Value chain / Nine-stage production and measurement corridor
Follow the thread across nine stages.
Cameroon harvests over three hundred thousand tonnes of seed cotton a year, yet exports about ninety-five percent of its ginned fibre raw. Value stays only where every physical conversion carries a measurement.
Eight physical conversion stages and one transverse data layer sit between field and garment. Scrub the chain: each stage changes the denominator, its loss mode and its record.
Stage 01 of 09 · zone A
Field
≈340,000 t
seed cotton harvested in 2024, after 314,455 t in the 2023 study year
Warp · mass in, mass out
Seed, rain and soil → unginned seed cotton (t)
Sourced via SODECOTON (more than 220,000 growers; about US$100m a year of rural cash).
Warp · primary loss mode
Moisture, trash and mixed grades picked into one lot
An engineering description of where the stage leaks, shown without a figure.
Weft · measurement signal
Producer-group lot ID, harvest date, moisture and village grade.
Stage 02 of 09 · zone A
Ginning
≈95%
of ginned lint exported raw, against about 5% processed locally
statutory monthly wage floor (SMIG) from February 2024, the labour-cost anchor for the sewing line
Warp · mass in, mass out
Cut bundles + trims → assembled garments
Core CMT scope (about 900 staff at full ramp, estimate); supervisor pipeline.
Warp · primary loss mode
Rework, rejects and line stoppages
An engineering description of where the stage leaks, shown without a figure.
Weft · measurement signal
Hourly output vs target, OEE (Availability x Performance x Quality), FPY.
Stage 08 of 09 · zone C
Finishing and dispatch
CEMAC
domestic and regional corridors; US AGOA is excluded from the base case
Warp · mass in, mass out
Sewn garments → inspected, packed orders
Own-brand domestic retail and eCommerce, institutional and CEMAC wholesale.
Warp · primary loss mode
Late or short orders, and failed final inspection
An engineering description of where the stage leaks, shown without a figure.
Weft · measurement signal
Orders On Time In Full (OTIF), final AQL pass rate, landed route cost.
Stage 09 of 09 · zone D
The data layer
6 KPI families
fed from shift one; Tier-2 forecasting waits for 12 months or more of clean history
Warp · mass in, mass out
Floor events → validated operational and ML records
Manual PMS spine in Phase 0 and 1; edge vision in Phase 1; ML forecast in Phase 2.
Warp · primary loss mode
Gaps and untrusted records that hide every other loss
An engineering description of where the stage leaks, shown without a figure.
Weft · measurement signal
Edge buffer uptime, schema completeness, forecast MAPE vs history.
Basis · ≈340,000 t Unginned seed cotton entering SODECOTON’s gins: 330,000 t in 2022, 314,455 t in 2023, ≈340,000 t in 2024, above 350,000 t projected for 2025.
Basis · ≈95% Ginned lint leaving Stage 02 exported raw against about 5% processed locally: the ratio applies to lint, not to seed-cotton weight.
Basis · 50% by 2030 SND30 local-processing target for the same lint by 2030: a policy ambition that needs spinning capacity, supervisors and market absorption.
Evidence · Stage 01 314,455 t (2023); ≈340,000 t (2024); >350,000 t (2025P) (R01-R04)
Evidence · Stage 02 ≈95% raw export vs ≈5% local; 50% target by 2030 (R03, R06, R07)
Evidence · Stage 03 Camtext SA US$300m / 180bn XAF; 12k + 3k jobs (R08-R10)
Evidence · Stage 04 60 m/min CV vs 60-70% manual; AI market US$523.4m (R28, R29)
Cameroon has no shortage of raw cotton. The structural loss happens at the post-ginning boundary, where pressed lint bales board export vessels instead of entering domestic spinning.
