Short answer. Production management software holds five things: the master data, items, bills of materials, routings, rates; the production orders; tracking at the station, where quantities, stops and scrap are declared at the moment they happen; material and work in process; steering. The third decides everything: without a declaration taken on the spot, the software contains only what was planned. With it, three answers: how many useful hours in a shift, which causes of stops cost the most, where each order stands. In this page's example, an eight-hour shift gives 4 h 22 of useful production.
This page is for those who run a factory or a manufacturing workshop, plastics, packaging, metalwork, food, materials, textiles, or who lead its production, and for those who own a unit in Algeria and follow it from France. The definitions come from public sources: overall equipment effectiveness, six big losses, downtime, work in process, Pareto principle, material requirements planning, bottleneck, quick changeover, andon. The factory of forty, its minutes and its pieces are a constructed example; the figures are fictitious. The page covers neither payroll nor labour rules, and quotes no hardware price.
The life of a production order
The figure follows an order in six steps.
Release. A production order says what to produce, how many, by when, at which station. In a workshop, it comes from the customer's order, to show that work is about to begin. Material. It leaves stock for that order, with its lot: that is what will say what the order consumed. Set-up. The changeover is timed like everything else: it is often one of the first causes of stops, and the least looked at. Production. It is the accented step: at the end of each hour or each pallet, the quantity is declared; at each stop, a cause is chosen from seven or eight.
Declaring in ten seconds. A tablet at the station, big buttons, no keyboard: quantity, stop, cause. If the machine has a counter, it is read directly and only the cause remains to be entered. Without a network on the shop floor, the station keeps recording, see the offline application. The check. Good pieces, scrap, rework: three figures per order, not one. Closing. Time spent, material consumed, good quantity: the closed order gives its real cost, the one detailed in the cost of a finished product. What this page does not cover. Calculating dates and load is in production planning; traceability of food lots, in the food manufacturing software page.
An eight-hour shift, four hours twenty of production
The figure runs through a shift minute by minute, then regroups the same minutes.
Overall equipment effectiveness. OEE measures how well a piece of equipment is used compared with its full potential, during the periods when it is scheduled to run. It is the product of three factors. Availability: the share of scheduled time in which the equipment is available to operate. Here, 345 minutes running out of 450 scheduled, break deducted: 76.7%. Performance: actual speed as a share of the theoretical maximum speed. 2,760 pieces out for 3,450 possible in 345 minutes: 80%. Quality: good units as a share of total units produced. 2,622 out of 2,760: 95%.
The product. 76.7% × 80% × 95% = 58.3%. In other words: 2,622 good pieces where 4,500 were possible, or 4 h 22 of useful production in an 8 h shift. None of the three alarms on its own. 77, 80 and 95 each read as “not bad”; it is their product that surprises. No ideal figure. You often read that an OEE of 85% is the world-class reference; the very source that reports it warns that this target is wrong in many cases, for example when not all losses are counted. The useful OEE is your own, month after month, on the same machine. What it is not. A mark given to the team: in the example, 70 of the 105 minutes of stops, changeover and waiting for material, are decided upstream of the station.
«None of the three figures alarms on its own; it is their product that surprises.»
The stops, cause by cause
The figure sorts one month of stops by duration.
Downtime. In industry the term is commonly applied to failures of production equipment; it covers unplanned stops, such as a breakdown, and planned ones, such as maintenance. The record at the station notes both, with a cause. Seven or eight causes. A list of thirty causes will not be used; seven or eight, chosen with the workshop, and an “other” box that is watched: if it grows, a cause is missing. The ranking. The Pareto principle says that, for many outcomes, roughly 80% of consequences come from 20% of causes; the source adds that this ratio is a rule of thumb, not a law. In the example, three causes out of seven make 74% of the 3,800 minutes.
What the ranking changes. Before the record, people said “breakdowns”; they are third, at 19%. The first two, changeover and waiting for material, are dealt with elsewhere than in maintenance. The changeover. Quick changeover, known by the acronym SMED, is a method for moving fast from making one product to making the next; before any method, it has to have been measured. Waiting for material. It points to stock and purchasing, see replenishment. Breakdowns. Each breakdown declared at the station can open a work request, followed as in maintenance interventions. Alerting. An andon system notifies supervisors and maintenance of a quality or process problem: a stop that exceeds ten minutes sends a message, without anyone having to phone.
«Before the record, people said “breakdowns”. They are third.»
What production management software holds
The figure stacks five layers; the third feeds all the others.
The master data. The bill of materials says what a product is made of; the routing, which stations it goes through and at what rate. Without a nominal rate per machine and per product, there is no performance to calculate. Material requirements. Material requirements planning, or MRP, is a production planning, scheduling and inventory control system; most are software-based. It is the historic core of production management software: from orders and bills of materials, what must be bought and made, and by when. Work in process. That is the name for partially finished goods waiting for completion. Too much of it between two stations points to a bottleneck.
