Call operations research
Estimating the review time for a weekly sample of recorded calls
A practical method for estimating how long a weekly call-review sample takes, including listening, scoring, notes, and coaching follow-up.

Table of contents
A weekly call sample is easy to promise and easy to underestimate. The review includes selecting calls, listening, scoring, writing notes, and following up with the agent. If the plan counts only listening time, the schedule will slip and the sample will shrink. This review describes a way to estimate the full task before committing to a review cadence.
Key takeaways
- Count the full review cycle, not only the call duration.
- Measure selection, listening, scoring, notes, and coaching separately.
- Use a small timed pilot before setting a weekly quota.
- Cap the sample to the time the reviewer actually has.
- Recalculate after the scorecard or recording system changes.
Research question and scope
This review asks how much reviewer time a weekly sample of recorded calls requires in a small phone team. The unit is one review cycle for one call, from selection to the completion of the coaching note. The review includes the time spent locating the recording, listening, scoring, and writing. It does not measure the cost of the phone system or the agent's time.
The scope covers recorded inbound calls reviewed by a supervisor or quality lead for a small team. It excludes live call monitoring, customer surveys, and formal disciplinary investigations. It also excludes the time needed to build a new scorecard from scratch. Fix the boundary before timing the work.
The review does not decide the right sample size for every team. That depends on the purpose of the review, the risk of the calls, and the reviewer's available hours. The review provides a time estimate that a team can use to choose a sample it can sustain. A smaller sample completed every week is usually more useful than a large sample that is abandoned.
Record the scorecard version, the recording system, the reviewer, and the call length distribution. A team that changed its scorecard during the period will produce a misleading estimate if the cycles are pooled. Split the sample by scorecard version and report each one. The reader should know which process produced which time.
Definitions and unit of analysis
Define a review cycle as the work needed to turn one selected recording into a completed score and a coaching note. The cycle starts when the reviewer opens the recording and ends when the note is saved. It includes any time spent re-listening to a section and checking a policy reference. It also includes the short pause before a difficult score is written.
Define a scored call as one with a completed scorecard and a short written note. A call that was listened to but not scored is not a completed cycle. A call that was scored without a note is partially complete. Record the status so the time estimate does not mix finished and unfinished work.
The unit is the call, not the hour. A one-hour block may contain three short calls or one long call. The estimate should be reported per call and per week. A per-hour figure hides the variation in call length and review depth.
Decide how to treat interruptions. A review cycle interrupted by a phone call or a meeting may take longer in clock time but not in active work time. Use a timer that can pause, or record both active and elapsed time. Report which measure the team will use for planning.
Sampling and timing method
Draw a small pilot from the calls the team already plans to review. Include a mix of short and long calls, simple and complex issues, and new and experienced agents. Do not choose only the easiest calls. The pilot should resemble the weekly sample the team intends to run.
Time each stage separately. Record the time to select the call and open the recording, the time to listen, the time to score, the time to write the note, and the time to deliver coaching. Use a simple timer and a shared form. The separation shows where the cycle is expensive and where it can be shortened.
Repeat the timing for at least ten calls before setting a quota. A single call may be unusually short or unusually difficult. Ten calls give a rough range without requiring a large study. Record the median and the longest cycle, not only the average. The longest call often drives the weekly schedule.
Include the setup time that happens once per review session. Opening the scorecard, checking the queue, and exporting the list are part of the work. Divide that setup time across the calls in the session or report it separately. A quota that ignores setup will always be missed.
Measuring the stages
Listening is usually the largest stage, but it is not the only one. A reviewer may listen to a call twice to check a policy answer or a readback. Record the number of passes and the reason for each. A second pass is not automatically waste; it may be the quality check that catches an error.
Scoring time depends on the number of fields and the clarity of the definitions. A scorecard with twenty subjective fields will take longer than one with eight specific checks. Time the scoring separately from the listening. If scoring dominates, simplify the scorecard before adding more calls.
Note-writing time depends on the template. A blank text box invites long notes and indecision. A short template with a few required fields is faster and easier to compare. Time the note separately. If the note takes as long as the call, the template is probably too open.
Coaching time is often left out of the estimate. A review that produces a score but no feedback does not change behavior. Include the time to prepare and deliver one coaching point. If the team uses a weekly group session, divide the session time across the calls discussed. Report the coaching model beside the estimate.
Analysis and limitations
The pilot cannot predict every week. Call volume, issue mix, and reviewer experience change the time. The estimate is a planning range, not a guarantee. State the range and the assumptions rather than presenting a single number as exact.
The reviewer may work faster during the pilot because they know they are being timed. This can make the estimate too low. Compare the pilot with the first two real weeks. If the real weeks take longer, use the real numbers for the quota. The pilot is a starting point, not the final answer.
The sample may not represent the calls that are hardest to review. A random sample can miss the long, complex calls that consume the most time. Add a short review of the longest calls in the period. Report them separately so the average does not hide the tail.
The scorecard may change after the estimate. A new field or a new policy can add time without changing the call mix. Recalculate the estimate after any material change. Do not carry an old quota into a new scorecard and assume it still fits.
Repeating the estimate
Repeat the timing exercise when the scorecard, recording system, or review owner changes. A new system may add search time or remove it. A new owner may have a different pace. Record the version and the date beside the estimate. The next reviewer should be able to see why the quota changed.
Use the result to set a sample the team can complete every week. A smaller sample with a completed note is more useful than a larger sample with a backlog. Test the quota for four weeks before changing it. A short trial is easier to interpret than a permanent rule.
Share the estimate with the reviewer and the team. Ask which stage feels heaviest and which fields are unclear. The reviewer often knows where the scorecard creates extra work. Their context can turn a time estimate into a better process. The QA review routine is the operational companion to this estimate.
Keep the review log visible during the week. A reviewer who can see the remaining sample and the time left is more likely to finish. A quota that is reviewed only at month end is easy to ignore. The record should support the work, not just the report. The reviewer calibration study covers the related problem of keeping scores consistent.