Downtime tracking software records when a machine or line stopped, for how long and why, and uses those records to calculate OEE (overall equipment effectiveness). The short answer: OEE is availability multiplied by performance multiplied by quality. Availability captures stops, performance captures slow running and minor stops, and quality captures scrap. The real value of the software is not the percentage on the screen but the list of reasons behind the lost time, ranked by duration.
Why do shift logbooks and spreadsheets fail at downtime tracking?
In many plants stops are written into a logbook or a spreadsheet at the end of the shift, from memory. That approach has three weak points. Stops of a few minutes never get written down, although added up over a day they can outweigh a single big breakdown. Reasons are free text, so "die", "die setup" and "die change" become three separate rows that no report can add together. And the recorded minutes rarely add up to the length of the shift; the gap is loss that nobody owns.
A stop without a reason code is just bad news. If the records do not add up to the shift length, the OEE you calculate from them is an estimate, not a measurement.
How is OEE calculated?
OEE is the product of three ratios, and each one exposes a different kind of loss. The calculation starts from planned production time rather than the whole shift: time when the line was never meant to run, such as breaks or periods without orders, is taken out first.
- Planned production time: shift length minus scheduled stops (breaks, no orders, scheduled maintenance).
- Availability: run time divided by planned production time. Run time is what remains after breakdowns, setups and waiting for material are subtracted.
- Performance: total count multiplied by the ideal cycle time, divided by run time. It shows how much slower than intended the machine produced while it was running.
- Quality: good count divided by total count.
- OEE: availability times performance times quality.
A worked example: in a 480 minute shift with a 30 minute break, planned production time is 450 minutes. With a 40 minute breakdown and a 20 minute changeover, run time is 390 minutes and availability is 86.7 percent. If 700 pieces came off a product with a 30 second ideal cycle time, performance is 89.7 percent; if 14 of them were scrap, quality is 98 percent. Multiplied together that is an OEE of roughly 76 percent. The numbers only illustrate the arithmetic; the same logic in code looks like this:
type Stop = { reason: string; minutes: number };
type Shift = {
shiftMinutes: number; // total shift length
stops: Stop[];
idealCycleSec: number; // ideal seconds per piece for the product that ran
totalCount: number;
scrapCount: number;
};
// Which reasons fall outside planned production is a company rule kept in a table.
const SCHEDULED = new Set(['break', 'no-order', 'planned-maintenance']);
function oee(s: Shift) {
const minutes = (scheduled: boolean) =>
s.stops
.filter((x) => SCHEDULED.has(x.reason) === scheduled)
.reduce((sum, x) => sum + x.minutes, 0);
const plannedTime = s.shiftMinutes - minutes(true);
const runTime = plannedTime - minutes(false); // breakdowns, changeovers, waiting for material
if (runTime <= 0 || s.totalCount <= 0) return null; // nothing to measure
const availability = runTime / plannedTime;
const performance = (s.totalCount * s.idealCycleSec) / (runTime * 60);
const quality = (s.totalCount - s.scrapCount) / s.totalCount;
return {
availability,
performance,
quality,
oee: availability * Math.min(performance, 1) * quality,
// Faster than "ideal" means the ideal cycle time is wrong: flag it, do not hide it.
idealCycleSuspect: performance > 1,
};
}Two details matter here. First, which reasons count as scheduled is a company rule stored in a table; the operator cannot decide it at the machine. A changeover may be planned, but it is still an availability loss. Move it outside planned time and OEE goes up while one of the most reducible losses disappears from view. Second, if performance comes out above 100 percent the line has not broken a record; the ideal cycle time is wrong. The code flags that case instead of hiding it.
How should downtime tracking software record a stop?
- Every stop is its own record: machine, start and end time, reason code, the work order running at the time and the shift.
- A record without an end time is an open stop. It closes when the machine runs again, and a machine can never have two open stops at once.
- The reason is picked from a two level list (group first, then reason). Free text is only an optional note.
- Minor stops below a threshold you define are not questioned one by one; they are counted and booked as performance loss.
- A stop with no reason has to be completed before the shift is closed. The share of the "other" code is tracked separately.
- Records are never deleted. A corrected duration or reason is stored together with who changed it and when.
Should the data come from the operator or from the machine?
There are three options. With fully manual entry the operator taps a large "stopped" button on a tablet and picks a reason; it is the cheapest setup, but it misses short stops. In the semi automatic approach a running signal or a part counter from the machine captures the start and end of each stop, and the operator only picks the reason. Fully automatic setups take the reason from machine alarms as well, but not every machine provides them. For most plants the second option is the balanced starting point: the machine says when it stopped, a person says why. How the signal is collected depends on the age of the machine and its controller, so hardware should be treated as a separate budget line. The ideal cycle time and the product currently running should come from the work order, which we covered in our production tracking software (MRP) article.
Which reports turn downtime data into loss analysis?
- Pareto by reason: stop reasons ranked by total duration. The improvement meeting starts at the top of this list.
- OEE trend per machine: how the same machine changes week over week.
- Performance per product: which products never reach their ideal cycle time.
- Shift comparison: the difference between shifts on the same machine and the same product.
- Minor stop count: the minutes look small, but the repetition shows where the problem sits.
These reports do not live alone. A stop caused by a breakdown should be linked to a maintenance work order; that flow is described in our maintenance management software article. Scrap should not be counted in two places either: it should come from the quality control software record. And if management wants to see these figures next to other indicators, the data can feed a business intelligence dashboard.
Common mistakes in downtime and OEE tracking
- Turning OEE into a league table across machines or plants. When the product mix and shift pattern differ, the comparison misleads; the meaningful benchmark is the history of the same machine.
- Tying OEE directly to bonuses. The records degrade: stops drift into scheduled codes and scrap gets reported late.
- Starting with dozens of reason codes on day one. Operators pick the first one in the list. Start with a handful and refine them based on the data.
- Entering the ideal cycle time once and forgetting it. It needs a review whenever the product, tooling or material changes.
- Looking only at the percentage. The percentage says how much was lost; the reason list says where.
Packaged MES or custom software?
Packaged MES and OEE products ship with the standard calculations and screens, and they deliver quickly in plants with modern, similar machines. Custom development earns its place when work orders have to come from your existing ERP, when the machine park mixes ages and brands, when the reason tree and reports are specific to you, or when stop records must share identifiers with your maintenance and quality systems. Whichever route you take, starting with a single bottleneck machine is the cheapest way to test the reason codes and the operator screen on the shop floor.
Conclusion
Good downtime tracking software records every stop with a start, an end and a reason code, takes the scheduled versus unscheduled split from a table, calculates OEE while showing the three ratios separately, and ranks losses by reason. The percentage alone changes nothing; the time spent on the reason at the top of the list does. To build stop recording into your work order, maintenance and quality processes, explore our corporate solutions page or describe your needs and request a quote.