There was a period, not that long ago in the arc of industrial history, when the problem I described previously did not even have a name. And without a name, it was very hard to solve.
Picture the manufacturing floor of the mid-1980s. ERP systems — or their ancestors — were planning production. They were good at managing materials, purchase orders, and financial transactions. The floor was executing production. But between plan and execution was a void: no real-time tracking, no active management of what was actually happening moment to moment, no system that could tell you at 10 am whether your 2 pm delivery commitment was going to hold. Supervisors managed by walking. Quality problems were discovered at final inspection, or worse, after shipment. Schedule adherence was measured in hindsight. The gap between what was planned and what was happening was real, significant, and entirely opaque.
I was one of the originators of what became the Manufacturing Execution System. The work was to do something specific: define that gap precisely, give it structure, and articulate what a system designed to fill it would need to do. My first book — Applying Manufacturing Execution Systems — was the vocabulary document. It gave manufacturers and vendors a shared language for something that had existed as a problem but had never been named as a solvable category.
That act of naming mattered more than it might seem. Vocabulary precedes investment. Once there was a name for the gap, there could be a market for solutions to it. Vendors could build products they could describe and sell. Manufacturers could write capital appropriation requests for systems they could now define. Integrators could build practices around implementations they could now scope. The Manufacturing Execution System became a global software category, and that category genuinely changed what plants could do.
Plants gained real-time visibility into production status. They gained production tracking that didn’t depend on a supervisor’s whiteboard count. They gained quality data captured at the point of occurrence rather than discovered at final inspection. They gained electronic work instructions, genealogy and traceability, labor tracking, and — critically — a live view of whether the plan and the floor were aligned. The gap between plan and execution narrowed in a way that had real economic consequences: faster response times, better yield, tighter delivery performance. That is a genuine accomplishment, and I want to say clearly that I am proud of the part I played in it.
But honesty requires the second part of that sentence.
MES was designed to execute within a defined process. It was not designed to learn from that process, to connect it backward to design intent, or forward to field performance. It was not designed to act autonomously when conditions deviated from the plan. A second book — Collaborative Manufacturing: Using Real-Time Information to Support the Supply Chain — addressed the next visible gap: the intelligence generated on the floor stayed inside the four walls. Suppliers, customers, and design teams operating upstream and downstream remained disconnected from real-time production reality. The collaboration gap was real, and naming it created space for solutions to it too.
But there is a third gap — one that neither MES nor collaborative manufacturing fully addresses. It is not an execution gap. It is a continuity gap. The failure to keep intelligence connected and active across the entire product lifecycle — from design through production through field service and back — across systems, across organizational boundaries, and across the divide between what machines record and what experienced people know. That is a different kind of problem. It requires a different kind of architecture. And it has a different name — which I will introduce next week.
| Engagement question: “For those who implemented MES in the early years: what did it solve that you hadn’t been able to solve before? And what did it not solve that you expected it to?” |
Michael McClellan www.productiondigitalthread.com