Lichtenburg Nut Company Business Plan — Quality, Food Safety and Aflatoxin Control

Aflatoxin testing and rejection protocols, HACCP and the quality regime a peanut processor must run.

Section 15 of 37

Quality, Food Safety and Aflatoxin Control

Jump to section

This section is longer than a plant of this size would normally warrant. That is deliberate. The company’s entire commercial premise, its price premium and its exit value rest on a capability that can be destroyed by a single failure, and an investor is entitled to see exactly how that capability is built.

14.1 The regulatory standard

Regulation R.1145, promulgated under the Foodstuffs, Cosmetics and Disinfectants Act, governs tolerance for fungus-produced toxins in foodstuffs and sets a ceiling of 10 parts per billion for aflatoxin in peanut butter and similar products. Enforcement is active: the National Consumer Commission has pursued manufacturers, imposed an administrative penalty and published recall notices naming affected retailers and products.

14.2 The company's control architecture

Table 35. Aflatoxin control architecture from field to despatch

Control point	What is done	Action limit	Consequence of failure
Grower selection and agronomy	Field officer support on harvest timing, windrow drying and storage; contracted growers only	Advisory	Higher incoming contamination rates; more intake rejection
Intake screening	Probe sampling of every delivery; moisture meter and rapid lateral-flow aflatoxin screen before offloading	Reject above 8% moisture or 15 ppb screen	Contaminated lot enters storage and can cross-contaminate
Drying and storage	Aerated silos; moisture held below 7.5%; temperature monitored	7.5% moisture	Aflatoxin develops in store, which is where much contamination originates
Colour sorting	Optical sorter with infrared channel removes discoloured, shrivelled and damaged kernels	Reject stream monitored; target below 4% of intake	The single most effective physical reduction step is lost
Blanching and visual inspection	Skin removal exposes defects; manual inspection station	Visual defect specification	Defective kernels reach the roaster
Post-sort laboratory test	In-house HPLC on a composite sample of every sorted lot	Internal action limit 6 ppb against a 10 ppb legal limit	A non-conforming lot is roasted and milled, contaminating a full batch
Finished goods positive release	HPLC and ELISA on every production batch; no despatch without a release certificate	6 ppb internal limit	Product reaches the market; recall
Confirmatory external testing	Accredited external laboratory on a defined sampling frequency and on any borderline result	Corroborates in-house results	In-house method drift goes undetected
Traceability and mock recall	WMS lot identity from grower delivery note to pallet; quarterly mock recall exercise	Full trace within four hours	Recall scope cannot be bounded, forcing a wider and more costly withdrawal

14.3 Cost of the quality system

Table 36. The cost of the quality system

Element	Type	Amount	Recovered through
Laboratory and aflatoxin testing suite	Capital	R3.6m	Price premium on private label and institutional supply
Colour sorter within the sorting and milling line	Capital	Included in the R24.5m roasting and milling line	Reduced downstream rejection and recall avoidance
Quality systems, certification and external laboratory	Operating	R1,650.0m a year by Year 5	Retailer approval status; a prerequisite for listings
In-process quality testing and consumables	Operating	R0.55/kg of output	Included in variable cost
Raw material rejection and downgrade	Operating	R1.8m a year at Year 5 volume	Partly recovered through sale of downgraded lots at 45% of value
Product recall insurance	Operating	Included within the annual insurance provision	Limits but does not eliminate the financial consequence of a failure

Total identifiable annual quality cost at Year 5		Approximately R1,653.5m	773.2% of net revenue

The honest limit of what the system can promise

No control architecture reduces the probability of an aflatoxin exceedance to zero. Sampling is statistical, contamination is heterogeneous within a lot, and the February 2026 recall occurred at a producer whose own routine internal testing detected the problem, which is to say, the system worked and a recall still happened. What the company can promise is a materially lower probability, faster detection, and a bounded recall scope through lot-level traceability. Investors should assume that a recall event will occur at some point over a ten-year horizon and should satisfy themselves that the company’s insurance and balance sheet could absorb one. On the base case balance sheet, an event on the scale estimated in Section 4.2 would consume roughly a full year of Year 5 EBITDA.