A warehouse relocation fails in the quietest way possible: not with a truck breaking down, but with an order that has nowhere to go. Think of it as changing a tire on a moving truck. The truck cannot stop, the tire still has to come off, and one wrong move rolls the whole thing over. The move is not a test of muscle. It is a test of whether the cutover math adds up.
Relocation Math First: Phased Move or Big Bang?
A phased move and a big-bang move differ less in freight cost than in who carries the stockout risk. A big-bang move is like pulling the power plug and plugging it back in somewhere else: the old site stops, the new site is not warm yet, and every order in between just waits. A phased move is like a rolling upgrade on a running line: slow movers go first, bestsellers go last, and at any moment one site can still ship. It is the same logic as swapping a conveyor motor: you cannot change it without stopping, and you cannot afford to stop, so you change it one section at a time.
Standard practice is to start planning 6 to 12 months ahead, with the physical move itself taking 1 to 4 weeks (source: ShipBob warehouse-move guide). That window is not guesswork. Rack positions have to be re-surveyed, WMS location maps rebuilt, and carrier routes changed one by one. Miss any of those, and cutover day will expose it.
The dividing line is simple: can your orders stop? A small site doing a few hundred orders a day can halt shipping for a weekend and cover it with pre-built stock, so a big-bang move is actually cleaner. A site doing several thousand orders a day, with service-level promises to big customers, cannot stop for a day without breaching contracts. For that site, a phased move is the only option. This is where the fake requirement shows up: someone will argue that a big-bang move is "cleaner and done in one shot." Cleaner for the moving company, maybe. The orders it strands are yours, not theirs. Convenient and cheap are not the same thing.
Illustrative example: an e-commerce warehouse shipping 4,000 orders a day. A big-bang move needs a 2-day shipping halt, losing 4,000 orders a day, plus 3 more days to clear the backlog. A phased move adds 6 weeks of parallel-running rent and labor, but not a single order stops. The lost orders and the customer trust burned in those 2 days usually cost far more than 6 weeks of parallel running. Run both numbers before you pick.
| Approach | Strength | Cost | Fits whom |
|---|---|---|---|
| Big bang | Short cycle, one site to run after | Shipping blackout, risk concentrated | Small sites that can pre-build stock |
| Phased | Orders never stop, risk spread across waves | Double running cost, longer cycle | High-volume sites with service promises |
| Natural switch | Almost no in-transit gap | Old site drains slowly, two sites to run for months | Fast movers who refuse stockout risk |
Equipment and systems have to reach the new site before the stock does. Conveyors, AMRs, and pack stations take a week or two to dismantle, install, and tune, and they must never arrive on the same day as the first real freight. The right sequence is equipment in first, powered up and dry-run with no load, the WMS dual-site configuration switched on, and a test order walked through the whole chain at the new site. Only then does the first real truck roll in. Shipping real orders before the equipment is proven means using your customers as test dummies for the install.
Two Sites, One Ledger
The easiest mistake in the parallel-running period is counting the same stock twice. The WMS deducts it from the old site but has not added it to the new one, and sales sees total stock shrink by a truckload. Or both sites add it, and overselling follows within hours. The fix is a separate in-transit status: once stock has left the old site and has not yet been received at the new one, it belongs to neither site's available stock. It is in transit, full stop. Reconcile three numbers at a fixed time every day: physical stock at the old site, stock in transit per truck, and stock received at the new site. The three must add up to the pre-move count. One carton short, stop and find it. Never carry the gap into the next day.
Illustrative example: a 3PL warehouse in East China, book stock 12,000 cartons, physical count 11,940. The 60-carton gap goes on a variance hold for investigation; the new site is not allowed to inflate its count to make the books balance. The first truck leaves with 2,000 cartons and the new site signs for 1,988, with 12 damaged and held for inspection, listed separately as in-transit variance. Stock you cannot account for should be admitted as short, never papered over by the system.
The double cost of parallel running belongs in the cutover math too: two rents, two crews, two carrier lanes, plus maintaining the WMS dual-site setup. Many owners see only that the new rent is lower than the old one and conclude the move saves money, then discover the first two months cost double. Savings come after the move. Parallel running is money spent to buy zero stockouts. Whether that money was well spent depends on how tight you kept the window: slow movers early, bestsellers late, and 4 weeks of overlap beats 8 weeks every time.
People have to move in waves just like the stock. Veteran staff know the bestseller locations and the exception handling by heart. They should stay at the old site until the last bestseller truck has left, then transfer to the new site to train the newcomers. Do it the other way round — pull the veterans to the new site first and leave rookies to close the old one — and the busiest days of the cutover land exactly when the old site is shortest on skill. Rosters, like freight, move in waves, never in one cut. Every hour of labor saved on the tail crew comes back doubled as error rate on cutover day.
Relocation Cutover Day: Connect First, Disconnect Later
Cutover day has one iron rule: connect the new site and prove it works before you disconnect the old one. Switch carrier routing to the new address, run real orders for a day or two, and walk the full chain — pick, pack, ship, proof of delivery — before you close order intake at the old site. The returns address moves even later: keep the old address live until the last return has been received. Deleting it early leaves returns with nowhere to go.
The counterintuitive call is this: the best money in the whole move goes to counting, not to the movers. Too many budgets put the big line against trucks and loaders, and leave two people to reconcile stock overnight. Start moving with dirty books and you carry the same errors into a brand-new building. The racking can be beautiful and the system stock still wrong. Money spent on a proper count is the highest-return line in the entire cutover budget.
After the switch comes a 4 to 8 week stabilization period (source: US Cargo Link warehouse-move checklist). Watch pick accuracy and order error rates every day until the numbers are back to pre-move levels. That is the finish line. Handing back the old site keys is not.
Was the money worth it? Look at one number: the daily on-time dispatch rate through the switch. Hold it above 99% for the whole period and every cent was well spent. Let it slip below 95% and the cutover math was wrong from the start. Every cent of parallel-running cost is insurance bought for that one number.
By Da Liu, Automation Engineer







