How do you measure whether call coaching works?

What should you measure?

Measure the behaviour you coached, not just the overall result. If you coached "confirm the customer's need before pitching", the first thing to track is how often reps confirm the need. That moves within weeks. Revenue moves too, but more slowly and for many more reasons.

That's how the research behind a well-known Harvard Business Review article measured it too: its authors defined performance as "a rep's gap to goal (i.e., percentage of quota attained)" (Dixon and Adamson, Harvard Business Review, 31 Jan 2011). Track both: the behaviour first, then the outcome.

Why isn't a before-and-after enough?

Because of regression to the mean. A paper in CBE—Life Sciences Education describes it as "a statistical phenomenon that creates patterns of change by chance alone": when you measure twice, extreme results the first time tend to sit closer to the middle the second time. It notes that, by chance alone, a large increase is expected in the lowest-scoring group and a large decrease in the highest (Furrow, CBE—Life Sciences Education, Jun 2019).

Coaching usually starts with the reps who scored lowest last month. Their scores would likely rise next month even if nobody spoke to them. So a before-and-after on those reps will flatter your coaching.

Coached reps, before48%

Coached reps, after71%

Not yet coached, before50%

Not yet coached, after58%

How do you set up a fair comparison?

Compare coached reps with similar reps who haven't been coached on that item yet. The same paper says that with matching between a control group and an intervention group, "one can observe whether an effect is larger for the intervention group" (Furrow, CBE—Life Sciences Education, Jun 2019). Another paper lists the main ways to control for the effect: account for the starting score, match the groups on their starting scores, or use a control group (Yu and Chen, Frontiers in Psychology, Jan 2015).

1. 1

   Pick one item

   The behaviour you're about to coach, from your scorecard.

2. 2

   Get a baseline

   Score it on every rep's calls for a few weeks before you start.

3. 3

   Coach half first

   Pair reps with similar baseline scores, and coach one from each pair. The other waits a few weeks.

4. 4

   Score it again

   Same item, same scorecard, same number of weeks.

5. 5

   Compare the change

   The difference between the two groups' improvement is roughly what the coaching did.

Coaching the second group a few weeks later means nobody misses out; you just stagger the start.

How many calls do you need?

More than most teams review. With only a few calls per rep, a score can swing a long way by chance: a rep who passes an item on 3 of 4 calls might really pass it anywhere from about 30% to 95% of the time (95% Wilson interval, our calculation using the method in the NIST/SEMATECH e-Handbook, as of 15 Oct 2026). Small changes are invisible at that size. See how many calls to review per rep.

When should you look at sales results?

After the behaviour has changed. If the coached behaviour went up and sales didn't move over the following months, you may be coaching the wrong thing. That's useful to know too: swap the item and test again.

How does Callyst help with this?

Callyst scores every call on your scorecard, so you get each rep's score on each item from all their calls, not a handful. That makes a baseline and a comparison group practical without extra listening time.