Collector Productivity Metrics That Actually Move Recoveries

Every collections operation eventually ends up with the same conversation: the team is working hard, the call volume is up, and the recoveries haven't moved. The instinct is to push for more calls. The honest answer is that the metrics being tracked aren't the ones that predict recoveries.

This guide is written for the operations manager or head of collections at a South African collection agency (and the same signals apply inside a law firm recoveries team). It sets out what to measure, why, and how to instrument it without adding another spreadsheet.

Why call volume is a lagging (and misleading) KPI

Call volume looks like productivity because it's easy to count. It rewards activity regardless of outcome and quietly encourages the wrong behaviour — short calls with the wrong people, no follow-through, and no honest capture of what actually happened on the line. A busy day and a productive day look identical in a call-count report.

The metric only becomes useful when paired with outcome signals. On its own, it tells you the team is tired, not that the book is moving.

The four productivity signals that predict recoveries

1. Right-Party Contacts (RPC) per hour. The share of dial time spent actually speaking to the debtor (or a decision-maker), not voicemail, wrong numbers or gatekeepers. RPC is the leading indicator: it moves before recoveries do.
2. Promise-to-Pay (PTP) conversion. Of the RPC conversations, what proportion end with a dated commitment? A high RPC rate with low PTP conversion is a coaching signal, not a data problem.
3. Kept-PTP rate. Of the PTPs made, what proportion actually pay on the committed date? This is the single best predictor of recovery run-rate, and the one most operations under-track because it needs the promise, the payment and the cross-reference to live in the same system.
4. Cycle time to first payment. From file allocation to first Rand received. Shorter cycle time compounds — earlier payments free up collector capacity for the next file, and the book turns faster overall.

Instrumenting the metrics without extra manual work

The reason most operations don't already track these signals is that the data lives in three different places — the dialler, the workflow tool and the payments system — and stitching them together weekly is a job nobody has. The fix isn't a new report; it's making sure the promise, the outcome and the payment all attach to the same account object.

• RPC comes from the dialler when call outcomes are captured against the account in a consistent set of dispositions.
• PTP conversion needs promises captured as structured objects (date, amount, channel), not free-text notes.
• Kept-PTP needs the payment reconciliation to attach against the promise, not just the account.
• Cycle time only works if allocation date and first-payment date are both recorded reliably.

This is why spreadsheet-based collections operations struggle with these signals: the four events are captured, but they never join up. Once the account object holds all four, the metrics become a query, not a monthly project.

Team-level vs. portfolio-level views

The same four signals need to be visible at two levels, for two different audiences.

• Collector level — used by the team lead for weekly coaching. Individual RPC per hour, PTP conversion and Kept-PTP rate, benchmarked against the team median rather than an absolute target.
• Portfolio level — used by the operations manager and executive for book-level decisions. Same four signals, aggregated by client, portfolio segment or product, so the pattern is visible before it becomes a recovery miss.

Turning metrics into a weekly cadence

A metric only helps if a conversation follows it. The cadence that gets the most out of these four signals is a 30-minute weekly coaching slot per collector, structured around one question: "which of the four signals moved most this week, and why?" Not a leaderboard, not a rank — a coaching conversation with data on the table.

At portfolio level, the same signals feed a 20-minute weekly operations review with the executive. Anything trending against target for two weeks in a row becomes an action, not a note.

What "good" looks like

Absolute benchmarks vary by portfolio type, book age and channel mix, so treat the following as directional ranges rather than targets:

• RPC rate: healthy books sit meaningfully above the average of "one RPC per handful of dials"; the exact ratio depends on data quality more than collector skill.
• PTP conversion: the coaching signal, not the recruitment one — improvements come from script and objection-handling work.
• Kept-PTP rate: the metric most operations discover is lower than they assumed once it's actually tracked. Improvement here moves recoveries directly.
• Cycle time: watch the trend, not the absolute number — a shortening cycle is the strongest signal that the operation is compounding.

Common failure modes

1. Turning the metrics into a leaderboard. Ranking collectors publicly on RPC or PTP conversion produces gaming, not improvement. The metrics belong in coaching, not on a wall.
2. Reporting monthly instead of weekly. Monthly reporting is too slow to coach against — the collector has already banked the habit by the time the report lands.
3. Missing the Kept-PTP link. Tracking PTPs made without tracking PTPs kept tells you nothing about recoveries. The two must be reconciled at the promise level.
4. Ignoring client-mix effects. A drop in RPC can be a portfolio issue, not a team issue. Segment the view before you draw a coaching conclusion.

The point of productivity metrics isn't to measure the team harder. It's to make the four moments that actually predict recoveries — the right conversation, the committed promise, the kept promise, the first payment — visible enough to coach against.