On this page
- Evidence snapshot
- Open the dispute window at the right moment
- Require the agent's specific objection
- Data and decision boundary
- Assign a reviewer who did not score the call
- Compare the scorecard line to the recording
- Decide uphold, adjust, or void
- Write the decision in the agent's file
- Copy-ready call and handoff lines
- Feed the pattern back into calibration
- Review dispute outcomes quarterly
- Questions managers ask
- Keep planning
- Sources
A score dispute is a test of the quality process, not a fight between an agent and a supervisor. When the same person scores the call and decides the dispute, the agent learns that appealing is pointless. A two-reviewer rule adds one independent listen and a written decision. The process below keeps the review fast enough to use and specific enough to trust.
Evidence snapshot
Open the dispute window at the right moment
Tell agents when a score becomes reviewable and how long the window stays open. A window that opens before the coaching conversation turns the review into an argument. A window that closes too fast punishes someone who was on leave. Publish the rule in the quality guide and the scorecard footer. The date should be visible on the score itself.
Start the window when the agent receives the score and the recording. Confirm the receipt so the clock has evidence. If the agent is absent, pause the window and record the pause. Do not let a supervisor close a dispute because the agent did not respond during vacation. The process should be firm about the deadline and fair about the reason.
Require the agent's specific objection
A dispute needs a specific line, a reason, and the evidence the agent wants reviewed. 'I disagree with the score' is not enough to assign a reviewer. Ask the agent to name the behavior, the timestamp, and what the scorecard says. A specific objection is easier to uphold or deny. It also protects the agent from a vague complaint that cannot be answered.
Give the agent a short form with three fields: scorecard line, what happened, and what should change. Let them attach a note or point to a timestamp. Do not require legal language or a long essay. The form should take five minutes, not an hour. If the agent cannot be specific, the coach helps them frame the question before the window closes.
Data and decision boundary
Use this table as a starting point, then match each row to the client's tools and call guide. The manager column stays outside the team member's normal authority.
| Data or request | Team member can | Manager keeps |
|---|---|---|
| Dispute window | Files within the published window | Owns the deadline and the pause rule |
| Objection form | Names the line, time, and requested change | Checks specificity before routing |
| Second reviewer | Listens independently and writes evidence | Assigns a reviewer without a conflict |
| Outcome | Acknowledges the decision | Approves adjustment or void |
Assign a reviewer who did not score the call
The reviewer should be someone other than the original scorer. A second reader reduces the chance that a scoring error survives because the first reviewer defends it. The reviewer needs the rubric, the recording, and the agent's objection. They do not need the original scorer's conclusion before they listen. Independence is the point of the rule, not seniority and the agent should see who reviewed the call.
If the site is too small for a fully independent reviewer, use a reviewer from another queue or a rotating peer. Document the exception and the reason. Do not let the original scorer review their own dispute just because no one else is available. A delayed independent review is better than a fast self-review. Set a maximum delay so the case does not stall.
Compare the scorecard line to the recording
The reviewer listens to the segment named in the objection, then the surrounding minute. They check the rubric's definition before deciding. A score can be wrong because the behavior was misheard or because the rubric was misapplied. Separate those two causes in the decision. The fix differs: one needs a correction, the other needs calibration and the report should name which one applies.
Write down the timestamp and the exact words that support the decision. A decision without evidence is just a second opinion. If the recording is unclear, say so and decide what the standard requires. Do not fill a gap with assumption. The written evidence is what the agent and the client will read later, so keep it short and concrete.
Decide uphold, adjust, or void
Three outcomes are enough: uphold the score, adjust the line, or void the evaluation. Uphold when the rubric and the recording support the original score. Adjust when the line was scored against the wrong definition. Void when the recording or the process is too flawed to judge. Write the outcome in plain words, not only a code so the agent can see the reasoning.
An adjustment should change only the line in dispute unless the review reveals a wider error. A wider error needs a separate note and a check of other calls from the same period. Do not quietly rewrite a whole scorecard to avoid a second review. Voiding should trigger a replacement evaluation when the call still matters. Record who approved the final outcome.
Write the decision in the agent's file
The decision note should include the objection, the evidence, the outcome, and the reason. Keep it short enough to read in one sitting. Give the agent a chance to acknowledge it and add a comment. If they disagree with the decision, record that too. A file that only contains outcomes cannot explain a pattern later or explain why a score changed.
Do not hide the decision from the next coach. A resolved dispute is useful context for the next review. Mark the line as disputed and resolved so it is not reopened by accident. If the outcome changed the score, update the report and the client summary. The record should match what the agent was told, including the date and the reviewer's role.
Copy-ready call and handoff lines
Acknowledging a dispute
I have your objection on the empathy line at 4:12. A reviewer who did not score this call will listen to that segment and the minute around it. You will get a written outcome with the evidence, not just a new number.
Explaining an adjustment
The reviewer found that the rubric requires a hold before the transfer, and the recording shows one. The score is adjusted on that line only. The rest of the evaluation stands, and the calibration team will see the example.
Feed the pattern back into calibration
Collect disputes by rubric line, not just by agent. A cluster on one line suggests the definition is unclear or the calibration drifted. Bring two anonymized examples to the next calibration session. Ask the group to score them before revealing the decisions. The discussion is more valuable than the vote, and the examples keep it grounded and the discussion stays on the rubric.
Update the rubric only when the group agrees on a definition that the examples support. Write the change and the effective date. Rebrief the scorers before the next sample. If the rubric changes often, the dispute count may rise before it falls. Track both the change and the trend, and review the examples again after sixty days so the improvement is visible.
Review dispute outcomes quarterly
A quarterly review should show the number of disputes, the share upheld, and the lines most often adjusted. Compare the pattern with calibration results and client audit findings. A high uphold rate may mean the agents misunderstand the rubric, while a high adjustment rate may mean the scorers do. Neither number is a verdict on its own and together they point to a question.
Ask agents what would make the process more credible. Common answers include faster decisions and clearer examples. Test one change per quarter so the effect can be seen. Report the change to the client with the evidence. The goal is a process agents trust enough to use, not a process that discourages disputes by making them slow or by making the outcome predictable.
Questions managers ask
Can the original scorer sit in the review?
They can provide context if the reviewer asks, but they should not decide the outcome. The independent reviewer listens first and writes the evidence before hearing the original rationale.
What happens when the reviewer cannot decide?
Void the evaluation when the recording or process is too flawed to judge, and record why. If the call still matters, schedule a replacement evaluation rather than guessing.
Sources
- Occupational Outlook Handbook: Customer Service RepresentativesBureau of Labor Statistics, 2026. Employment, wage, training, and outlook figures for customer service representatives.Source 1
- ISO 18295-1:2017 Customer contact centresInternational Organization for Standardization, 2017. Requirements for customer contact centre processes, people, and performance.Source 2
- Fact Sheet #64: Call Centers under the Fair Labor Standards ActU.S. Department of Labor, Wage and Hour Division, 2008. Hours-worked, rest-period, and recordkeeping rules for call center employees.Source 3
- GAO-20-283G: Assessing Data ReliabilityU.S. Government Accountability Office, 2020. A practical framework for documenting data sources, testing reliability, and stating limitations.Source 4
