White HorseAsk a shift supervisor a simple question: is the plant making money right now, and if not, why? In most operations the honest answer is "give me an hour". That delay is not a lack of data. SCADA has the signals. The historian has the history. The maintenance system has the work orders. Finance has the cost. Each holds a slice of the truth, and none of them talk. That is a visibility problem, and it shows up in recognisable ways. Here are five.
1. Everyone has a different number for yesterday
Production reports arrive from the control room, the plant and the weighbridge, and the three totals do not match. Nobody is lying. Each number was built from a different source, at a different time, with a different definition of "tonne" or "available". When a morning meeting spends its first twenty minutes agreeing what happened, rather than deciding what to do, the data is not the problem. The missing common layer is.
A quick test: pick one KPI from last week and ask three people for it. If you get three answers, you have the first sign.
2. The answer lives in a spreadsheet one person maintains
Every site has one. A workbook that pulls from a historian export, a maintenance report and a manual log, and produces the number the managers actually use. It works because one person understands it. When they go on leave, the number goes quiet or turns wrong without anyone noticing.
This is a sign because the spreadsheet is doing the job a system should do: joining sources that were never joined. It is a workaround that has become infrastructure.
3. Alarms and work orders never meet
An alarm fires on a conveyor drive. Later, a work order is raised in the maintenance system. Weeks afterwards someone asks how often that drive trips and what it costs. The alarm record and the work order are in different systems, with no shared asset identity, so the question takes a manual exercise to answer, or does not get answered.
If you cannot go from an alarm to the work orders raised against that asset, and back, the decision about whether to repair, replace or redesign is being made on a guess.
4. You can see the plant but not the money
Plenty of operations can show throughput, availability and trip counts on a screen. Fewer can show what an hour of a given asset's downtime cost, or what a unit of output contributed, in the same view and at the same time. The cost sits in ERP and finance on a monthly cycle; the operating picture sits in SCADA on a second-by-second one. By the time they meet, the moment to act has passed.
If the operating picture and the cost picture only meet in a month-end report, you are managing on the average, not the moment.
5. Every new question is a new project
Someone asks: which of our pumps spend the most time running outside their normal range before they fail? It is a reasonable question. Answering it takes a data request, an export, a clean-up and an analyst, and it arrives weeks later. The next question starts from zero.
When every question is a project, it is because the data was never organised around assets and decisions. It was organised around the system that produced it.
What to do about it
None of these signs mean you need to replace anything. SCADA, historian, maintenance and ERP are doing their jobs. What is missing is a live, organised layer above them that connects to what is already there, lines the data up by asset, and makes it answerable.
If two or more of these sounded familiar, the useful first step is small and specific: pick one asset and ask whether you could answer, in minutes, whether it is making money right now and why. Our first step with an operation is an assessment that looks at exactly that. If you would like to talk it through, start a conversation at whclips.co.za/contact.html.
White Horse Clips builds the layer a mine's systems never had.
← All notes