A 10-Point Call Scoring Rubric + Templates for CSRs and Sales Reps

Chris Mechanic
Chris Mechanic
Co-founder, Mecha AI
Think you can stump Jill? Talk to her →

I hope you enjoy reading this blog post. If you want us to just build your AI agents for you, click here.

FACT: If you reviewed EVERY call that happens… And gave your reps detailed feedback… Your numbers would improve DRAMATICALLY.

But for most operators, there’s not enough hours in the day.

And that’s where a call scoring rubric comes into play.

A rubric is just a set of rules. It probably lives in your head right now.

Once you get it down on paper, ANYONE on your team can score calls, just like you would. Which makes it scalable beyond yourself #leverage

And you’ll eventually graduate to a call scoring AI or something even more scalable.

But having a template is still useful.

Here’s an 80/20 rubric you can make your own


Key Takeaways

  • Ten yes or no checks are enough. The rubric below fits on one page, and a rep’s score is the number of yeses.
  • Outbound calls need five swapped checks. Permission, the reason for calling, “not interested,” the booking ask and voicemail replace their inbound cousins.
  • CSRs and sales reps get the same card with different definitions. One column of the table covers each role.
  • A 7 is good. An 8 is great. A 9 or 10 is exceptional. Recheck your scoring before you celebrate a perfect score.
  • Five calls per rep per week is enough to spot a habit. Coach the lowest scorers on one check at a time and rescore that check the next week.

Which Calls Go in the Denominator

Before you score anything, clean the denominator. A rep who answers a vendor, a wrong number and a billing question had zero booking chances on those three calls.

Tag every call you pull as one of three types: qualified new lead, existing customer, or other. Compute booking rate on the first group only.

Booking rate = qualified calls booked ÷ qualified calls received.

Existing customers, vendors, recruiters and wrong numbers still get scored on greeting and pacing, because your reps handle them. They sit outside the booking rate.

Call Scoring Rubric Template for CSRs and Sales Reps

Listen once, at normal speed. Mark each line yes or no. Add up the yeses. The same ten checks work as a call quality scoring card for both roles, and the right-hand column shows how each one changes for a sales rep.

#CheckYes means (CSR)Yes means (sales rep)
1GreetingCompany name, rep name, and a clear offer to help in the first few secondsRep and company named, and the caller knows who they reached
2Need discovery firstThe rep asked about the problem before mentioning availability or office hoursThe rep asked about the project, the why and the decision-makers before pitching
3Info captureName, callback number, address, the problem and the timeline are on the ticketScope, budget range, timeline and who else decides are in the notes
4Booking attemptThe rep asked for the appointment with a specific day or windowThe rep asked to set the in-home visit or estimate with a specific day or window
5Price question handledA range with context, or a move to book the visit firstBudget raised without flinching, and a range given only after discovery
6Handoff handledIf transferred, the rep explained why and the other person picked upIf the call goes to a closer or estimator, the caller heard who and when
7Next step confirmedThe caller heard what happens next, who is coming and whenThe caller heard what happens next and what the rep owes them
8Cues caughtThe rep acknowledged urgency or frustration the caller voicedThe rep acknowledged hesitation or an objection before answering it
9PacingNo dead air long enough to make the caller ask “are you there?”The caller finished talking before the rep started
10Wrap-upRecap, thanks, and a clean closeRecap, thanks, and a confirmed time on the calendar

Some lines need a sentence of explanation.

Check 2: need discovery first
This one matters most when your calendar is full. “We’re booked out two weeks” lands very differently after the caller has described a leaking water heater than before. Ask about the problem first, then deal with the calendar.

Check 5: price questions
Either answer scores a yes: a range with context (“most of these run between X and Y, and the tech confirms on site”), or “let me book the visit and you will have a firm number then.” A flat “I cannot give prices over the phone” scores a no, because a caller with no number has every reason to dial the next company.

Check 6: handoffs
If there was no transfer, mark it yes. A handoff that rings out is a no. Keep a weekly count of transfers attempted and transfers picked up, because a rep can do everything right and still lose the caller to a phone nobody answered.

A scorecard with 40 line items gets filled out once, in January, and then lives in a shared drive next to the org chart.

-> Curious how many calls your team never reaches? Run the free Phone Leak Analyzer and find out in 60 seconds.

Inbound vs Outbound Calls and What Changes on the Card

An inbound caller already wants something. An outbound call interrupts someone’s afternoon. Five checks change to match.

Inbound checkOutbound swapYes means
1 GreetingPermission and openingThe rep named the company and asked if now is a good time
2 Need discovery firstReason for calling stated earlyThe caller heard why you called within the first sentence or two
4 Booking attemptBooking askThe rep asked for a specific day or window after a short conversation
5 Price question”Not interested” handledThe rep acknowledged it, asked one easy question, and accepted a no
6 HandoffVoicemailThe rep left a short message with a reason and a callback number

The other five checks carry over unchanged.

Outbound calls in home services usually fall into three jobs: a quote follow-up, a reactivation call to an old customer, and a speed-to-lead callback to someone who just filled out a form. The reason for calling is different each time, so score it against the right one. Our post on following up on quotes covers the timing and wording for the first job.

Scoring CSRs vs Sales Reps

CSRs
A CSR owns booking, capturing the details on the ticket, and handling existing customers who call about a visit, a bill or a reschedule. Weight checks 3, 4 and 6 most heavily. Our post on dispatcher training explains where the CSR chair ends and the dispatcher chair begins, which tells you which calls belong on a CSR’s card.

