Four weeks out from an SAP EWM go-live, someone will put a slide in front of you with a training tracker on it. Percentages, mostly green, one or two ambers with a note about a make-up session. The room relaxes a little. The number looks like readiness. It is not readiness. It is the one thing that was easy to count, so it became the thing everyone counts.
This is about the gap between that green tracker and a floor that can actually run, and the read-out you should ask for instead.
Why the tracker feels like readiness.
A training completion figure is comfortable for three reasons. It is a single number, it trends in the right direction as the go-live nears, and it can be reported without anyone having to stand on the floor and watch a shift. That is exactly why it hides risk. It measures the input, not the outcome.
Attendance is not competence. Sitting through a two hour session on inbound putaway means an operator has seen the screens once, in a quiet room, with a trainer clicking through the happy path. Competence in that classroom is not competence at peak either. Peak is a chock-full staging lane, a truck early, a printer out of labels, and a supervisor asking why the wave has not released. The tracker says one hundred per cent. The floor has never once run that scenario on the new system.
Readiness is proven on the floor at peak, not on a tracker in a slide.
The five dimensions readiness actually has.
Real operational readiness is not one number, it is five separate questions. A green answer on one and red on the others is not eighty per cent ready. It is not ready, because the floor fails at whichever dimension is weakest.
Data
Master data is cleansed and reconciled to physical stock. Storage bins exist in the system and on the floor with the same labels. Handling unit types, packaging specs, and storage type controls match how the goods actually move. If the system thinks bin 04-12-03 holds forty cartons and it holds thirty-one, no amount of training saves the first cycle count.
Process
Every routine and, more importantly, every common exception has a trained, timed recovery path. Short pick, damaged handling unit, failed goods receipt, a wave that will not release: someone on shift knows the exact steps and roughly how long they take. If the answer to any of these is “raise a ticket”, that is not a process, it is a queue.
People
Operators can run their own area unaided, and supervisors can triage. Unaided is the word that matters. Not “with the trainer standing behind them”, not “with the superuser two bays over”. On their own, at pace, for a full shift.
Environment
RF coverage reaches the back of the racking and the dock. Printers are mapped, loaded, and have spare label stock. Chargers work and there are enough of them. The physical bays are labelled to match the data. This is the dimension programmes forget because it does not live in SAP, and it is the one that stops a shift dead at 6am.
Contingency
There is a rehearsed answer to the day something breaks. Not a document called a fallback plan, a rehearsed answer. If EWM is unreachable for two hours, does the floor know how to keep receiving and picking, and how to catch the system up afterwards? Has anyone actually walked it through?
The go-live minus four weeks conversation.
At the four week steering meeting, ask for the read-out in operational evidence, not RAG status. A status someone set to green so the meeting would be shorter tells you nothing. Evidence tells you everything.
Evidence sounds like this: “We ran a full simulated peak shift last Tuesday, three operators short on purpose. Inbound held up. Replenishment fell behind by forty minutes because the wave template was wrong, we have fixed it and we re-run Thursday.” That is a readiness read-out. “Training is at ninety-four per cent and we are on track” is not. One describes the floor. The other describes the tracker.
If nobody can point to a shift the operation has run on the new system, under load, without the project team hovering, you do not have a readiness number. You have a hope.
The move-the-date test.
Here is a clean test of whether the partner in the room is working for your operation or for their milestone. Ask what happens if the floor is not ready at week zero.
An outcome-focused partner will tell you, in writing, that they will recommend moving the date, and they will show you the evidence behind that recommendation. A milestone-focused one will compress readiness quietly, push through the original date, and present the wobble six weeks later as a change-management issue or a user-adoption problem. The date held. The operation did not.
Moving a go-live is expensive, unpopular, and sometimes exactly right. The tell is not whether someone is willing to move it. It is whether they put the recommendation and the evidence on paper before the date, rather than an explanation after it.
A checklist a sponsor can actually use.
For the last four weeks, stop asking for the training percentage. Ask these instead, and expect specific answers, not reassurance.
Has the operation run a full peak-volume shift on the new system, end to end, with the project team hands-off? On what date, and what broke? Can you name three operators who ran their area unaided for a whole shift? Has master data been reconciled to a physical count in the last fortnight, and what was the variance? Is RF coverage confirmed at the dock and the back of racking, not just the office? Are printers mapped and stocked, with spare labels on site? For the five exceptions that happen every day, can a supervisor state the recovery steps without opening a manual? And if EWM is down for two hours on go-live morning, has the floor rehearsed carrying on without it?
Seven questions. If more than one or two come back vague, your readiness is a tracker, not a fact, and you still have four weeks to change that.
Who actually owns readiness.
The trainer does not own readiness: their job was the sessions, and a full tracker means they did it. IT does not own it either: their job was to make the system stand up. Readiness belongs to the operation, the site manager and shift supervisors who have to run a real shift on the Monday.
Treated as the training team's number, readiness gets reported as training completion, because that is the number they own. Put it back where it belongs. Ask the operation whether they can run their floor on the new system at peak, unaided, on day one. If the honest answer is not yet, that is not a training gap. It is the truth, four weeks early, which is exactly when it is still cheap to fix.