Customer Service Skills: The 8 We Score When Hiring

By Miki Furman, Founder and CEO. Last updated: 2026-09-03.

What are the most important customer service skills?

There are a great many articles listing eleven or fifteen customer service skills. Empathy, active listening, patience, clear communication. They are not wrong. They are also written almost entirely by companies that do not hire for the role, and they describe qualities rather than a way of choosing between two applicants on a Tuesday afternoon.

We hire customer service agents at volume across the Caribbean and Colombia. What follows is the scoring rubric we actually use, published in full: the weights, the reasoning, and the parts we do not trust. It is version 1.0, dated 2026-08-07.

The 8 customer service skills we score, ranked by weight

Eight of the ten lines on the rubric are skills, and together they carry 85 of the 100 points. The other two lines are readiness checks and sit after the eight. Ranked as a block, voice quality is 40 points and nothing else comes close; ranked line by line, dialer experience is the single heaviest item. Every weight is our own, from rubric version 1.0 dated 2026-08-07, and none of it has been validated against job outcomes yet.

1. Dialer experience, 15 points. Working a predictive or power dialer at volume without the pace wearing you down. Self-reported on the application: over six months scores 15, under six months 7, none 0.
2. B2B appointment setting, 12 points. Getting a business buyer who did not ask to be called to accept a meeting on someone else's calendar. Self-reported: over a year 12, under a year 6, none 0.
3. Fluency, 10 points. Speaking without stalls, restarts or hunting for words. Measured by a model from the applicant's voice recording, as one quarter of the 40-point voice block.
4. Pronunciation, 10 points. Being understood the first time by a caller in the US or Canada. Measured from the same recording.
5. Prosody, 10 points. Rhythm, stress and intonation, which is the difference between reading a script and talking to a person. Measured from the recording.
6. Accuracy, 10 points. The words come out right, with no slips, garbled phrases or wrong forms. Measured from the recording.
7. Campaign-specific knowledge, 10 points. Knowing the client's product, buyer and objections before day one. Self-reported, checked in the interview, and the lightest of the four blocks because it is the field applicants overstate most.
8. General calling experience, 8 points. Time on the phone with strangers in any role, inbound or outbound. Self-reported: experienced 8, some 4, none 0.

The two readiness checks: a working home office scores 8 and being able to start this week scores 7. Neither is a skill, and the readiness section below explains why they are scored anyway.

The 100-point scale, in full

Voice quality, 40 points

A straight average of four sub-scores, each out of 100: fluency, pronunciation, prosody and accuracy, scaled to 40 points. A model listens to the applicant's recording and scores it.

It carries the most weight because it is the only input an applicant cannot simply claim. For an outbound dialing role, how someone actually sounds is closer to the job than any checkbox on a form.

Two honest caveats. The input is objective, the judge is a model, and the model is uncalibrated against job outcomes. And a candidate with no recording is not scored zero. They are flagged as voice_missing and excluded from ranking, because zero and unknown are different facts and burying someone for a missing file would be a scoring error dressed as a result.

Relevant experience, 35 points

Dialer experience outranks knowing the client's industry, and that surprises people. The reason is transferability. Familiarity with a client's subject matter is teachable during the paid training days. Comfort on a dialer at volume is not, and it is what separates the people who are still there in month three. If closing rather than setting is what you are after, our guide to high ticket sales jobs is straight about what that industry actually sells.

Readiness, 15 points

Workspace ready is worth 8 points, able to start this week is worth 7. Neither is a skill. Both are strong predictors of whether an application turns into a person on a shift, which is the thing a recruiter is actually deciding.

Campaign-specific knowledge, 10 points

Scored last and weighted least, because it is self-reported, unverifiable, and the thing candidates most often overstate.

What the weights say about customer service skills

Read as a list of skills, the rubric says something the generic articles do not.

• How you sound outweighs what you know. 40 points against 10 for subject-matter knowledge. On the phone, delivery is the job.
• Transferable beats specific. Dialer comfort at 15 points beats campaign knowledge at 10, because one travels between employers and the other does not.
• Logistics are not a footnote. 15 points for having a workspace and being able to start. A brilliant applicant who cannot start for six weeks does not solve the problem the recruiter has.
• Self-reported claims get discounted on purpose. 45 of the 100 points come from things the applicant asserts, and those are exactly the fields weighted lowest per claim.

The parts we do not trust

This section is the reason the rubric is worth publishing. Most hiring frameworks are presented as though they work.

It has never been validated against job performance. As of 2026-08-07 the database held no outcome data: no training-pass records, no quota attainment, no retention figures. Every weight above is a considered guess rather than a measured coefficient.

Self-report dominates by count. Four of the five input categories are things the applicant tells us. Voice is the only measured one, which is precisely why it carries the largest single block.

It is a filter, not a verdict. Use it to go from 64 candidates to a top 10 worth interviewing. Do not use it to argue that rank 1 beats rank 7. The scale does not discriminate that finely and nothing proves it should.

The scoring bug that made us write the rules down

On 2026-08-07 a scoring script used the wrong option strings for one field. It looked for 6_months_or_more when the real value was over_6_months. Every candidate scored zero on dialer experience, and the output still looked like a clean, plausible ranking. Nobody would have caught it by eye.

Four rules came out of that, and they are the part of this document most worth stealing:

1. Unknown answer values raise an error. They never score zero silently. Failure has to be loud.
2. Missing is not zero. Absent data flags a candidate, it does not sink them.
3. Weights live in one place, at the top of the scoring script, not scattered through the logic.
4. Ties are reported as ties. The script prints how many candidates share a score rather than implying an order the data does not support.

There is also a timezone check that scores nothing. The shift is fixed Eastern, so the script prints local shift hours for anyone whose location implies a large offset. It exists so nobody is interviewed for a shift that lands at 02:00 their time without a human having seen it first.

What would make this rubric real

Five things recorded per candidate: interviewed yes or no, offered yes or no, passed the paid training days, still active at 30 days, hit target in the first full month. With two or three campaigns of that data you can test whether score predicts outcome, drop what does not, and reweight what does.

Until then this is a stated set of assumptions, and the honest thing is to keep saying so.

Customer service skills for a resume vs skills an employer scores

A resume lists customer service skills. A hiring rubric scores them, and listing is not scoring. Recruiters read "client service skills", "customer service skills" and "customer support" as the same line, and since every one of them is self-reported the line on its own is worth close to nothing. What moves an application is demonstration: a voice recording a model can score, dialer and setting history with months attached, workspace and start-date answers that survive a five-minute call, and paid training days that either pass you or do not. Put the skills on the resume, then prove them on the application, where the recording is 40 points and no line of a resume is scored at all.

If you are applying for a customer service job

Read the weights as advice, because that is what they are.

• Record the voice sample. It is 40 points and it is the only part nobody can claim on your behalf. Skipping it does not score you low, it takes you out of the ranking entirely.
• Lead with dialer time. If you have worked a predictive or power dialer at volume, say so first and say how long.
• Answer the logistics honestly. Workspace and start date are 15 points and they are trivially verifiable on a call.
• Do not inflate campaign knowledge. It is worth 10 points, it is the field most often overstated, and it is checked in the interview.

Call Force Global is hiring across the Caribbean and Colombia for remote voice and support roles. The application takes about ten minutes and includes the voice recording described above.

Frequently asked questions