Coin Toss Calibration
Would chance explain this QA penalty? Re-roll the sample and see.
The job
Tests every "below standard" verdict a client gives agents from a few sampled calls against the sample size and the agent's wider scores. Penalties stand only where chance cannot explain them, and agents the sample missed still get coached.
The moment
A client has marked outsourced agents below standard from a handful of sampled calls. Each verdict is shown as a coin toss, to test whether chance could explain it.
The client's vendor manager watches agent 7's verdict come up fail in only 41 of 100 redraws of the same eight-call sample.
The outsourcing provider then flags two agents the client's sample missed. The vendor manager withdraws 9 penalties on the spot and agrees a minimum of 30 scored calls.
The running total shows $186,000 in service credits avoided this quarter, and 3 missed agents coached.
What it does
- Log client verdicts
- Stamp Stands / Coin toss / Needs more calls
- Refer a missed agent to coaching
- Propose the minimum sample-size clause
- Send the calibration pack
- Drill evidence calls
What you see
Coin-flip court: one coin per verdict inside a sample-size band, a 100-redraw re-roll animation, and a two-sided 'what the sample missed' panel
What it moves
Service credits in dollars and QA penalties avoided each month, and the share of client penalties based on at least the agreed minimum number of scored calls.
Built for
- QA, training & knowledge
- CX leaders & VoC
- Partners & ecosystem
- Contact-centre operations
