Case study: a home-made feed whose quality changes from one batch to the next, find the cause

This lesson is part of Making feed on the farm, expert level specialisation. It opens to every registered student, from free registration. Levels 1 to 5 and the tracks remain readable by everyone.

An entrepreneur from the Agnibilékrou region has been making his own feed for a year, three houses of 2,000 broilers staggered, a one-tonne-a-day workshop, two workers. He calls in November: "One flock in two is good. The other, the droppings are soft, the weight drags, the feed conversion climbs. Same formula, same maize, same everything. I don't understand." He is lucky to have followed lesson 8: his five registers exist and his retained samples are in their boxes. This lesson goes through the investigation from start to finish: the flock figures, the clues in the registers, the analyses of the retained samples, the reasoning that puts the causes in order, the decision, the correction plan, and what should have been done before. It is the typical exercise of the specialisation exam.

• ✓Read the figures of four flocks and recognise that an alternating problem points to the feed before any other cause
• ✓Cross the manufacturing and dispatch registers with the analyses of the retained samples to isolate the defective batches and what sets them apart
• ✓Rank the causes by their mechanism and their weight, and rule out those that explain nothing
• ✓Write a correction plan with figures that can be verified, and name what a quality system would have caught earlier

1. The situation: four flocks, one workshop, a problem that alternates
2. The clues: what the registers say
3. The reasoning: from clues to causes, by their mechanism
4. The decision and the correction plan
5. What should have been done before
6. Going further: the exam and the specialisation certificate

• •Flocks 1 and 3 good (FCR 1.78 and 1.81), flocks 2 and 4 failed (FCR 1.92 and 1.96, weight − 10%, droppings in waves) under the same conditions: the alternation points to what differs, the feed batches; register 4 + register 2: Tuesday batches conforming, Saturday batches out of range on four columns (protein 18.8%, sodium 0.24%, moisture 14.7%, CV 30%).
• •Mechanisms: soya G at 42.1% not recalculated → − 1.04 point of protein; fish meal H at 3% sodium with salt unchanged → sodium 0.24%; wetted maize "to be used quickly" → moisture 14.7%, bags that smell; mixing 3 to 4 min without pre-dilution → CV 28 to 33%, droppings in waves; scale + 6% → weak effect without alternation; water, season, chicks, disease ruled out.
• •Causes ranked: 1 short mixing (35%), 2 soya not recalculated (30%), 3 salty fish meal (20%), 4 wetted maize (10%), 5 scale (5%); four out of five born on 20 September, when two suppliers came in without a report and the head stopped being there on Saturdays.
• •Plan: bags blocked and retained sample to mycotoxin analysis, formula recalculated and dated the same day (G2-2026-11), job sheet with written time and monthly CV, report before manufacture, wetted lot to "DOUBTFUL", roof and pallets, daily test weight, head or trained second present, quarterly analyses, target FCR ≤ 1.82.
• •What should have been done before: analyse before manufacturing, read the mass balance and the "done" time of the first Saturday, isolate the wetted maize, check the scale every morning, send the retained samples to the laboratory on day 24 of flock 2; a register you do not reread protects against nothing; the exam awards the specialisation certificate issued privately by NEXTEKCI.