QA retrospectives produce one meaningful output: an updated coaching priority list based on what actually changed in the previous cycle. Teams that skip this structured review end up recycling the same coaching priorities each month regardless of whether they worked, which is how coaching programs stay busy without improving outcomes.
This guide covers how to structure a QA retrospective so it changes what coaching does next cycle, not just how the last cycle is documented.
What you need before you start: Criterion-level QA scores from the completed coaching period (minimum 4 weeks of data), the coaching priority list from the start of that cycle, and a QA lead who attended or reviewed sessions during the period.
What is a QA retrospective in a contact center?
A QA retrospective is a structured review that evaluates the outcomes of a completed coaching cycle. It asks three questions: what improved, what did not move, and what regressed. The output is an updated coaching priority list for the next cycle, based on QA scoring evidence rather than manager intuition.
How is a QA retrospective different from a performance review?
A performance review evaluates an individual rep's results against targets. A QA retrospective evaluates the coaching program itself: whether the coaching delivered produced the score movements expected, and where the approach needs to change. The subject of a retrospective is the coaching system, not the individual rep.
Step 1: Set a Cadence That Matches Your Coaching Frequency
Run retrospectives monthly for teams with weekly or bi-weekly coaching sessions, and quarterly for teams with monthly coaching cycles. Weekly retrospectives create noise: the data window is too short to distinguish a real trend from a single bad week.
A monthly retrospective covers 4 to 5 weeks of coaching data, which is long enough to see whether a coached behavior actually changed in subsequent calls. A quarterly retrospective gives you 3 full coaching cycles to compare, which is the minimum needed to distinguish genuine coaching impact from regression to the mean.
Step 2: Pull Criterion-Level Score Data, Not Composite Scores
The inputs for a useful retrospective are specific. You need score trends by individual evaluation criterion over the coaching period, not overall QA scores. You also need score change data for reps who received coaching on a specific criterion compared to those who did not.
SQM Group contact center benchmarks indicate that QA programs using criterion-level tracking identify coaching gaps significantly faster than programs using composite scores alone. Composite scores mask which behaviors improved. A rep's overall score can hold steady while empathy improves and compliance worsens simultaneously.
Pull criterion-level scores for the top three coaching priorities from the completed cycle. For each criterion, calculate the average score at the start versus the end of the cycle. A movement of 3 or more percentage points on a criterion after focused coaching is meaningful. Movement under 1 point suggests the coaching approach is not reaching that behavior.
Insight7's QA dashboard surfaces criterion-level score trends per rep and across the full team. Managers filter by time period, criterion, and rep group to see which coached behaviors moved. The platform also shows coaching sessions assigned versus completed, so the retrospective data includes whether coaching was actually delivered before evaluating whether it worked.
Step 3: Sort Results Into Three Buckets
Before the retrospective meeting, sort criterion-level data into three categories. Improved means the criterion score rose by 3 or more points across coached reps. Did not move means the score is within 1 point of where it started. Regressed means the score dropped.
Each bucket requires a different response in the next cycle:
- Improved criteria can move to maintenance coaching with fewer sessions
- Criteria that did not move need a coaching approach change, not more of the same sessions
- Regressed criteria become the priority reassignment for the next cycle
A common pattern: a criterion shows no movement because the definition is ambiguous, not because the coaching failed. If "empathy" is defined as a yes/no on whether the agent used the customer's name, coaching to "improve empathy" will not produce score movement because the criterion is not measuring what the coaching targets.
Step 4: Separate Systemic Issues from Individual Rep Problems
If a criterion did not move for the majority of your team, that is a systemic signal. The coaching approach, criterion definition, or session frequency needs to change. If the same criterion improved for most reps but stayed flat for three specific reps, that is an individual performance issue, not a systemic failure.
Insight7 automatically generates practice sessions for reps based on QA scorecard feedback. Supervisors review and approve before deployment, so human judgment stays in the loop while the data surfaces the systemic pattern. Fresh Prints, a referenceable Insight7 customer, noted that the ability for reps to practice the specific gap identified by QA immediately after a session was a qualitative shift: coaching recommendations became actionable the same day, not at the next scheduled session.
Step 5: Update Coaching Priorities for the Next Cycle
The retrospective produces one output: an updated priority list for the next coaching cycle. Cap it at three criteria per role type. More than three means sessions are spread too thin to move any individual criterion meaningfully.
For each updated priority, document two things: the specific coaching approach (role-play, call review, side-by-side, or AI practice session), and the threshold score movement that will count as success at the next retrospective. Setting the success threshold before the cycle starts prevents rationalizing flat results afterward.
Decision point: If a criterion has been a coaching priority for two consecutive cycles without movement, escalate to criteria definition review before the third cycle. Persistent non-movement usually means the rubric is ambiguous rather than the coaching is inadequate.
If/Then Decision Framework
If your retrospective is producing flat priority lists cycle after cycle, then add criterion-level tracking before running another session. Composite scores cannot produce specific enough findings to change coaching approach.
If coached criteria improve for most reps but stay flat for a consistent subset, then separate those reps into an individual performance track rather than continuing systemic approach changes.
If a criterion definition has not been updated in more than 3 months, then review it before attributing non-movement to coaching failure.
If your coaching program has no documented success thresholds per criterion, then set them before the next cycle starts. Without pre-set thresholds, every outcome can be rationalized.
Step 6: Keep Retrospective Documentation Functional
The retrospective documentation needs to be short enough that coaches will actually reference it before their next sessions. A one-page format works: three buckets (improved, flat, regressed), three updated priorities for the next cycle, and the coaching approach assigned to each priority.
The test for whether your documentation format is working: ask a coach two weeks into the next cycle to name the top coaching priority for their team without looking it up. If they cannot, the documentation is not functional.
FAQ
How to measure the success of a coaching program?
Measure criterion-level score movement on coached behaviors across consecutive coaching cycles. The primary metric is score change per criterion per period, not overall QA score. After three retrospective cycles, teams should see coached criteria move by at least 5 percentage points over the full period. The number of criteria on the "did not move" list should shrink each cycle as ambiguous rubric definitions are corrected. Insight7 surfaces this data in a format designed for retrospective-ready reporting.
What are the 5 C's of coaching?
The 5 C's of coaching are commonly described as Context, Contract, Conversation, Commitment, and Close. In a QA retrospective context, they map to: setting context with QA data (Context), agreeing on what success looks like for the next cycle (Contract), reviewing criterion trends in the retrospective meeting (Conversation), documenting updated priorities and approaches (Commitment), and scheduling the next retrospective with pre-set thresholds (Close).
QA managers and coaching leads running monthly or quarterly reviews: see how Insight7 surfaces criterion-level coaching outcome data for retrospective-ready reporting at insight7.io/improve-quality-assurance/.
