Production measures must expose losses, not compress them

Sponsors often evaluate a new garment factory by multiplying installed sewing stations by annual working hours and dividing by an assumed garment work content. That calculation describes theoretical nameplate capacity under uninterrupted steady-state conditions, not saleable throughput.[S11] Production ramp-up research shows that a newly launched process is initially poorly understood, simultaneously causing low production rates and low yields.[S11] In an apparel sewing line, where thirty to forty interdependent stations pass bundles sequentially, a gap between theoretical capacity and saleable output arises at four boundaries before a carton enters the warehouse: calendar time lost to unscheduled shifts, planned time lost to stoppages, running time lost to sub-standard speed or imbalance, and completed pieces lost to rework or rejection.[S06][S07]

Historical evidence from Cameroon's existing textile sector illustrates the risk of treating installed capacity as an undifferentiated block. As shown in Exhibit 21, the inherited sector baseline from July 2023 records wide variation across domestic facilities, with Garoua operating at 33.7 percent of capacity, CICAM 1 at 48.9 percent and CICAM 2 at 10.83 percent.[R12] While those figures are Grade-C sector observations of an established enterprise rather than GFT operating results, a single plant-level utilisation percentage cannot tell an operations director whether missing output resulted from power interruptions, spare-parts shortages, unbalanced work allocation or quality rejects.[R12][S07] Reporting one aggregate utilisation number compresses distinct engineering and commercial problems into an unactionable figure.[S0501]

Exhibit 21 establishes only that wide gaps between installed capacity and actual utilisation have been observed in the national industry and associated with financial loss, reported at about CFAF 2 billion in 2020.[R12] It cannot establish root causes inside each plant, because an aggregate utilisation ratio merges demand deficits, supply delays, breakdowns and efficiency losses into one denominator.[S07] To prevent that ambiguity, GFT separates the time and piece accounting hierarchy into non-overlapping layers aligned with ISO 22400-2.[S01]

Before any efficiency or OEE percentage is calculated, the shift ledger must account for every minute of the working day. In Cameroon, labour regulations establish a standard 40-hour working week and a 48-hour maximum week, with overtime compensated at 1.2 times the regular rate.[R16] Under an illustrative five-day schedule of eight-hour shifts, each line begins with 480 scheduled calendar minutes per shift (2,400 minutes per week).[R16] Unscheduled shift time due to unreleased orders is a capacity-allocation variance owned by commercial planning, not an operating loss owned by the line supervisor.[S22]

Once a shift is scheduled and attended, three subtractions convert calendar time into productive-equivalent time:[S01][S06]

Locking these conservation rules into the shift record prevents the failure identified by NIST's manufacturing diagnostics subcommittee, which warned that multiple competing definitions and calculations of OEE circulate across industry and make unstandardised figures incomparable.[S0501]

Overall Equipment Effectiveness detects production losses and maximises equipment effectiveness by decomposing performance into three multiplicative components: availability, performance efficiency and quality rate.[S06] Applied to a sewing line or cutting room, the composite score equals the product of those three sub-KPIs:[S06]

$$\text{OEE} = \text{Availability} \times \text{Performance} \times \text{Quality}$$

When availability is $\text{Actual Run Time} / \text{Planned Production Time}$, performance is $(\text{Total Pieces Completed} \times \text{SMV}) / \text{Actual Run Time}$, and quality is $\text{First-Pass Good Pieces} / \text{Total Pieces Completed}$, intermediate terms cancel upon multiplication.[S06] The product simplifies to $(\text{First-Pass Good Pieces} \times \text{SMV}) / \text{Planned Production Time}$, the ratio of good standard minutes earned to planned production minutes.[S06][S0501] That cancellation guarantees an exact conservation measure of productive time, but displaying only the cancelled product hides whether time vanished into stops, slow pace or defective stitching.[S06][S07]

Exhibit 22 shows how GFT structures the shift bridge so every loss category remains visible between the opening 480-minute scheduled shift and closing productive-equivalent time.[R16][S06]

In Exhibit 22, starting with 480 scheduled minutes per station and deducting 30 break minutes leaves 450 planned production minutes.[R16][S06] Deducting 45 minutes of illustrative availability losses leaves 405 run minutes ($405 / 450 = 90.0\text{ percent}$ availability).[S06] An illustrative 80.0 percent performance rate deducts 81 minutes of speed loss, leaving 324 standard-earned minutes, and an illustrative 95.0 percent first-pass quality rate deducts 16.2 minutes spent on defective garments, leaving 307.8 minutes of good time ($307.8 / 450 = 68.4\text{ percent}$ OEE).[S06][S10] While this waterfall reconciles shift arithmetic, it cannot by itself identify which station acted as the constraint or whether speed losses came from skill gaps or starvation; those questions require component drill-downs.[S04][S07]

Because multiplication is commutative, a line with 90.0 percent availability, 80.0 percent performance and 95.0 percent quality achieves the same 68.4 percent OEE as a line with 80.0 percent availability, 95.0 percent performance and 90.0 percent quality, yet needs intervention from different teams.[S06][S22] Exhibit 23 illustrates three garment lines in an illustrative plant review that all report a composite OEE near 68 percent (68.4 percent, 68.0 percent and 68.3 percent), with diverging component profiles.[S06][S0501]

In Exhibit 23, Line A runs reliably (90.0 percent availability) with clean work (95.0 percent quality), but operates at an illustrative 80.0 percent performance rate during ramp-up, requiring line balancing and coaching.[S04][S11] Line B runs near standard pace when active (95.0 percent performance) with 95.5 percent quality, but loses 25.0 percent of planned time to stoppages (75.0 percent availability), requiring maintenance, cutting coordination or setup reduction.[S06][S07] Line C maintains 92.0 percent availability and 95.0 percent performance, but loses 21.9 percent of output to defects (78.1 percent quality), pointing to machine tension, pattern faults or speed-biased incentives.[S10][S0503] Exhibit 23 is an illustrative comparison rather than an empirical benchmark and cannot replace event logs, but it shows why GFT never prints a composite OEE figure without its three constituent ratios and denominators.[S01][S06]