Here is the conclusion first: the most dangerous thing in the first week of a WMS go live is not a software bug. It is a problem nobody can find an owner for, a severity nobody can agree on, and a rollback nobody dares to approve. A WMS go live war room exists to solve exactly those three things: where the people are, who decides the severity, and who signs off on a rollback. That is what WMS go live war room preparation is really about, and the full preparation list is below, but lock this in first: the night before go-live, freeze the people, the access rights, the severity rules, and the rollback plan, then run the same check sheet every day of week one, and you will actually sleep. Going live without that preparation is a mistake I have watched three times, and the script is the same every time.

A War Room Is Not a Room. It Is a Shift Roster

Many people think a war room means clearing out a meeting room and setting up a few laptops. Wrong. The core of a war room is a shift roster: who is on site, who is on call remotely, who can make decisions, when shifts change, and whose phone rings at 2 a.m. when something breaks. Our way of doing it: two shifts a day for the first week, one team on site during the day and one on call remotely at night, and the person in charge must be someone present who can actually decide, not a liaison who can only pass messages along. The customer, the implementation team, and the warehouse floor each supply one decision-maker. Miss one of the three and the war room is decoration.

One client insisted the night before go-live that the system had been fully tested and a war room was unnecessary. On the morning of day two, wave release froze and pick lists stopped coming down. Dozens of people stood around waiting while it took half an hour to reach the implementation consultant, another hour to reach a developer, and another half hour to get the customer's boss to approve a rollback. A whole morning gone. This is the mistake I have seen three times, and the script never changes: the system was not the problem. The problem was that when something broke, nobody could decide immediately. So every cell on that roster needs a name and a phone number, not a department name. A department name does not answer the phone at 2 a.m.

Severity Levels in the Go Live War Room: How to Grade P0, P1, and P2

Problems will arrive in week one. That is guaranteed. What matters is grading them. P0 means the system is down or nothing can ship; someone must respond within fifteen minutes and the owner walks in. P1 means part of the system is broken but there is a workaround; a temporary fix is due the same day. P2 covers experience and reporting issues; log them and schedule the fix. Write the grading rules down the night before go-live and stick them on the wall. Nobody improvises severity on the spot. The worst failure mode is everything getting called P0. Shout P0 often enough and nobody believes it, and when a real P0 arrives, nobody moves.

The escalation path needs the same treatment in writing. A floor operator reports to the shift lead, the lead grades it, a P0 pulls the owner straight into the group chat, and if it is not fixed in half an hour it escalates to the project lead. Every level needs a named person with a phone number. Never write "contact the vendor." That sentence helps nobody. Daily rhythm: a fifteen-minute stand-up half an hour before the end of the day, covering only three things: how many issues came in today, how many are still open, and what scares us most about tomorrow. If the stand-up runs past fifteen minutes, someone is dodging blame. Stop it and move on.

The Night-Before Checklist: Every Item Is Mandatory

The night-before checklist is a ritual I never skip on any project. Data: is the master data frozen, has anyone changed it since the freeze, and does it need re-checking? Access: can every role log in, or will someone discover a wrong password on go-live morning? Documents: are all test orders cleared out, because test data leaking into the live database is the most common accident I see. Interfaces: are the connections to the ERP and the parcel carriers alive, and is someone watching them? Printing: print one label from every label printer and every shipping-label printer, and never trust "it worked last time." Every item gets a tick and a signature. An item nobody signed is an item nobody checked.

The rollback plan deserves its own section, because it is the easiest thing to fake. A rollback is not a failure. It is insurance. First define the triggers: under what conditions do we roll back, and who has the authority to call it? Then write the steps down to the level of which step switches back to the old system and which step tells the floor to stop the wave, with an owner for every step. Most important is the rehearsal: walk through the rollback at least once the night before go-live. Do not discover on the real day that the old system's accounts expired months ago. A rehearsed rollback stays calm when it is actually used.

What to Check Each Day of Week One

Day one is about outbound flow: wave release, picking, verification, packing, and shipping. Check every step for blockages and count how many orders are stuck. Day two is about inventory: does the system quantity match the physical quantity, and is anyone working the discrepancy list? Day three is about interfaces: ERP confirmations, carrier labels, inventory sync. Look for dropped orders. From day four, check efficiency: picks per person, the share of exception orders, and whether those numbers are better or worse than before go-live. Write the daily results into a short report. Keep it to five numbers or fewer and watch them closely. That daily rhythm is the part of WMS go live war room preparation nobody photographs, but it is the part that keeps week one from drifting. At the end of week one, hold a review meeting and sort everything into three buckets: system defects, processes that were never explained clearly, and training gaps. Record them separately and fix them separately.

One more trap: do not add new features in week one. Customers always want to strike while the iron is hot, but my rule is that week one is for fixing defects only. Features can wait until week two, after the system is stable. Week one has a single goal: orders ship, the books balance, and nobody burns out. This is a mistake I have seen three times, and every time the team added a feature in week one, week two turned into a chain of fires. Another client is the classic case: on day three they asked for one extra report, the change broke inventory sync, and the whole afternoon went to firefighting.

One last line: the better prepared the war room is, the more boring week one feels, and boring is the best possible sign of a successful go-live. Good WMS go live war room preparation does not make week one exciting. It makes it forgettable, and forgettable is exactly what you want. Is your war room roster posted yet? If not, post it today. Do not wait until the fire starts to look for people.

By Lin Xiao, WMS Implementation Consultant