Building Sales Coaching Scorecards Your Reps Won't Ignore

Win rate is an outcome. It's also the wrong thing to coach directly, because you can't debug a number, only a behavior. Average new-logo win rates run around 19%¹, which gives you a useful floor to compare a rep against, but the scorecard itself should track things a manager can actually observe and coach: whether discovery calls surfaced a clear budget and timeline, whether the next step after each call was scheduled before the call ended, or whether a mutual close plan existed before a deal reached the final stage.

Each behavior on the list should answer one question: if a rep got noticeably better at just this one thing, would deals move faster or die less often. If the answer is unclear, the behavior doesn't belong on the scorecard yet.

Behaviors worth scoring are specific and observable, such as:

• Whether the discovery call surfaced a clear budget and timeline before it ended.
• Whether the next step was scheduled before the call ended.
• Whether a mutual close plan existed before the deal reached its final stage.
• Whether getting noticeably better at that one behavior would plausibly make deals move faster or die less often; if not, it doesn't belong yet.

A custom field set on the deal or activity record, scored on each qualifying call or milestone, works better than a separate spreadsheet because it stays attached to the deal it's rating. In Close, this can hang off the call logging workflow directly, since call recordings and notes already live there; in Pipedrive, it fits as a checklist field triggered at specific pipeline stages, so a manager scoring a discovery call sees it right alongside the deal's other context.

Either way, the score should be entered by the manager during a live or recorded call review, not self-reported by the rep, or it becomes a compliance exercise instead of a coaching one.

A scorecard reviewed quarterly is a performance review, not coaching. Coaching happens weekly, in short 1:1s where the manager and rep look at two or three recent scored calls together and agree on one specific thing to try differently on the next one. The scorecard's job is to make that conversation concrete instead of a general "keep pushing" pep talk that leaves the rep with nothing specific to change on Monday morning.

Keep the review short on purpose. Fifteen focused minutes on two calls beats an hour spent scrolling through a dashboard neither person prepared for.

For example, a manager reviewing two recent discovery calls with a rep notices that neither one confirmed the decision-maker before the call ended. Instead of a general note to qualify harder, they agree on one change: ask who else signs off on the purchase before hanging up. The following week they score the next two calls on that single behavior only. If the score improves, they add a second behavior. If it stays flat, they change the intervention, such as a role-play session, instead of repeating the same advice. Each review ends with one specific change, which is easier for a rep to remember and easier for a manager to verify.

If a rep's scores on a specific behavior stay flat for several weeks despite direct coaching, that's a signal to change the intervention, not repeat it louder. Sometimes the fix is a role-play session away from live calls, where the pressure of a real prospect isn't clouding the feedback; sometimes it's pairing the rep with a peer who scores well on that exact behavior for a joint call. Continuing the same weekly note with no change in approach is how a scorecard becomes background noise the rep learns to tune out entirely.

Once a few months of scored calls accumulate, look across reps for patterns rather than just within one rep's trend line. If most of the team scores low on the same specific behavior, that's not five individual coaching problems, it's a training gap or a process gap worth addressing once for everyone, whether that's a short workshop, an updated call framework, or removing a step in the process that's making the behavior harder than it needs to be.