Industry 4.0: drop the slogan, decide what you will measure
Most digital transformation projects stall long before the technology choice. They stall at deciding which data is worth collecting.
Most of what gets written under the Industry 4.0 banner is true and useless. Machines talking to each other, data flowing in real time, decisions becoming automated. All correct. None of it tells you what to do on Monday morning.
Start with what you cannot see
When something goes wrong on a production line or in a warehouse, the first question is usually when it happened. If nobody can answer, you have found data that is not being collected.
The examples are concrete. Which shift changed the die. Which rack did the pallet end up on after it was received. How many hours has the machine run since its last calibration. In most facilities the answers exist somewhere, but on paper, in a spreadsheet, or in one person head. Not in a system.
Automatic identification is the cheapest step
The most expensive way to collect data is to have a person type it. It is slow, it is error-prone, and it is the first task dropped on a busy day. RFID, barcodes and sensors are the first rung of the ladder for that reason: they move the record off the person and into the process.
If a pallet passing through a doorway creates its own record, the accuracy of that record no longer depends on how busy the shift was. That is the whole point.
Data without integration, integration without a report
The most common place these projects stall is the connection to ERP or WMS. An RFID dashboard sitting on its own screen, which nobody opens, gets abandoned within months.
The reverse fails too. Pipe the data into ERP and let it land in a table nobody reads, and you end up in the same place. Write down in advance who makes which decision from this data. Which screen does the warehouse manager open in the morning, and who gets notified when a threshold is crossed.
Start small, but start with a baseline
Two criteria are enough when picking a pilot: the result can be stated as a number, and it finishes within six weeks. If it cannot be stated as a number, nobody can defend it later. If it takes six months, the organisation priorities will have moved by the time it lands.
Measure the current state before you begin. How long a count takes, what the stock discrepancy percentage is, how many assets are unaccounted for. Without those numbers all you can say at the end is that things improved, and that does not win the next budget.
The technology decision comes last
Which frequency, which tag, which reader. Good questions, and the last ones to ask. Get clear first on which decision needs which data. After that the technology choice tends to make itself.