Part IX · Operating Well

RCN Procurement: Turning a Kernel Sale into a Maximum Buying Price

Quick answer

Do not buy on a market view. Derive a maximum landed RCN price from kernel you have already sold: revenue, less conversion cost at 1:1 against RCN tonnage, less your target margin, divided by the tonnes you need. Twenty dollars of conversion cost is twenty dollars of buying power, one for one — which is why it is the only one of the five profit drivers you set yourself.

Everything in this chapter happens inside a window of three to four months, and every decision taken in it is irreversible for a year.

Build the buying plan before the season opens

Volume. Annual kernel target ÷ production KOR. A plant selling 700 t of kernel at 23% needs roughly 3,000 t of RCN, plus a margin for the loads you will reject at the gate.

Timing. West African harvest runs roughly February to June; East Africa peaks around October; Cambodia and the southern hemisphere arrive at other points. Prices generally bottom during peak arrivals and rise as origin stocks deplete — but a falling market and a rising one demand opposite behaviour, and the plan has to say which you are assuming.

Price bands. Not a market view. A maximum landed price derived from kernel you have already sold.

Quality thresholds. A minimum cutting-test KOR, a maximum intake moisture around 9% w/w, a nut-count band matched to your calibration classes, and a foreign-matter limit. Trade convention — a convention, not a standards-body figure — treats under 40 lbs/80 kg as typically rejected, 40–46 as marginal and requiring negotiation, 47–50 as a standard purchase, and 50–54 as premium.

Worked example: converting a kernel sale into a maximum RCN price

This is the calculation that keeps an origin processor alive, and it runs in the opposite direction to how most people buy.

You sell 100 t of kernel forward at a blended realised price of US$6,810/t. Revenue: US$681,000.

At 23% production KOR you need 100 ÷ 0.23 = 435 t of RCN.

Conversion cost enters at 1:1 against RCN tonnage, so the whole band carries through and the answer is a band rather than a point. On the manual/African basis of US$559–579/t:

Low: 435 × 559 = US$243,165 High: 435 × 579 = US$251,865

Target net margin at 6% of revenue — the midpoint of a 4–8% net range — is US$40,860.

Low conversion: 681,000 − 243,165 − 40,860 = US$396,975 ÷ 435 t = US$913/t landed High conversion: 681,000 − 251,865 − 40,860 = US$388,275 ÷ 435 t = US$893/t landed

Your maximum landed RCN price is US$893–913 per tonne.

Note what that band is made of. Twenty dollars of conversion cost is twenty dollars of buying power, one for one. That is why conversion cost sits among the five metrics that predict profit — it is the only one of them you set yourself.

Then compare, in both directions

Run the answer against two reference points before the season opens.

Against origin farmgate plus assembly. Farmgate price, plus transport, field logistics and bagging, gives your realistic landed cost from local buying. The headroom between that and your maximum price is your actual trading room — and in a normal year it is thin enough that a single bad conversion-cost month eats it.

Against import parity. Compare with what the same nuts fetch delivered into Asia. The spread between origin farmgate and origin RCN landed in Vietnam is substantial, and that gap is the whole reason to process at origin.

It is also the whole reason an origin processor who buys at import parity dies. Your business is capturing part of that spread, not paying it. A buyer who chases volume up to the import-parity price has converted a processing business into a trading loss with a factory attached.

Suppliers and agents are credit decisions

You will buy through some combination of farmer groups, village aggregators, licensed buying agents and traders. Each is a credit decision as much as a supply decision, because advances are normal — and an advanced agent who sells your nuts to someone else has taken an unsecured loan from you.

Run a scorecard per supplier, updated every load:

Contracted KOR against delivered cutting test
Moisture at the gate
Foreign matter
Weight variance
Lateness
Defaults

Two seasons of that data is worth more than any relationship. Origin-side default is documented rather than theoretical — traders have reported raw-material shipments arriving late or in smaller volumes than agreed.

Write the terms down: pricing basis tied to the cutting test, the moisture deduction formula, your right to reject at the gate, retention against outturn where you buy on a stated KOR, and what happens to an advance if delivery falls short.

At the gate

The cutting test is the contract, so it has to be run the same way every time: a documented sampling rule, a fixed sample mass, the same method, and the result recorded against the lot before the truck is tipped.

Reject at the gate or accept with a stated deduction — but never accept quietly and argue later. Once a load is in the stack it has lost its identity, and with it any claim you might have had.

Log every load into the same record the mass balance is built from. The gate log and the monthly KOR reconciliation are the same dataset seen at two ends, and a plant that keeps them separately will find they disagree by more than the margin.

Covered in detail elsewhere on this site

This chapter assumes the ground these pages already cover, and does not repeat it.

Frequently asked questions

How do I decide the maximum price to pay for raw cashew nuts?

Work backwards from kernel you have already sold. Take the revenue, subtract conversion cost applied at 1:1 against RCN tonnage, subtract your target net margin, and divide by the tonnes of RCN needed at your production KOR. That gives a maximum landed price — a discipline, not a forecast.

Why does conversion cost matter so much to buying power?

Because it enters the calculation one-for-one against RCN tonnage. Twenty dollars per tonne of conversion cost is twenty dollars per tonne of buying power surrendered. It is also the only one of the five profit-predicting metrics that you set yourself rather than receive from the market.

What quality thresholds should a buying plan set?

A minimum cutting-test KOR, a maximum intake moisture around 9%, a nut-count band matched to your calibration classes, and a foreign-matter limit. Trade convention treats under 40 lbs per 80 kg as typically rejected, 40–46 as marginal, 47–50 as a standard purchase and 50–54 as premium.

How should I manage buying agents and aggregators?

As credit decisions, not just supply decisions — advances are normal, and an advanced agent who sells your nuts elsewhere has taken an unsecured loan from you. Run a scorecard per supplier updated every load: contracted KOR against delivered cutting test, moisture at the gate, foreign matter, weight variance, lateness and defaults.