Library · Payment orchestration · How to run this
Cascade latency can abandon the sale
Slow cascade hops can abandon before auth returns. Type hops, latency ms, abandon %, and AOV $. Lost $ = hops x (abandon % / 100) x AOV.
Pair: Cascade latency cost on Orchestration. Fee path: retry cost waterfall. Hop count: failover auth count.
What you type
Hops - cascade hop / attempt count in the latency window.
Latency ms - average hop latency (for the cut you measured).
Abandon % - share of those hops that abandoned before a response.
AOV $ - average order value on that path.
Lost $ = hops x abandon rate x AOV. Shorten the cascade or kill slow MIDs when lost $ is material - do not invent an abandon chip.
What not to do
Do not invent an abandon percent. Do not fold attempt fees into this latency hole. Do not treat hop count alone as lost $. Do not average every BIN when one route owns the latency.
Related: Orchestration, retry cost waterfall, failover auth count, cascade pollution.
Procedure
- Type hops, latency ms, and abandon %.
- Type AOV $.
- Shorten the cascade when lost $ is material.
Failure modes
- Inventing an abandon percent.
- Folding attempt fees into this latency hole.
- Treating hop count alone as lost $.
Worked example
None on this page yet. When one lands it will be a real case: what broke, the number, and what changed. Not a composite.
Glossary
- Cascade latency lost $
- Hops × (abandon % ÷ 100) × AOV $ on slow cascade hops.
Work with us — We run this job on your accounts. Services
FAQ
- Is this retry cost waterfall?
- Related Solidgate job. [Retry cost waterfall](/manuals/solidgate-retry-cost-waterfall) is attempt fees vs net approvals. This page is hop latency abandon $.
- Is this failover auth count?
- Related hop count. [Failover auth count](/manuals/failover-auth-count) is auths per recovered order. This page is latency-driven abandon on those hops.
- What abandon % should I invent?
- None. Type abandon % observed on slow cascade hops.