Treasury Operations · control & efficiency analyser

Four-Eyes Ledger

Enter daily volumes and the minutes each maker, verifier, authoriser and releaser spends per transaction, plus the EOD checklist. The ledger returns a control-adjusted efficiency score, the control-to-processing ratio, the cost of each control layer per catch, and ranked recommendations.

Transaction types

Type Vol / day Makermin Verifiermin Verifierreject % Authorisermin Authoriserreject % Releasermin Releaserreject % Rework% of vol Reworkmin each
Set a step to 0 minutes if it doesn't exist for that type (for example maker = 0 where the Bloomberg–FLEXCUBE integration creates the deal). Reject % = share of items the layer sends back.

Team and end-of-day

Minutes per day by role

Where you sit on the control curve

Control layer economics

Hours per catch = layer hours per year ÷ items it rejects per year

Recommendations

How the numbers are calculated

Total time = Σ volume × (maker + verifier + authoriser + releaser minutes) + Σ volume × rework % × rework minutes + EOD minutes × people.

Required time = Σ volume × (maker + the first control layer present) + the part of EOD the system cannot answer. The first control layer is the four-eyes minimum and is never treated as waste.

Efficiency = required ÷ total. 100% means every minute is either doing the work or the one independent check on it. What sits between your score and 100% is additional control layers, rework, and manual EOD checks the system could answer.

Control ratio = all control minutes ÷ maker minutes, shown as x : 1. Where deals are system-created (maker = 0) the ratio is shown per human minute of input across the whole book.

Layer verdicts: a layer that catches fewer than one item a month, and is not the first check, is a redundant candidate; a layer rejecting under 1% is review — tier by risk; otherwise it is earning its place. These follow the inverted-U finding (JIAR 2021) that efficiency rises with control up to a point and falls beyond it; the point is where a layer's cost exceeds the loss it prevents.

Savings are hours per year using your working-days setting. They assume a merged or tiered layer keeps 30% of its time for high-risk items, a sequencing fix removes 80% of rework, and automating a system-answerable EOD question removes its full share of the checklist.