Does agent matching work when collector turnover is high?

Why turnover matters to matching

Predictive pairing, BestPair’s term for agent matching, works from the floor’s own results: which collectors have collected on which kinds of accounts. Experienced collectors carry most of that record. When one leaves, what the floor learned about them leaves too, and the accounts they did best on go to colleagues with less of a track record.

New starters bring the opposite problem. Their first weeks rest on too few contacts to say much.

Turnover affects how much history each collector has, and it can be planned for.

New collectors start at floor average

A new collector is treated as floor average and gets a normal mix of accounts, not the leftovers. That matters for fairness as much as for results. Nobody new is parked on the accounts the rest of the floor would rather not work.

The Floor Assessment (Stage 1) also shows where to start new collectors on your floor. Using it for coaching and hiring.

How often the model is updated

Once live pairing starts, the model is updated about once a month, and weekly in the first couple of weeks.

The roster carries each collector’s start and end dates, which is how joiners and leavers are picked up. Stage 2 runs on daily data files, with assignments delivered by a daily file or the dialer’s router. Files travel by encrypted SFTP to a folder BestPair provides, never by email. BestPair never asks for consumer names, Social Security numbers, phone numbers, email or street addresses, or bank or card details. Account IDs are your own, so BestPair cannot tell who a consumer is. The roster is data about your own staff, and BestPair processes it as personal information. The full data list.

What a wave of leavers does

When several experienced collectors leave in the same month, the BestPair group loses some of its best-known pairings. Expect that, and keep end dates on the roster current, so any change in the monthly result can be traced to a cause instead of argued about.

The comparison stays fair through all of this. Accounts are split at random between the BestPair group and the control group, so turnover lands on both in the same weeks. What the monthly result shows is how much pairing adds with the floor you actually have, churn included. How to prove a routing change worked.

When there is too little history

A floor where few collectors stay long enough to build a record, or where the whole floor is new, gives pairing little to work from. That is better learned before anything changes. The Floor Assessment reads up to 24 months of your call and payment history and says plainly whether there is enough to go on.