In-home sales reps and comfort advisors
Their calls run longer and the outcome is a set appointment or a signed job. Weight discovery (check 2), budget and timeline (check 3), setting the appointment (check 4) and objection handling (check 8). A sales rep who skips discovery and pitches in minute one will score low on three checks at once.

Compare reps within a role, CSR against CSR and closer against closer.

How Many Calls to Score Per Rep

My recommendation is five calls per rep per week, drawn at random from the qualified new-lead pile, plus any call that ended in a complaint.

Say you have four reps and an average call runs four minutes. Five calls each is 20 calls, and 20 calls times 4 minutes is 80 minutes of listening, plus marking time. One person can do that on a Friday afternoon.

The honest counterpoint. A yes or no card measures behavior. It does not measure whether the caller was in a good mood or whether the lead was any good. With two reps, five calls a week is ten data points: enough to spot a habit, too little to rank people. Judge the pattern over a month before you act on one week.

What to Do With the Lowest Scorers

Pick the single check the rep misses most often. Coach only that one.

Pull two of their calls where they missed it and one from your best rep where they hit it. Play all three. Then hand over a script to practice. Our call script library has a front-desk phone answering script to start from. Edit it until it sounds like your company.

Rescore that same check the following week. When it holds for three weeks in a row, move to the next one.

If the same check is low across the whole team, look at your process. A rep cannot follow a script nobody gave them. Fix the script, then rescore.

How We Score Our Own AI Agent’s Calls

We built a call reviewer that scores our AI agent’s calls. It reads each call and rates six dimensions from 1 to 10, then gives an overall score from 1 to 100. It also notes the cues the agent missed and which details it captured, without a score. The six questions work just as well for a human rep’s recording, and they overlap with the ten checks above.

DimensionWhat it asks
Greeting and framingDid the caller understand right away who they reached? Was it warm or robotic?
Information sequencingDid the agent gather the caller’s needs before revealing constraints like “the office is closed”?
Pacing and timingUsing the transcript timestamps, were there awkward pauses, rushed parts, or interruptions?
Conversational qualityDid the word choices sound natural, with acknowledgments and empathy where the moment called for them?
Call endingDid the wrap-up confirm next steps, and did the caller seem cut off or wanting to say more?
Sales opportunityWas there a chance to build rapport, capture budget or timeline, or create urgency that the agent missed?

Three parts of the design carry over to your own card.

The calibration line
The reviewer’s instructions say to be honest and critical: a 7 is good, an 8 is great, and a 9 or 10 is exceptional. Without a line like that, scores tend to creep upward until a 9 means very little.

The sequencing rule
Our reviewer’s instructions treat leading with a constraint (“we’re closed right now”) as a cause of hangups, and gathering the caller’s need first as the way to keep them on the line. This is check 2 on the rubric above, and it gets a whole dimension in our reviewer.

The single biggest improvement
Each review ends with one specific, actionable change, such as the exact sentence to say after the caller describes the problem. One fix per call is coachable. A long list is not.

Scoring Software and AI Tools

Software that records, transcribes and grades calls now exists, and AI can do the grading quickly. The input they need is the same short rubric you just read, so a clear yes or no card makes any tool work better.

Score by hand for a few weeks first. You learn what good sounds like in your company, and you can tell later whether software scores agree with your ear.

For calls your team cannot get to, an AI agent can answer when your lines are busy and warm-transfer to your team with full context when the caller asks for a person or the call needs human judgment. For routine status and scheduling calls, AI customer service and dispatch frees your reps for the calls worth scoring.

Questions About Scoring Calls

What is call scoring?
Call scoring is rating a recorded call against a fixed list of behaviors so you can compare reps and weeks. A ten-line yes or no rubric and a spreadsheet are enough to start.

What should a call scoring rubric include?
A call scoring rubric should include a greeting, need discovery, info capture, a booking ask, a price answer, a confirmed next step and a clean close. Add checks for pacing and for cues the caller voiced. Keep it to about ten lines, so it gets used.

Can AI score calls?
Yes. AI can read a transcript and grade it against a rubric quickly. A human still decides what the rubric rewards and spot-checks the scores.

How is call quality scoring different from call monitoring?
Call monitoring is listening to calls, live or recorded. Call quality scoring writes down what you heard in a form you can compare, so the listening leads to coaching.

The Final Word

Your phone team decides how much of your ad spend turns into booked jobs. If you record calls, the raw material already exists, and a call scoring rubric turns it into coaching.

  1. Tag each call: qualified new lead, existing customer, or other.
  2. Score five qualified calls per rep per week against the ten checks.
  3. Swap in the five outbound checks for outbound calls.
  4. Coach one missed check at a time, and rescore it.

Start this Friday with 20 calls and a spreadsheet.

-> Want to hear what a well-run front-desk call sounds like? Try Jill, our voice AI demo agent. She greets you, asks about your project and books an appointment, the way she would for a real customer.

Chris Mechanic
About the author
Chris Mechanic
Co-founder, Mecha AI

Chris Mechanic is the co-founder of Mecha AI, building voice AI agents purpose-built for home services companies doing $5M+. Before Mecha, Chris co-founded WebMechanix, a performance marketing agency that generated millions of leads for home services brands, and saw firsthand how missed calls and slow response times cost contractors millions in lost revenue.

Share

Talk to Our AI Agent Right Now

Experience a voice AI agent for yourself. You'll be a believer in about 90 seconds.

Talk to Our AI Agent ↗