Why field-level diffs matter more than match rates

Most OTC reconciliation vendors lead with a headline match rate. 94%. 97%. 99.2%. The number is comforting, and it is real. It also hides the thing that actually costs an ops desk its day: the shape and quality of the remaining breaks.

A 97% match rate on 2,000 daily trades leaves 60 breaks. If each one takes an analyst 20 minutes of screenshot-stitching and email exchange before they can even begin to investigate the actual disagreement, that's 20 hours of work, just in identification and triage, before anyone has worked a single break on its merits.

What the match rate doesn't tell you

The match rate is a summary. It collapses a lot of operational reality:

• Reason mix. 60 rate/spread breaks are very different from 60 leg-direction breaks.
• Aging distribution. 60 breaks aged under 30 minutes is a healthy queue. 60 breaks aged 4+ hours is an SLA problem forming in front of you.
• Counterparty concentration. 50 of 60 breaks being with one counterparty is a process conversation, not 50 independent tickets.
• Recurrence. The same field breaking with the same counterparty every morning is a booking convention issue, not a matching issue.

A match rate tells you none of this.

What ops teams actually need

An analyst working breaks doesn't need a percentage. They need answers to a small set of questions, for each break, fast:

1. What actually disagrees? Not "there is a break" — which field, with which values on each side, and is that field identity-bearing?
2. How old is it? Where is it in the SLA band?
3. Who else is looking at it? Has the counterparty opened it yet, attached a note, or proposed a resolution?
4. What changed? Is this an amendment, a novation, a late attestation, or a fresh disagreement?

None of these are served by a headline number. They are served by field-level diffs, aging buckets, and a shared operating surface.

Match rate as health, not as product

Match rate is a fine KPI for management reporting and week-on-week health tracking. It is a poor primary UX for the team actually working the queue. Design the tool for the analyst's day. Report match rate. Don't lead with it.