Compare each rep's average call score with the median, mode, and count of scored calls before choosing a coaching focus.
Key takeaways
- Write down the filter before calculating scores.
- Compare the average with the median and mode.
- Check the scored call count before drawing conclusions.
- Listen to the calls behind an unusual pattern.
You score each recorded call against your rubric, collect the scores for each rep, and calculate one figure. That figure is easy to put on a slide, easy to sort, and easy to explain. It is also easy to misread. Two reps can share the same average and still need very different coaching, because the average says nothing about how their scores are spread or how many scores it is built from.
This post covers three values to place beside each rep's average: the median, the mode, and the count of scored calls. Each one answers a question the average cannot answer by itself. Together they help you decide whether the average is a fair summary of a rep's work and what to do next.
Before looking at the three values, it helps to be exact about what the average is.
The average is the mean of the calls that match your filter
The average is the arithmetic mean of the scored calls that meet the conditions you set.
The Microsoft Support page for AVERAGEIFS describes the function as one that "Returns the average (arithmetic mean) of all cells that meet multiple criteria." The page covers the function in Microsoft Excel: "This article describes the formula syntax and usage of the AVERAGEIFS function in Microsoft Excel." The phrase that matters for a sales manager is "meet multiple criteria." An average for a rep includes only the rows that pass every condition in the formula.
That has a practical consequence. If the rep column is right but the date range, the call type, or the rubric version is wrong, the average will look precise and still be wrong. Before you compare any values, write down which calls went into each number. A short note is enough: the rep names, the start and end dates, the call types you included, and the version of the scoring guide you used.
Rubric changes deserve special attention. If you revised how a criterion is scored partway through a quarter, scores from before and after the change may not mean the same thing. You can still average them, but say so next to the number, or split the period and report each part separately.
The median is the score in the middle
The median is the score in the middle when a rep's scores are put in order from lowest to highest.
The AVERAGEIFS page gives this example: "For example, the median of 2, 3, 3, 5, 7, and 10 is 4." In that list, 4 sits halfway between the two middle scores, 3 and 5.
The median is useful beside the average because it responds differently to unusual calls. A single very strong or very weak call can move the average noticeably. The median is much less affected by one outlier, because it only cares where the middle falls. When a rep has an even number of scores, the median sits halfway between the two middle values.
Read the two numbers against each other:
- If the median is well below the average, a few strong calls are probably lifting the average. The typical call is weaker than the average suggests. Check which calls sit at the top before concluding the rep is strong across the board.
- If the median is well above the average, a few weak calls are probably pulling it down. Look at the low scores before deciding the rep has a broad problem. Sometimes one bad call has a cause that is not the rep's skill, such as a poor connection or a call that ended early.
- If the median and the average sit close together, the average may be a fairer summary of the center. You still need the mode and count, and scores can sit far from the middle on both sides.
These readings are starting points for review, not verdicts. The median tells you where to look, and the calls tell you what happened.
The mode is the score that appears most often
The mode is the score that appears most often in a rep's list of scores.
The same AVERAGEIFS page gives a matching example: "For example, the mode of 2, 3, 3, 5, 7, and 10 is 3." In that list, the score 3 appears twice and every other score appears once, so 3 is the mode.
The mode is a different kind of signal. The average and the median summarize the whole range of scores. The mode shows the score a rep gets most often, which points to a habit rather than a single call. If a rep's most common score on a criterion is low, you have a repeated pattern you can coach. If the most common score is high, the rep may be consistent on that criterion, even if a few calls dragged the average down.
The mode has limits, and you should state them. A list of scores can have two values that tie for most common, or no value that repeats at all. In those cases, report the tie or say there is no clear mode. Do not pick one value just to fill the cell.
When the mode and the median point in different directions, look closer. The most common score may sit below the middle score, which means a low score repeats often while a few higher scores push the middle up. Find the calls behind each value before drawing a conclusion.
The count of scored calls shows how much evidence sits behind each number
The count of scored calls tells you how many calls sit behind each average, median, and mode.
Two averages can display the same number and carry very different weight. An average built from a handful of calls can change a lot when one more call is scored or one is removed from the filter. An average built from a longer run of calls is usually steadier, because one unusual call has less effect on the average. The count tells the reader how much evidence supports the rest of the row.
The count also guards against a quiet error. If a filter leaves no matching calls, the formula does not return a score. The AVERAGEIFS page gives this case: "Because there are no scores greater than 95, #DIV0! is returned." Treat that error as a sign that the filter matched nothing. It is not a score, and it is not zero. In your comparison table, show a dash for that rep's average and show the count as zero, so the reason stays visible.
Set your own minimum count before you make decisions based on a rep's numbers. Choose it based on how many calls you can realistically score each period and how much a single call could change the result. Write the threshold down and apply it to every rep. If a rep falls below it, the honest answer is that you do not yet have enough calls to judge. The next step is to score more calls, not to rank the rep.
The count also sets context in conversation. When you share numbers with a rep, starting with the count tells them the figures describe a specific body of work, not a verdict on their whole career.
Filter first, then average
Average only the calls that meet a condition, and say which condition you used.