The bottleneck. A bottleneck is a process whose limited capacity reduces the capacity of the whole chain. An hour lost on the bottleneck is an hour lost for the factory; so that is where tracking starts. Where is the order? Each declaration moves the order forward: released, in set-up, in production, 1,200 pieces out of 3,000, in check. The answer to the customer is read on screen, without going down to the shop floor. The acronyms. Production management software plans and follows the orders; a manufacturing execution system, or MES, tracks and documents the transformation of materials into finished goods; an ERP brings it all together with purchasing, sales and stock. For a mid-sized factory, one piece of software often does all three; what matters is that tracking at the station exists.
Six big losses, two per factor
The figure sorts the losses by OEE factor.
Availability. Waiting and breakdowns: the machine is stopped. These are the losses people see, hence the ones they talk about. Performance. Minor stops and reduced speed: the machine runs, but puts out less than its rate. In the example, 69 minutes over the shift, without a single stop having been noted. Only the comparison between pieces out and the nominal rate makes them appear. Quality. Scrap and rework: pieces made for nothing, or made twice.
Where to start. With the longest loss on the bottleneck machine, and one at a time. One month of records is enough to choose it; more is needed to know whether the action worked. By shift, without ranking. The same station read by shift shows gaps; they come as often from the product released or the material received as from the people. Causes are compared, not shifts. In the morning. Ten minutes in front of the screen: yesterday's OEE per machine, the three longest stops, what is decided. The same figures go up to management in the dashboard.
Eight weeks, in the example
Eight weeks for a workshop of a few machines, one of them a survey.
- What management receives: production orders followed from release to closing; quantities, stops and scrap declared at the station; OEE per machine, per shift and per product; stops sorted by cause; the progress of each order; the real cost per order.
- What the factory no longer does: learn yesterday's output the next day on a sheet, argue about causes without figures, go down to the shop floor to answer a customer, discover a delay on delivery day.
- What is measured: availability, performance, quality and OEE; minutes stopped per cause; duration of changeovers; work in process between stations; orders delivered on the date; gap between planned and actual time.
- What Symloop does: the survey on the shop floor, the master data, the production orders, the declaration screens at the station, offline included, reading counters where possible, the shop-floor boards, then the first morning review with you.
The survey. One week on the shop floor: your machines, your routings, the production sheets already being filled in, and the causes of stops as the workshop names them. The bottleneck machine first. Tracking starts on one or two machines, those that limit the factory. The tablets. One per machine, fixed at the station, with the current order on screen. The morning review. The last milestone: the first daily meeting held in front of the screen.
What remains afterwards. An OEE per machine and per shift, stops sorted by cause, orders whose progress is known, a real cost per order. Off the shelf or custom. Production management and shop-floor tracking software exists, and some of it is very complete; it fits when your processes fall within its frame and your teams can absorb its complexity. Custom work is justified when tracking at the station must be very simple, tied to what you already have, or when the process has its particularities. The durations are those of a constructed example.
- Is there production management software for a small or mid-sized factory?
- Yes. It must hold master data, production orders, tracking at the station, material and steering. The point to check: are quantities and stops declared at the station, when they happen, or re-entered the next day?
- What is production management software?
- The software that plans what must be made and bought, releases the production orders and follows their progress. Its historic core is the calculation of material requirements.
- How do you follow production in real time?
- With a declaration at the station, on a tablet: quantity produced, stop, cause. If the machine has a counter, it is read directly. The progress of each order and the hour's output can then be read on screen.
- How do you calculate OEE?
- Availability times performance times quality. Or, more simply: good pieces produced, divided by the pieces possible during the scheduled time at the nominal rate. In the example, 2,622 out of 4,500, that is 58.3%.
- What is a good OEE?
- There is no figure that holds for everyone: the 85% target often quoted is disputed by the very source that reports it. The right reference is your own OEE, on the same machine, from one month to the next.
- Why do my machines produce less than planned?
- Three families of causes: the machine is stopped, waiting, breakdowns, changeovers; it runs below its rate, minor stops, reduced speed; or it makes pieces to throw away or redo. One month of records at the station says which weigh most.
- How do you reduce machine stops?
- By recording them first, with a cause, for a month. Sorting by duration points to two or three causes; in the example, the first ones are not breakdowns but changeovers and waiting for material.
- How do you know where an order is on the shop floor?
- Each declaration moves the production order forward: released, in set-up, in production with the quantity done, in check, closed. The answer to the customer is read on screen.
- Do the machines have to be connected?
- Not to begin with. A well-made manual declaration is enough for quantities and causes. Reading a counter or a running signal comes next, where it saves an entry; it depends on each machine.
- How long does it take to equip a workshop?
- Eight weeks in this page's example, for a few machines, starting with the bottleneck machine, up to a first morning review held in front of the screen.