How do I calculate the Guaranteed Analysis in my pet treats?

All pet food labels require a guaranteed analysis on the label to advise the purchaser of the product’s nutrient content. At minimum, guarantees are required for minimum percentages of crude protein and crude fat, and maximum percentages of crude fiber and moisture. The only exception is for products that do not and are not intended to provide protein, fat or fiber (for example, vitamin and mineral supplements), in which case the product is exempt from guarantees for those components. In all cases, though, a moisture guarantee is required. Guarantees for other nutrients are normally voluntary, although additional guarantees may be required to support claims made on the label. For example, claims such as “with calcium” or “high in vitamin E” would require minimum calcium or vitamin E guarantees, respectively.

Unless the product is formulated through the use of sophisticated computer software with a complete and accurate database on nutrient content of all ingredients used in the product, the best means of determining appropriate guarantees is by laboratory analysis. A “proximate analysis” of the finished product includes testing for the four nutrient components mentioned above in addition to ash. Results are typically reported in percentages “as fed” (AF). If unsure as to the units reported, it is prudent to ask the laboratory for clarification.

Important to note is that a single analysis of a product may not provide sufficient data to determine reliable guarantees. Guarantees are declared as either “minimums” or “maximums,” meaning that if a feed control official obtained a sample of the product and tested it in his or her official state lab, the crude protein and crude fat must be NO LESS THAN the stated percentage on the label, while the crude fiber and moisture must be NO MORE THAN the label declaration. Each batch of finished product is going to vary in composition due to variation in ingredient composition, mixing rates, and amount of water driven off during the cooking process. Unfortunately, a single analytical value does not give any indication as to the expected batch-to-batch variation. As a result, the composition of the batch that was tested may be very different from the one the feed control official analyzes!

Failure to meet the guarantees may result in possible enforcement action against the product. It is prudent, then, to test multiple batches of each product. How many? In general, the more analyses conducted, the better one can estimate batch-to-batch variation and hence more reliably set guarantees. Understanding economic limitations of start-up companies, it is prudent to conduct at least two analyses of each variety to start. As more batches are produced and analyzed, more data may indicate the need to adjust the guarantees on the next label printing.

In the following example, four batches of product are tested and the results tabulated. The best means of determining appropriate guarantees from the data is by statistical analysis, calculating the average and standard deviation for each nutrient (these calculations can be done easily in MS Excel). Assuming the variation will have a “normal distribution” (i.e., it follows a “bell curve”), the average plus or minus two standard deviations will encompass approximately 95% of the population. In other words, based on this estimate of variation, by setting the minimums below the low end and maximums above the high end, 95% of the time the product will meet the guarantee.

Alternatively (and far less complicated), finding the lowest value among the crude protein and crude fat results, and then further rounding down from that value to the next whole percentage can be done to set the minimum guarantees. For crude fiber and moisture, the same can be done, except in those cases, the highest value is rounded up.

In the following example, if the manufacturer relied solely on data from the first batch to set guarantees, there likely would have been problems down the road. The first batch measured 25% crude protein. However, one of the subsequent batches was much lower at 23%. If that later batch happened to be analyzed by a state, the product would have been found deficient. The average crude protein was found to be 24.8%. Subtracting two standard deviations from the average is 22.25%. Thus, a prudent minimum crude protein guarantee for this product is 22%.

Example: Calculating guarantees from proximate analysis data

 Batch  Crude Protein
% AF
Crude Fat
% AF
Crude Fiber
% AF
% AF
% AF
 1  25  15  4.5  7.8  3.2
 2  26  14.3  3.2  8.9  3
 3  25.2 13  4  8.1  2.9
 4  23  16  3.8  9.1  3.3
 Average  24.8  14.58  3.88  8.48  3.10
Standard Deviation 1.28 1.26 0.54 0.62 0.18
Average +/- 2SD 22.25 12.05 4.95 9.72 3.47
Min/Max 23 13 4.5 9.1 3.3
Recommended Guarantee 22 (min) 12 (min) 5 (max) 10 (max) 3.5 (max)

NOTE: Ash guarantee not required on pet food labeling, but is typically included in “proximate analysis” lab reports