Every raw pet food brand on earth answers one engineering question, whether it says so or not. What kills the Salmonella? There are exactly four answers in commercial use. Knowing them turns a wall of premium packaging into a one question sorting exercise.
Answer one: nothing
The most common answer, in Thailand and worldwide, is nothing at all. Careful sourcing, a clean facility, freezing, and hope. Freezing stops bacteria multiplying but kills none of them. The FDA's survey of 196 commercial raw pet foods found Salmonella or Listeria in 38, close to one in five, and every product was sold frozen. The US recall record shows how this ends: advisories and warning letters against untreated raw brands year after year, including product that reached homes. Good sourcing lowers the odds. It cannot make them zero.
Answer two: test and hold
A step up. Hold each batch in the freezer, send samples to a laboratory, release only the negatives. Serious brands run this honestly, and it catches real contamination. Its weakness is statistical. A test describes the sample, not the batch. Contamination is not spread evenly through a production run, so a clean 25 gram sample can sit beside a colonised one. Brands running test and hold programmes have still ended up in recalls. Testing without treatment reduces risk. It does not remove the mechanism.
Answer three: high pressure processing
HPP is the established kill step of the American raw category. Sealed packs are submerged in water and squeezed at pressures up to about 6,000 bar, which ruptures bacterial cells while leaving food structure largely intact. It is cold, validated and effective. Controlled studies show Salmonella reductions from 1 to 9 logs depending on pressure and time. The best US brands run every batch through it and publish results by lot number. Its constraints are physical: flexible packaging that tolerates compression, treatment on chilled rather than frozen product, and machinery that concentrates in large plants. Where it is available, it is a legitimate answer.
Answer four: electron beam
The newest answer in pet food, and the one built for frozen product. The sealed, already frozen pack rides a conveyor through a beam of accelerated electrons for a few seconds. No radioactive source, no heat, no water, no pressure. The arithmetic is regulated: 1 kGy removes one log of Salmonella, published poultry work shows total bacteria below detection at 3 kGy, and kill steps run at 5 to 10 kGy. Nutrition holds at those doses. Batch testing at an ISO/IEC 17025 accredited Bangkok laboratory measured taurine, vitamin A and peroxide value effectively unchanged in a Thai raw cat recipe before and after treatment, with only thiamine showing a loss of about 17 percent, the kind of loss formulation buffers exist to absorb. Frozen in, frozen out, treated through the final pack.
How to judge any answer
The method matters less than the verification. Pau & Sun, the team behind a world championship cattery, apply the same three tests to any brand regardless of technology. Name the kill step. State the validated log reduction. Show the per batch laboratory certificate tied to the lot number printed on the pack.
HPP with lot level certificates is a good answer. E-beam with lot level certificates is a good answer. The only wrong answers are the first two on this list, and they are the answers most of the market is quietly giving.