95t of every 100 t of lint exit raw
5t stay for Stages 03 to 08
Schematic flow, not a facility. Heights are shares of 100 t of ginned lint: the 95 / 5 baseline is programme-annex data , the 50 percent figure is the SND30 policy target , and positions between them are illustrative arithmetic. Seed cotton and cottonseed come from SODECOTON's nine gins and two oil mills .
The ratio applies to lint leaving Stage 02, not to seed-cotton weight. Positions between 5% and 50% are illustrative arithmetic.
Denominator ruler · the thread cut into the four quantities this page counts
010203040506070809
Stage 01 counts unginned seed cotton by weight: 330,000 t in 2022, 314,455 t in 2023 and about 340,000 t in 2024 across more than 220,000 growers. Multiplying it by garment prices would overstate value.
Stage 02 counts ginned lint. The 95 percent raw-export and 5 percent local shares apply to this lint, and so does the 50 percent SND30 target for 2030; the slider above moves along this denominator only.
Stage 03 counts yarn. The Camtext SA integrated pole targets about 12,000 t a year (US$300 million, 180 billion XAF), with about 12,000 jobs at Dibamba and 3,000 at Garoua.
Stages 06 to 08 count garments: an illustrative nameplate of 3.0m garments a year for a focused cut-make-trim and own-brand venture, staffed by about 900 workers at full ramp. Planning estimates, not operating results.
Source note: why the four denominators differ
An industrial value-chain model fails if it multiplies raw field harvest weight directly by garment selling prices. The 330,000 tonnes harvested in 2022, 314,455 tonnes in 2023 and roughly 340,000 tonnes in 2024 measure unginned seed cotton entering SODECOTON’s nine ginning plants . Ginning separates the seed, which feeds two cottonseed oil mills, from spinnable lint fibre; an illustrative 40 to 42 percent ginning outturn is used in the Cotton Value-Flow Explorer tool. The 95 / 5 split and the 50 percent target apply to that lint.
Four zones, four steps. Each step lights one zone of the thread, draws its machinery, swaps its figure and hangs its card beneath it.
Zone A of D · Origin custodyZone B of D · Quality at the loomZone C of D · Cloth into brand valueZone D of D · The data layer
Zone A · 01 Field · 02 Ginning
≈340,000 t
seed cotton harvested in 2024
2022330,000 t
2023314,455 t
2024≈340,000 t
2025P>350,000 t
Zone B · 03 Spinning · 04 Fabric
>90%
vision defect detection at up to 60 m/min
By eye60 to 70%
Visionabove 90%
Zone C · 05 Wet · 06 Cut · 07 Sew · 08 Dispatch
XAF 43,969
statutory monthly wage floor (SMIG) from February 2024
Cameroon SMIGabout US$76
Cambodia garment minimumUS$204, benchmark
Zone D · 09 Data layer
6 KPI families
captured from shift one
Hourly production
First-pass quality
OTIF delivery
Labour cost
People capability
Water and energy per unit
Clean history toward Tier-2 forecastingthreshold: 12 months or more
Zone A · Stages 01 and 02
Field agronomy and saw ginning establish origin custody.
Northern Cameroon’s cotton belt supports more than 220,000 growers and nine SODECOTON ginning plants. Capturing producer-group lot identifiers at village collection and stamping HVI fibre parameters on every pressed lint bale anchors traceability before a single thread is spun.
Partnered spinning and vision inspection protect quality early.
Spinning turns ginned lint into yarn cones; weaving and knitting form greige cloth. Manual inspection is 60 to 70 percent accurate and misses 20 to 30 percent of defects; Tier-1 computer vision inspects fabric at up to 60 metres a minute with above 90 percent accuracy across more than 40 defect types.
Wet processing, cutting, sewing and dispatch convert cloth into brand value.
Stage 05 meters water, dye and energy per unit. Stage 06 nests cutting markers around mapped fabric defects. Stage 07 runs cut-make-trim lines on SMIG-compliant productivity bands, the statutory floor being XAF 43,969 a month from February 2024, and Stage 08 fulfils domestic and CEMAC orders.
