Which Enterprise? The Five Gates

#Seventeen tonnes imported for every tonne grown

In 2023, South Africa's whole freshwater aquaculture sector — 140 farms across nine provinces — produced 451.62 tonnes of fish, down 70.73% on the year before. That same year the country imported 7,879.6 tonnes of tilapia at a landed value of about R12.39/kg (DFFE Aquaculture Yearbook 2024). Imports of one species beat total domestic freshwater production roughly seventeen times over. Of the farmers the department managed to survey for that year, 52% reported income below R100,000, and not one reported above R1 million.

Read those numbers the way the previous lesson taught you to. The department's words are "out of the total number of farmers surveyed", and it calls its freshwater data collection voluntary and incomplete. The 70% drop may be farms closing or farms not returning a questionnaire. This describes who answered, not a proven ceiling — and it still points hard in one direction.

This lesson is about the choosing, not the husbandry. Three questions settle most enterprises before an animal arrives: who buys this, what does my site do in July, and may I lawfully sell what it makes. All three can be answered at a desk in an afternoon, for nothing.

#Five gates, in an order that is not negotiable

The aquaculture research produced a decision sequence sharp enough to generalise. This course runs it on every animal in every module, from cattle to bees to fish. The gates are the same; only the answers change.

#Gate 1 — three names, in writing, before any capital moves

The rule from the aquaculture work is blunt: if you cannot name three buyers, stop. Not three market segments. Not three types of customer. Three names, with phone numbers, who have told you what they will take and at what price.

Then the harder half: get it in writing before capital moves. A written offtake is not a contract you will litigate. It tests whether the demand is real. People who happily say "ja, I'd definitely buy that" go quiet when asked to put a monthly volume on paper. That silence is Gate 1 doing its job, six months and a great deal of capital early.

Selling direct to the public raises this gate rather than skipping it. Four things must be true before a rand goes on branding: a registered abattoir near enough to take small lots, an unbroken cold chain, correct labelling, and committed demand rather than surveyed interest. Fail one and the answer is no.

Be careful with the story that grass-fed or free-range earns a premium. South African carcass classification pays for youth and fat, not for how the animal was raised. In the week ended 17 July 2026, A2/3 beef traded at R69.15/kg and C2/3 — an older animal, exactly the slow grass-finished profile — at R60.73/kg (RPO). That is R8.42/kg against you, about R2,000 on a 240 kg carcass. The premium exists only off the commodity rail, and the research behind this course found no published South African figure for what a customer will pay extra for grass-fed beef. If someone claims that premium, ask to see the invoice.

#Gate 2 — the constraint that does not negotiate

Gate 2 is where your farm file from the previous lesson earns its keep. Here is the worked refusal.

Delta Farms, at Meyerton on the Highveld, is considering tilapia in a dam. The national water quality guideline gives Mozambique tilapia a growth optimum of 28–30 °C and a lower lethal range of 9–11 °C, and states flatly that "the costs involved in controlling temperature under pond-culture conditions are prohibitive and not feasible" (DWAF Volume 6, 1996). Now put the site against it, on a 30-year record computed independently rather than lifted off a weather website.

Gate 2 closes, before the first load of clay is moved. Honesty travels with that. Those are air temperatures used as a proxy: no measured Highveld pond or dam water series was available to the research, and real water is damped — a deep dam runs warmer than air on a winter night, a metre-deep earth pond barely at all. The direction is sound; the exact months are not. The fix is cheap, and it is the best-value action in this module: log your dam at 06h00 weekly for a winter and bring data to the decision.

Gate 2 is not only about fish. On veld it is grazing capacity, and it refuses just as firmly. In Gauteng it runs 4 to 20 hectares per Large Stock Unit, mean 6.8 (DAFF grazing capacity norms) — a 4 ha/LSU polygon and a 20 ha/LSU polygon are both "Gauteng", so use your own, never the mean. Take 100 hectares at 6.8 ha/LSU: 14.7 LSU. A cow with her calf plus her share of the bull is costed at 1.6 LSU in the Land Bank's beef budget (Land Bank 2023/24), so it carries about nine breeding cows. Nine cows do not support a household, and no enthusiasm converts them into eighteen.

#Gate 3 — may you lawfully sell what this animal makes?

This is the gate South African beginners skip most often, and the one that turns an enterprise into an offence.

Apply Gate 3 as a sentence, for any animal: can I lawfully sell what this animal produces, and through exactly what route? Rabbits show why the route half matters. Rabbit sits in Schedule 1, so s7 applies in full: household eating is lawful, selling the meat is not, unless it went through a registered abattoir. The research behind this course could not establish that a registered rabbit abattoir exists anywhere in South Africa. Until someone shows you an open route with a telephone number on the end of it, a rabbit meat business is a plan to break the law.

The instrument changes with the animal — for fish, the invasive-species listing under the National Environmental Management: Biodiversity Act 10 of 2004 plus a provincial nature-conservation permit; for game, a certificate of adequate enclosure under the Game Theft Act 105 of 1991; for bees, annual registration with the department. Find yours before you buy the animal.

#Gate 4 — inputs, and what happens when they move

The same test on a Land Bank 500-bird broiler batch returns five cents of gross margin per rand of cost, with feed at about 69% of that cost, so a 2% move in the feed price erases roughly 30% of the margin (Land Bank 2023/24). That is not a business with a buffer. It is a feed-price arbitrage.

The same budget puts winter lick at 55% of total variable cost in an extensive beef enterprise, with bought feed about three-quarters of the whole. Hold onto that, because it sets up most of this course: every practice that carries grazeable forage into July attacks the largest line on the income statement. Those are 2023/24 prices — take the structure, not the numbers.

Gate 4 has a second half here: power. If a four-hour outage kills your entire stock, you do not have a business, you have a bet against Eskom.

#Gate 5 — the trough, and the drought

Can you fund the deepest point of the cash curve, and then survive the drought that arrives during it? Livestock and fish are long-cycle: nothing returns until first sale, and the trough sits where the winter feed bill is still running and no income has landed. Do not fund a multi-year animal with twelve-month money, and assume you are self-insuring the drought — the drought reserve is a financing line, not a virtue.

#The scheme test

#Now write the shortlist

Most learners stop at Gate 1 or Gate 2 on most of their candidates. That is the sequence working, not failing. A gate is cheap; a pond, a fence, a flock and a permit application are not. Go and find your three names, and go and read your own winter temperature.