Digital transformation starts with the process, not the software

Most manufacturing digital programmes begin with a system. An MES. A dashboard. A connectivity project. A set of KPIs that look complete in a slide. The software is selected, the integrator is appointed, and the plant is asked to adapt.

That sequence is backwards.

A factory is not waiting for software. It is already running — with a mix of formal process, informal workarounds, machine constraints, material shortages and the judgement of people who have been doing the job for years. If you do not understand that system first, the new software will only make the existing confusion more visible, or worse, more official.

What the process actually is

The documented process is rarely the real one. The real process includes the extra inspection that started after a customer complaint, the buffer that exists because changeover takes longer than planning assumes, and the operator who knows which machine drifts after lunch.

If you ask only managers, you will get the intended process. If you stand on the floor for a few shifts, you will see the one that produces the goods.

Digital transformation has to start there: production sequence, capacity, quality loops, WIP, machine behaviour and the decisions people make when the plan no longer fits. Until those are clear, there is nothing useful to digitize.

Software cannot compensate for an unclear problem

I have watched plants collect thousands of machine signals and still be unable to answer a simple question: why did this order miss the shift target?

The data was not missing. The problem was never named. Availability, performance and quality were being measured because those are the standard measures, not because someone had decided which decision the numbers were supposed to change.

A useful digital project begins with a narrower question. Where does time disappear between two operations? Which stoppages are planned and which are not? When the plan changes mid-shift, who sees it, and how long does that take? Those questions tell you what to measure. The software comes after.

The shop floor already has information

Operators, supervisors and planners are not waiting to be informed by a new system. They already have information — some of it in their heads, some of it on paper, some of it in a spreadsheet that nobody officially owns.

A good digital approach does not replace that overnight. It makes the useful parts of it shared and timely. The test is simple: after the system is live, can a person make a better decision during the shift than they could before?

If the answer is that management now has a nicer report at the end of the month, the project has not reached the floor.

Start small enough to be wrong cheaply

Large programmes try to model the entire plant on the first attempt. That is usually too early. The first digital slice should be small enough that you can be wrong without embarrassing the organization.

One line. One constraint. One decision that currently waits too long. Once that is working, the next slice is easier because the plant has seen that the work is about their process, not about installing a platform.

This is slower to announce and faster to become useful.

The combination that matters

I started my career as an industrial engineer in a garment factory. Later I spent decades in software. The useful part of that combination is not a slogan. It is a habit: understand the work, then decide what technology can honestly support.

Manufacturing digital transformation fails when it is treated as an IT rollout. It works when it is treated as an operations problem that happens to need software.