The data layer turns shift logs into a cold-start forecasting foundation.
A greenfield factory begins with zero sales history. Stage 09 captures hourly output, downtime causes, first-pass yield and resource meters at the edge, feeding six KPI families from shift one while it accumulates the clean history that Tier-2 demand forecasting needs.
Without an explicit boundary per stage, losses hide between departments. Hover, focus or press a stage, or let a row cross the reading line, and the read-out follows.
Showing 3 columns: stage, loss mode, signalShowing 2 columns: stage, material in and outShowing 2 columns: stage, programme scopeShowing 2 columns: stage, evidence anchorShowing all 6 columns
Nine-stage value-chain architecture: material denominators, loss modes, programme scope, measurement contracts and evidence anchors.
Stage
Material in and out
Primary loss mode
Programme scope
Measurement signal
Evidence anchor
Seed, rain and soil → unginned seed cotton (t)
Moisture, trash and mixed grades picked into one lot
Sourced via SODECOTON (more than 220,000 growers; about US$100m a year of rural cash).
Producer-group lot ID, harvest date, moisture and village grade.
314,455 t (2023); ≈340,000 t (2024); >350,000 t (2025P) (R01-R04)
Own-brand domestic retail and eCommerce, institutional and CEMAC wholesale.
Orders On Time In Full (OTIF), final AQL pass rate, landed route cost.
Cameroon US$94.70/person (R26); AGOA closed since 2020 (R18)
Floor events → validated operational and ML records
Gaps and untrusted records that hide every other loss
Manual PMS spine in Phase 0 and 1; edge vision in Phase 1; ML forecast in Phase 2.
Edge buffer uptime, schema completeness, forecast MAPE vs history.
McKinsey AI inventory -35 to -40%, error -20 to -50% (R27)
Source note: how the ledger reads
Loss modes are engineering descriptions of where each stage typically leaks, shown without a figure. Evidence anchors are programme-annex references; sector AI benchmarks (R27, R29) are literature context rather than achieved local results, and the 3.0m-garment nameplate and 900-person workforce are planning estimates. Without an explicit boundary per stage, factory losses hide between departments: ginning trash becomes spinning ends-down, unmapped weaving slubs become sewing rejects, and power trips masquerade as operator slowdowns.
09 THE DATA LAYER → 01 FIELDThe transverse record feeds back into the first warp thread of the interlude loom ↓
Interlude · the idea in one line
Production is the warp that makes the garment.Production is the warp that makes the garment.Measurement is the weft that makes it viable.Measurement is the weft that makes it viable.
Lit: the eight conversion rows, 01 Field to 08 Dispatch.Lit: row 09, the data layer that crosses every stage.Lit: every crossing, the nine points where a measurement is taken.
Architecture A · full vertical build-out: heavy upstream capital, decade-scale payback
Basis · Capex Camtext SA integrated project (180 billion XAF) for about 12,000 t a year.
Basis · Payback Return on investment, with optimal capacity around 2033.
Basis · Restructuring Estimated restructuring of the incumbent CICAM; one relaunch plan cites 70.2bn.
Dial positions 1 and 2 are illustrative mid-points between the two architectures.
Proposed architecture B · focused CMT, partnered upstream: contained Phase 1 capital
Basis · Nameplate Garments a year: an illustrative nameplate (planning estimate, not an operating result).
Basis · Workforce Workers at full ramp (estimate).
Basis · Candidate supply channels Illustrative planning assumption: SODECOTON, the SODECOTON-CICAM pole and Camtext SA are potential raw-material, yarn or fabric sources, subject to diligence, capacity and agreement. These sources do not establish GFT supplier agreements.
Dial positions 1 and 2 are illustrative mid-points between the two architectures.
0 of 9
Built from scratch upstream
3 of 9
Partnered or metered upstream
4 of 9
Owned by GFT in Phase 1
Counts are read from the strands: arithmetic on the roles, not a cost.
