A new nutritionist takes over the mill's broiler starter formula. It has a minimum of 3 percent for fish meal. Nobody quite remembers why. A customer may have asked for it once, or it may have been a former consultant's preference. When the nutritionist looks at the shadow price report, this one constraint turns out to add about 140 TRY a tonne to the formula. On monthly production of 1,500 tonnes that is 210,000 TRY a month. The numbers are illustrative, but the situation is one we meet very often in the field.
Constraints of habit
Every formula collects constraints over time. A problem comes up and a maximum is set for an ingredient. A customer asks for something and a minimum is added. Years go by, the problem is solved, the customer is gone and the ingredient's quality has changed, but the constraint stays in the formula. Each of these constraints may look small on its own. Together they can pull the formula well away from least cost.
The easiest way to find them is to look at shadow prices. Every binding constraint, one the solution runs up against, has a shadow price. It shows how much cheaper the formula would become if the constraint were relaxed by one unit. Constraints with high shadow prices are the most expensive decisions in the formula, and the reason each one is there needs to be known.
That does not mean every expensive constraint should go. Some, for pellet quality, animal health or legal limits for instance, are necessary. But a constraint whose reason nobody knows should either be justified or removed. Reviewing the constraints of every formula together with their shadow prices once a year is a good habit for bringing these silent costs to light.
Mixing up as-fed and dry matter
Another common mistake in the field is entering ingredient values and animal requirements on different bases. A lab result comes in on a dry matter basis and is written into the matrix as it is, when it should be entered as fed. Or a ruminant requirement is defined on dry matter while the ingredient values are as fed, and no conversion is made.
Take an example, with illustrative numbers. The lab result for barley at 88 percent dry matter shows 12.5 percent crude protein on a dry matter basis. If that value goes into the matrix without being converted to as fed, the barley's protein shows as 12.5 percent instead of the real 11 percent. With 40 percent barley in the ration, the ration's crude protein looks 0.6 points higher on paper than it really is. A mistake like this is often noticed only when animal performance drops.
The way to prevent such mistakes is to record clearly in the matrix the basis of every analysis value and to let the system, not a person, do the conversions. Treating a value that differs unexpectedly from the previous analysis as a warning is a good habit too.
Unit mix-ups are a similar source of error. Energy is held in kcal in one place and MJ in another. A mineral comes in as a percentage in one analysis and as mg/kg in the next. One ingredient's price is entered per tonne and another's per kilogram. Some of these errors are noticed at once because they lead to absurd results. The more dangerous ones are those that shift the result a little and do not catch the eye.
Digestible instead of total
Formulating amino acids on total values is an outdated approach for many species today, but it lives on in some formulas. Total amino acids do not see differences in digestibility between ingredients. A by-product with low digestibility can seem to compete with soybean meal on total lysine while supplying the animal with far less lysine. Moving to digestible amino acids reveals the real value of these ingredients and often improves both performance and cost.
When no solution is found
Sometimes no formula can be found that meets all the constraints at once. The most common reason is constraints that contradict each other. A minimum set for one ingredient, for example, can force another nutrient above its maximum. In that case, seeing which constraints clash is the first step to solving the problem.
But not every infeasible result is a real contradiction. Sometimes constraints are defined in a very narrow range, for example a minimum and maximum for a nutrient at almost the same value, and small rounding differences in the calculation make the limit look broken. In that case the constraints should be widened to a meaningful range or the program's tolerance setting reviewed. Forcing a formula that looks infeasible by relaxing constraints at random can create new and better hidden problems.
Records and reasons
All of these mistakes share one root: the reasons for decisions are not recorded. When a constraint is set, an analysis value changed or an ingredient taken out of the matrix, leaving the reason as a note keeps the next nutritionist from having to ask the same questions all over again. Formula version history shows which change was made when. Notes explain why.
In OptiFormula, shadow price and parametric analysis reports show each formula's most expensive constraints and ingredient price ranges. Formula version history and the cost archive make it possible to follow past decisions and their results. With these tools, a formula review done once a year turns up a much larger saving than expected in most mills.
The most expensive mistakes in formulation are usually the ones that do not catch the eye. The formula is solved, feed is produced, animals grow and everything seems fine. The only way to find the silent costs is to question, regularly, the constraints and values nobody has questioned for years.
