Finishing
99.75 percent is 2.6 log, not 3. Each log is another tenfold kill, and the last one is the expensive one.
— log
The scale the specification is written on
AATCC 100 and ISO 20743 differ in inoculum, contact time and neutralisation, so a log reduction from one is not directly comparable with the other and neither transfers to JIS L 1902 without care; quote the method with the number. The zero-contact count must come from the treated specimen at time zero rather than from an untreated control, because a finish that inhibits growth without killing will otherwise be credited with a kill it did not perform - and a specimen where the organism simply failed to survive the contact period on the untreated fabric produces a large apparent reduction from a test that proved nothing. Log reduction is a ratio and carries no information about absolute risk: three log from a large inoculum leaves more organisms than two log from a small one. The dose model treats efficacy as a threshold, which is a simplification of a dose-response curve that is steep but not vertical; near the threshold small dose differences move the log reduction substantially, and that region is exactly where a marginal formulation sits. Nothing here addresses regulatory approval of the active, which in most markets is a separate and harder gate than efficacy, or the question of whether an antimicrobial finish delivers any hygiene benefit in use as opposed to on a specimen.
Antimicrobial Log Reduction & Durable Dose — free, with the formula and a worked example, at Textile School.