Before you buy the WMS

The shortlist arrives before the process does. Three vendors. A scorecard. A site visit to a warehouse that looks nothing like yours, running a catalogue that looks nothing like yours, with a supervisor who has been on this system for five years and still makes it look easy. You leave wanting the easy.

Then you buy it. Then you discover that the easy was a process, not a product. The product was just sitting on top of a job somebody had already decided.

I've been in those site visits. I've also been in the month after go-live, when the scorecard is a souvenir. This is not an argument against a WMS. It's an argument about what has to be true before the purchase is a good idea.

Before you buy the WMS, buy the sentences. If you cannot say how the work runs, you are shopping for a narrator, not a system.

A WMS is a very strict intern

It will require a scan. It will require a location. It will require that the stock in the bin and the stock in the screen are the same story. That strictness is the value. It is also why a messy operation hates it, then works around it, then says the WMS "doesn't fit how we work."

How you work might be the problem. A WMS will not politely absorb two definitions of available, a pack bench that isn't a pack bench, and a returns process that is actually a pile. It will either freeze, or it will be overridden by a superuser until it is a very expensive notepad.

The question is not whether the intern is clever. Most of the decent ones are clever enough. The question is whether you are ready to be that strict on a Tuesday when a marketplace order is late and a director wants an override.

The checklist nobody wants

Can you walk one SKU from receiving to despatch without calling the person who "just knows"? Including putaway. Including the pick face that isn't really a pick face. Including the carton that doesn't match the master data.

Do you know what "in stock" means? On hand, available, reserved, inbound, damaged, returns pending, marketplace reserved. If two of those are the same number in a spreadsheet, you are not ready. A WMS will make you pick, and the first week will be a theological argument dressed up as implementation.

Do you have locations that exist in real life? Not "somewhere in A aisle." A WMS that is told to put something in a place that is actually a walkway will do it, then you will spend a month hunting.

Do you have the ugly five exceptions written down? Short pick. Damaged. Unidentified return. Barcode that belongs to two SKUs. Order that changes after it is already waving. If those are still vibes, the WMS will open a queue named exception and leave it there, forever, as a monument to the brief you didn't write.

Vendors will not ask this hard enough

They will ask about volumes, integrations, and whether you need VAS. They will demo the happy path. They will tell you their implementation team has seen it all. They have seen a lot. They have not seen your folklore, and they will not extract it unless you put it on the table.

A 3PL selling you "their WMS" is the same trap in a different shirt. You are buying their standard. If your product is not their standard, you will either change your product at the pack bench, or you will live in a set of custom workflows that only one person at the partner understands. Ask them to configure your ugly five, in the first workshop, not as a phase two that never happens.

If the workshop cannot survive a failed barcode, the go-live will not survive a peak.

What before actually looks like

Map the real process on the floor. Kill the parallel version. Master data that a human would trust. Locations that exist. A decision on whether you will be strict. If you will not be strict, do not buy a strict system. Buy better discipline in the spreadsheet and hire the supervisor. That sentence is unfashionable. It is also cheaper than a Ferrari in the carpark.

Pick the WMS for the process you can already explain, not for the roadmap slide. Wave, batch, cluster, putwall, whatever. Those words are not strategy. They are mechanics. Mechanics come after the job is named.

At 500 parcels a day you might not need it yet. At 5,000 you might be late. The trigger is not a round number. The trigger is when the hero cannot hold the locations, the exceptions, and the truth about stock in their head. Buy then. Not as a substitute for the hero. As a way of not needing them to be in every aisle.

The goal was never a modern warehouse

The goal was a warehouse that tells the truth about what it can ship, then ships that, without inventing a second process after 4pm.

Do the unfashionable work first. Write the sentences. Walk the floor. Then sit with vendors. You will ask better questions. They will not be able to sell you a narrator. You will be shopping for a system, which is the only purchase that was ever worth the meeting.