Filters you might use include the rep, the type of call, the date window, and the scoring guide version. Each one changes the question you are answering. An average across all calls answers "how does this rep usually do across everything I scored?" An average for one call type answers "how does this rep do on this kind of call?" Both can be useful, but they are not interchangeable, so label each number with its filter.
Be careful with filters based on the score itself. If you average only the calls that scored above a certain level, or only those below one, you have changed the question again. The result describes that band of calls, not the rep's overall work. To see the high and low ends, compute them as separate, clearly labeled rows rather than replacing the full average.
For a fair comparison, use the same filter for every rep. If one rep has calls from three months and another has calls from one month, the numbers are not measuring the same thing. Matching the window and call types matters more than finding the perfect window.
Put the three values side by side
Place the average, median, mode, and count in one table, with one row for each rep and the same filter applied to every row.
A simple layout works well:
| Rep | Average | Median | Mode | Scored calls |
|---|---|---|---|---|
| Example | Mean of matching calls | Middle score | Most common score | Number of calls used |
Build the table in this order:
- Choose one filter for the period and write it at the top of the table.
- Calculate the four numbers for each rep using only that filter.
- Read each row before reading the column. Ask whether the median sits near the average, whether the mode falls in the same region, and whether the count is large enough to trust.
- Only after reading the rows, look at how reps compare with each other.
The order matters. If you sort by average first, the average becomes the whole story and the other values become footnotes. If you read the row first, the average becomes one part of a picture that also includes spread and the amount of evidence.
Keep the table easy to update each period. The goal is to watch the same rep change over time, using the same filter, so you can tell whether coaching is making a difference. A single snapshot shows where a rep stands. A series of snapshots shows whether something is moving.
Illustrative example: matching centers can hide spread
Suppose a rep has illustrative scores of 70, 70, 90, and 90. The average is 80, and the median is 80. The modes are 70 and 90, with a count of 4. The matching average and median do not mean calls scored near 80: none did. Review calls from both ends before choosing a coaching focus.
Use the three values for coaching before ranking by average
Use the three values to choose one coaching focus for each rep, not to rank the team on a single number.
A coaching conversation works best when it is specific. Pick one behavior the rubric measures, such as how a rep opens a call, how they ask about budget, or how they set the next step. Use the median and the mode to find out whether that behavior is a repeated gap or a single bad day. Then listen to two or three calls together with the rep and talk about what you both hear. Keep timestamped call notes so you can return to the moments you discussed.
After the conversation, set a short follow-up. Ask the rep to focus on the chosen behavior for a set period, then score a new set of calls with the same rubric and filter. Compare the new row with the old one. If the median or the mode has moved on the behavior you chose, that may suggest the coaching helped. If nothing moved, that is useful too, and you can try a different approach. Use coaching follow-up tasks to carry the chosen focus into the next review.
Be honest with reps about what the numbers can and cannot show. An average summarizes scored calls. It does not capture everything a rep does, such as building relationships with a champion inside a prospect's company or work that happens between calls. The table tells you where to start the conversation, and the rep's own account of the calls is part of the evidence.
Calibrate your scoring before you compare reps
Calibrate your scoring so reps are measured against the same standard, because a rep's average is only as consistent as the scores behind it.
Calibration means that you, and anyone else who scores calls, review the same recordings and discuss where your scores differ. Start with a few calls that each person scores independently. Compare the results, talk through the criteria where you disagreed, and write down the interpretation you agreed on. Repeat the exercise periodically, especially after the rubric changes or a new scorer joins.
Calibration matters for the three values in different ways. A difference between scorers can shift an average, move where the median falls, and change the mode. If two reps are scored by different people, their rows may reflect the scorers as much as the reps. Keep track of who scored which calls so you can check for that.
You do not need a perfect process to get value. A short monthly session can reduce the most obvious disagreements. Keep a simple note of the decisions you made, so the next scorer can apply them the same way.
Start with the table in your next review
Start by adding the median, mode, and count to your next review, and keep the average where it is.
You do not need new software or a new scoring process. You need one written filter, one table with one row for each rep, and a habit of reading each row before ranking the team. Much of the value comes from reading the values together, not from the calculation itself.
Over a few review cycles, the table will show which reps have a broad pattern, which reps have a few outlier calls, and which reps you simply have not scored enough to judge. That information is far more useful than an average alone. It also gives you a better conversation with reps, with your own leaders, and with yourself about what the scores mean.
The average will still be on the page. Now it will sit beside the values that tell you whether to trust it.
Today, choose one rep and calculate the median, mode, and count for the calls behind their current average.
Related guides: Log call feedback for every rep: 4 vendor rules to check · Automated call scoring rules from 4 publishers · Call handling: 9 rules to check before coaching.
Questions about comparing call scores
Does a matching average and median show consistency?
No; matching centers can hide scores spread far apart. Check the individual scores before describing the rep as consistent.
What should you do when scores tie for the mode?
Report every score tied for most common. Keep the tie visible when you review the calls.
What comes first when the count is too small?
Score more calls before making a decision about the rep. Use the minimum count you chose for the review period.