1Demand anchored at home and in CEMAC, not on AGOA
2Supervisor-training pipeline funded
3Sustainability backed by metered data (05 and 09)
Phase 1 owns cut, make, trim, finishing and the brand. Spinning and fabric are partnered upstream steps and wet processing is metered, so first-phase capital stays contained. Pick a figure to read its basis.
Step 2 of 4 · dial towards A
Building upstream turns three stages into capital.
Dial strip B towards A and Stages 03, 04 and 05 flip to built from scratch. The panel now reads architecture A: heavy industrial capex, base-load power and a decade-long payback.
Step 3 of 4 · three conditions
The verdict is a conditional go on three conditions.
Revenue anchored on domestic Cameroon and CEMAC demand rather than closed US AGOA access, a funded supervisor-training pipeline, and a sustainability position substantiated with metered data.
Step 4 of 4 · the programme tests each
Three deliverables hang from the strand.
Each piece of the programme answers a condition: the feasibility study tests all three, the performance system instruments Stages 06 to 09, and the digital strategy starts where technology is proven and data history is thin.
The value chain is not switched on all at once. Each phase has a condition precedent that must be met before capital is committed to the next.
Gates passed 0123 of 3
Which stages each phase lights is a reading of the phase text; gate and gauge positions are ordinal, not measured scores.
Phase 0 · Set up · Opens at the start
Lock the siting decision, SODECOTON offtake and the manual data spine.
Read the condition
Resolve the Douala-Dibamba corridor versus northern Garoua cotton-belt location trade-off, secure traceable SODECOTON cotton and partnered spinning terms, recruit the seed supervisor cohort, and establish the manual performance-management tracking sheets.
Phase 1 · Launch and stabilise · Gate: condition precedent for Phase 1 capital
Run the first CMT lines with daily visual control and Tier-1 vision inspection.
Read the condition
Commission cutting, sewing and finishing lines toward the illustrative 3.0m-garment ramp. Operate the six KPI families daily and deploy Tier-1 line-scan computer vision (above 90 percent accuracy across more than 40 defect types at up to 60 m/min) on incoming fabric rolls.
Phase 2 · Scale · Gate: 12 months or more of clean SKU-level records
Add sewing lines, widen into CEMAC and activate Tier-2 demand forecasting.
Read the condition
Once 12 or more months of clean SKU-level sales and production records accumulate, introduce machine-learning demand forecasting and inventory optimisation (where sector literature reports 20 to 50 percent forecast-error reductions and 35 to 40 percent inventory-cost reductions at scale).
Phase 3 · Optimise · Gate: national pole capacity matures
Integrate upstream pole partnerships and physics-constrained digital design.
Read the condition
Deepen local yarn and fabric integration as national pole capacity matures toward 2030 to 2033, and adopt multi-model, physics-constrained design-to-specification workflows across weave, shade and silhouette.
1234 of 4
Physical capacity
Siting, offtake and a recruited supervisor cohortFirst CMT lines toward the illustrative 3.0m-garment rampMore sewing lines and CEMAC regional ordersUpstream yarn and fabric integration toward 2030 to 2033
1234 of 4
Digital maturity
Manual performance-management tracking sheetsSix KPI families daily; Tier-1 line-scan visionTier-2 forecasting on 12 months or more of clean recordsPhysics-constrained design-to-specification workflows
Basis of preparation: programme-annex figures, estimates labelled, not investment, legal or tax advice.
Figures on this page are drawn from GFT’s Cameroon textile programme annex (prepared 31 January 2024; source refreshes flagged 2026) . Planning inputs such as the 3.0 million garment annual capacity and 900-person workforce are explicitly labelled illustrative estimates, and sector AI benchmarks (R27, R29) are literature context rather than achieved local results. Intermediate positions on the lint-split instrument are illustrative arithmetic, and the tool sparklines sketch each tool's response shape only. Nothing on this site constitutes investment, legal or tax advice.