moves.
4. Close the project record, so its status says complete rather than trailing open.
5. Deactivate the facility's locations rather than deleting them, so the history stays queryable.

Point five matters more than it sounds. The project's records are the evidence base for any claim, dispute or warranty question that surfaces later — and those surface after the building is gone.

The strongest argument for running a project warehouse on a real system is not any single feature. It is that the second project starts from the first project's operating model instead of from nothing.

What carries forward:

None of that survives on a spreadsheet that lived in a building you no longer have.

That is what the positioning on the FF&E and project logistics page means in practice: the facility may change from project to project, but the operating model should not have to.