ERP and WMS projects often look healthy in conference rooms and fall apart on the floor. Receiving cannot post. Picks stall. Labels print wrong. Supervisors invent a paper path to keep trucks moving. Once that paper path works, the new system has already lost, even if IT calls the cutover a success.
The floor is the real go-live
Finance and leadership care about clean postings. The building cares about whether a pallet can move without a workaround. We design cutovers around floor roles first: receiving, putaway, picking, packing, shipping, cycle count. If those roles cannot complete a full shift in the new system under realistic volume, you are not ready, no matter what the project plan says.
Parallel run like you mean it
A parallel run is not a checkbox. It is old and new producing comparable outcomes for a defined window, with differences triaged daily. Include ugly orders, returns, and partial receipts. Time the work. Listen for where people leave the system. Those exits are your training curriculum and your defect list.
Mid-market multi-site programs should pilot one site, then replicate the playbook. SMB manufacturers and distributors can often run a shorter parallel if scope is tight. New businesses implementing first-time WMS should resist big-bang heroics; start with the flows that ship revenue this month.
Degraded mode is part of the design
Write what happens if scanning fails, if the label printer dies, if inventory is wrong at a bin. Who can authorize overrides. How you catch up without duplicate movements. Tape that plan where supervisors stand. If degraded mode only exists in a slide deck, the floor will invent a worse one under pressure.
Change management for supervisors, not only end users
Supervisors make or break adoption. Train them before the crew. Give them a short list of "never go back to paper unless" rules. Staff a floor support rotation for the first weeks, not only a help desk queue. Celebrate retired workarounds the same way projects celebrate go-live.
Cutover readiness checklist
- Each floor role completed a timed rehearsal at realistic volume.
- Exception types have owners and queues.
- Label, scanner, and print paths proven on production hardware.
- Inventory authenticity checked on sample locations.
- Degraded-mode one-pager posted and practiced.
- Parallel differences trending down, not sideways.
What to do this week
Walk the floor with the project plan in your pocket. Ask each supervisor where they still expect to use paper on week one. Turn those answers into rehearsal scripts. Do not debate cutover dates until those scripts run clean. Shipping through a bad go-live costs more than moving the weekend.
Data readiness the floor can feel
Bad master data shows up as cannot find item, cannot trust qty, cannot ship. Clean items, units, locations, and customer ship-to data for the flows in scope before you chase exotic features. A shorter scope with honest data beats a wide scope that forces workarounds on hour one.
Cycle count accuracy on a sample of high-velocity locations is a go or no-go signal. If the sample is messy, cutover will teach everyone the wrong lesson: the system is wrong, paper is safer.
Integration points that strand shipments
Label formats, carrier rate shops, ecommerce order intake, and accounting postings each create floor pain when they fail. Test them on production-like hardware. Include peak print volumes. Include the order types that only appear on Tuesdays. Projects that only test happy-path wholesale orders discover their real customers in week one of production.
Support model for the first fifteen business days
Floor walkers with authority to decide. A visible board of top defects. A daily fifteen-minute stand-up that includes operations, not only IT. A rule for when to use degraded mode. Without that, every issue becomes a private Slack thread and learning never compounds.
Change management is not a go-live speech. It is supervisors coaching the new path while the old path is intentionally harder to use. If the old terminal is still in the same spot with the same muscle memory, people will return to it under stress. Remove or block it on the cutover plan, with a documented override.
Role rehearsal script outline
Receiving: appointment, ASN or blind receipt, putaway to a directed location, exception for damage. Picking: wave or order pick across a hot SKU and a slow SKU, short pick path, short ship confirm. Packing: label, weight, void and reprint. Supervisor: override, recount, release hold. Time each. Note every time someone asks a neighbor instead of the system. Those questions become job aids.
Run the script on the equipment people will actually hold. A rehearsal on a conference room laptop does not count. If scanners are new, battery and wifi failure modes belong in degraded mode docs.
What to do this week
Schedule the role rehearsals on the calendar with named supervisors. Freeze scope changes that affect floor flows until rehearsals pass. Publish the degraded-mode one-pager draft and walk it with the people who will use it. Cutover dates should move for failed rehearsals, not the other way around.
Want more of this in your inbox? Subscribe to our newsletter.
If go-live plans skip the floor rehearsal, we should fix that before any cutover weekend.
Book a Call