Should calls that were never going to book count against a rep's score?
Not in the booking metric, but yes in the handling score. A wrong number or a vendor call should be classified out of booking rate entirely. A caller outside your service area should still be handled well and scored on it. The working rule: exclude from outcome metrics anything the rep could not influence, and score handling on every call where a human actually spoke to someone.
Two metrics, two different denominators
Booking rate and handling quality answer different questions and should never share a denominator. Booking rate asks what share of winnable opportunities the rep won. Handling quality asks whether the rep did the job well on the call in front of them, winnable or not.
Collapsing them punishes whoever draws the worse queue. It also lets a genuinely weak rep hide behind a favorable mix, which is the same error in the opposite direction.
Four tiers worth classifying
- Not a customer contact. Wrong numbers, robocalls, vendors, internal calls. Excluded from both metrics.
- Contact, no opportunity. Existing customer confirming an appointment, a billing question, a returned call. Handling scored, booking excluded.
- Opportunity you cannot serve. Out of area, service you do not offer, equipment you do not work on. Handling scored, booking excluded, but track the volume because it tells you something about your advertising targeting.
- Bookable. Everything else, including price shoppers inside the service area. Both metrics apply.
Where exclusion quietly goes wrong
Any exclusion rule can be abused, whether by a rep tagging dispositions or by a classifier that is too generous. The safeguard is to audit exclusion rates by rep. If one person's calls are classified unbookable at twice the rate of their peers, that is not a coincidence and those calls should be listened to.
The second safeguard is to require a reason code rather than a checkbox. "Out of area" with the city captured is auditable. "Not a lead" is not. Consistent classification across every call, applied the same way every time, is one of the practical arguments for automated call classification over rep-entered dispositions.
The category that causes the most arguments
Existing customers with a warranty or callback complaint are the hard case. There is no new revenue to book, the call is often unpleasant, and the rep's handling matters enormously to whether that customer stays. Scoring them on booking rate is meaningless; leaving them unscored tells the team those calls do not count.
Give them their own small rubric focused on ownership, accurate commitment, and follow-through, and report them separately. What happens on those calls tends to show up later in retention and review scores, which is why it belongs in customer intelligence rather than buried in a sales metric.
Topics: call classification · booking rate · fairness · scoring · QA
Have a version of this question about your own business?
The useful answer usually depends on which systems you run and how they're connected. That's a conversation, not a blog post.