Short answer: A QA score doesn't change anything by itself. To turn scores into coaching, look past the total score to the specific behaviors that failed, pick one behavior per agent to work on, coach it using the actual call, have the agent practice the fix, and check the next batch of scored calls to see if it stuck. Do it every week, in short sessions, not once a month in a long review.

I've seen a lot of QA programs that score thousands of calls and change nothing.

The scores go into a spreadsheet. The spreadsheet goes into a monthly report. The agent gets a number and a "keep it up" or a "needs improvement." Next month, same number.

That's not a QA problem. It's a coaching problem. QA tells you what happened. Coaching is the part that changes what happens next.

Here's the process I'd use to connect the two.

Why Scores Alone Don't Coach Anyone

Tell an agent "you scored 78 this month" and watch what happens. They want to know which calls. Then they want to argue about which calls. Then they leave the meeting not knowing what to do differently.

A total score mixes too much together. Greeting, verification, empathy, hold handling, compliance and closing all get blended into one number. The agent can't act on a blend.

Behaviors are different. "On 7 of the 30 calls we scored, you moved to the fix before acknowledging the problem" is something an agent can actually work on tomorrow.

So step one is to stop coaching to the score.

Step 1: Break the Score Into Behaviors

Pull every scored question for the agent over the last week or two. For each question, count how often they passed and how often they missed.

Then sort by misses. You'll usually see one of three patterns:

This only works if you have enough scored calls per agent to see a pattern. With 4 or 5 calls a month, one bad call looks like a trend. (We covered why in How Many QA Evaluations Do You Need Per Agent?)

Step 2: Pick One Thing

This is the hardest part for supervisors. They see six things the agent could do better and they want to cover all six.

Don't. Pick one. Two at most.

People change one habit at a time. If you hand an agent six things, they'll work on the one that's easiest, and that's rarely the one that matters most.

How to choose:

  1. Compliance first. Always.
  2. Then whatever hurts the customer most. Wrong information, unresolved calls, transfers that shouldn't happen.
  3. Then whatever shows up most often. A behavior missed on a third of calls matters more than one missed once.

Write the one behavior down in plain words. "Acknowledge the caller's problem before you start solving it." Not "improve empathy."

Step 3: Coach on the Real Call

Pick one call where the agent missed the behavior, and one where they nailed it if you have one.

Play the moment. Not the whole call. Just the 30 seconds that matter.

Then ask questions before you give answers:

Agents usually know the right answer once they hear themselves. When they find it, it sticks better than when you tell them.

Playing the call where they got it right matters too. It proves they can do it. Now it's about doing it every time.

Keep the session short. Fifteen minutes on one behavior beats an hour on ten.

Step 4: Practice Before the Next Shift

Knowing the fix isn't the same as doing it on a live call with a frustrated customer.

So have the agent practice. Role-play the moment with a supervisor or a peer. Say the new words out loud a few times. It feels awkward, and it works.

The goal is simple: the first time they try the new behavior shouldn't be on a real customer.

A five-minute role-play with a supervisor or a peer is plenty. Just don't skip practice.

Step 5: Check That It Stuck

This is the step most coaching programs skip, and it's the one that tells you whether coaching works at all.

After the session, watch that one behavior on the agent's next batch of scored calls. Did the misses go down?

Write down what you coached and when. If you don't track it, you can't tell good coaching from busy coaching. Behavior change after coaching is one of the QA numbers I'd put in front of leadership.

Make It Weekly, Not Monthly

Monthly coaching sounds organized. In practice, it means an agent repeats the same mistake for weeks before anyone talks to them about it.

A weekly rhythm is easier on everyone:

Day What happens
Monday Supervisor reviews last week's scored calls by behavior for each agent
Tuesday to Thursday 15-minute one-on-ones: one behavior, one call, one practice
Friday Quick check: did anything get escalated? Any compliance misses to follow up?
Next Monday Look at the coached behavior on the new calls. Keep going or move on.

Adjust the days to your operation. The point is a short loop. Score, coach, practice, check, repeat. For new hires in nesting, tighten it to same-day feedback.

Look for Team Patterns Too

If eight agents on the same team miss the same behavior, that's not eight coaching problems. It's one training problem, or a script problem, or a form problem.

Roll the behavior misses up by team and by call type every so often. When the same miss shows up everywhere, fix it once at the source instead of coaching it eight times. Complaint calls are a good place to start: tagging each one with a root cause shows you where the fix belongs.

Keep QA From Feeling Like Punishment

Agents who fear QA game it. They argue scores, they hide calls, they stop trying new things.

A few things help:

Where OttoQA Fits

Coaching runs on data. The more calls you score, the clearer the pattern for each agent, and the less time supervisors spend arguing over a 5-call sample.

OttoQA scores calls against your own QA forms, so the behaviors you coach are the questions you already use. No seat licenses, no IT, and setup takes about 2 weeks. There's a $1,000 one-time setup fee, and you can start with a 200 to 500 call proof of concept: we score 200 to 500 of your calls against your form.

Bottom line... a score is the start of the conversation, not the end of it. Coach one behavior, on a real call, every week.

FAQ

How do you turn QA scores into coaching?

Break each agent's scores into individual behaviors, pick the one or two behaviors missed most often, coach them using a real call where the miss happened, have the agent practice the fix, and check the next scored calls to see if it improved.

How often should call center agents be coached?

Short, frequent sessions work better than long monthly reviews. A weekly 15-minute session focused on one behavior keeps the feedback close to the calls and gives you a fast way to see if coaching is working.

What should a QA coaching session include?

One behavior, one real call, and one practice. Play the key moment, ask the agent what happened and what they'd do differently, role-play the better response, and agree on what you'll look for on their next calls.

How do you know if agent coaching is working?

Track the specific behavior you coached on the agent's next batch of scored calls. If the misses go down, move to the next behavior. If they don't, change the approach or look at the script, training or form.

Should QA scores be used for discipline?

Be careful. Small samples swing a lot, and agents who fear QA stop improving. Keep coaching sessions about getting better, and base any serious decisions on a large, fair set of calls.

About the author

Tom Laird is the founder of OttoQA and CEO of Expivia, with about 30 years in contact centers. He hosts the Advice from a Call Center Geek podcast. Connect with Tom on LinkedIn.