Latency, in payments technology, is the time that passes between a system sending a request and receiving the answer, and the subject is covered in full under payment processing latency. It is measured in milliseconds. When a shopper taps a card, the delay before the terminal shows an approval is the sum of the latency of every system the request crosses: terminal, gateway, acquirer, card network and issuer.
The general term comes from computing and telecommunications, where it describes any such delay. One component is network latency, the time data spends travelling between two points. The rest is processing time inside each system, including fraud checks, database lookups and calls to other services. Payment companies usually report latency as percentiles, not averages, because a small share of very slow responses causes most of the harm.
Latency matters because payment messages are subject to time limits. A payment authorization that takes too long is timed out and may be declined or answered on the issuer's behalf, and instant payment schemes set a maximum time for each transfer. Providers therefore commit to response times in a service level agreement, alongside uptime.