01One number hides three owners
When someone says acceptance rate is down, the first question is: which one? A single headline percentage mixes declines that belong to three different owners. The issuer said no. A fraud rule said no before the issuer was ever asked. Or the provider failed somewhere in between.
Each of those has a different fix and a different person who can make it. Report one blended number and all three get the same response, which is usually a ticket to the PSP asking why approvals are low.
02The three rates I track side by side
General acceptance is approved transactions over everything that entered the channel. It is the widest view and the one most dashboards show.
Merchant-side acceptance is the rate as the merchant experiences it. It matters because it is the number the merchant will quote back to you, and it does not always agree with what upstream reports.
Anti-fraud-adjusted acceptance removes transactions stopped by fraud rules, leaving the attempts that actually reached an authorisation decision. This is the fairest measure of how issuers and routes are performing.
03Reading the gaps
The gaps between the three are the diagnosis. A healthy adjusted rate with a weak general rate points at fraud rules declining too much. A weak adjusted rate on one BIN range across every route points at the issuer. A weak rate on one MID while the same BINs perform elsewhere points at the route or the provider.
A gap between upstream and merchant-reported results is its own finding. It usually means the two sides are counting different things, and that has to be settled before anyone argues about performance.
04Roll-ups are where it goes wrong
Averaging percentages across rows is the classic mistake. A BIN-level rate has to be rebuilt from approved and attempted counts, never from the average of its MID-level rates. I found exactly this kind of roll-up error understating BIN-level merchant acceptance in a daily report, and fixing it changed which BINs looked like problems.