Coaching from evidence, not ride-alongs

A ride-along coaches one call. Your team makes thousands. Whatever you learn sitting in on a Tuesday afternoon is a sample of one rep, on one day, in a mood they know is being watched.

The sampling problem nobody names

Call coaching has a measurement problem that everyone works around and almost nobody states plainly. A manager can sit through perhaps three or four calls a week with any real attention. A team of eight reps making forty dials a day generates sixteen hundred calls in the same week. The coaching sample is a quarter of one percent, it is not random, and it is biased in the least useful direction: managers sit in on the reps they are already worried about, on the accounts that are already important, at times of day that suit a calendar.

So the feedback loop closes on the wrong evidence. A rep hears about the call the manager happened to watch, not the pattern across the two hundred they made that month. The pattern is where the coachable behaviour lives.

What changes when every call is scored

When scoring runs across every call rather than a sample, the unit of coaching changes from the anecdote to the distribution. You stop saying "on that call you talked too much" and start saying "across sixty discovery calls your talk time is sixty-eight percent, and on the eleven that advanced to a next step it was forty-one." The second sentence is coachable because it carries its own evidence and its own counterfactual.

The scoring itself is unglamorous and that is the point: did the rep set an explicit next step with a date, did the stated objection get a response or a deflection, how long was the longest monologue, was a decision-maker named. These are observable events in a transcript, not judgements about tone. Anything requiring taste stays with the manager, where it belongs.

Coach from the tape, not the recap. The tape does not round up, and it does not remember the call more kindly than it went.

Why the clip matters more than the score

A score tells a rep where they stand. It does not tell them what to do differently, and a number without a moment attached is the fastest way to make coaching feel like surveillance. The useful artifact is the thirty seconds either side of the moment the call turned: the objection that landed, the pause that was filled too quickly, the question that opened the account up.

That is why the clip is the deliverable and the score is only the index. A manager reviewing a queue of moments can coach eight reps in the time it used to take to sit through two calls, and every conversation starts from the same recording both people can hear.

The honest limits

Automated scoring is good at the countable and poor at the contextual. It can tell you a next step was not set. It cannot tell you the rep was right not to push because the buyer had just mentioned a bereavement. It will mark a short call as low-engagement when it was a decisive qualifying out, which is a good outcome badly measured.

So the scores are inputs to a conversation, not a verdict, and they are visible to the rep as well as the manager. A coaching system a rep cannot inspect is a performance-management system wearing a friendlier label, and reps work out the difference within a week.

The gain is not that a machine coaches better than a good manager. It is that a good manager stops spending their scarcest hours deciding which calls to listen to, and starts every conversation already knowing which thirty seconds matter.

What a manager actually does with this on a Monday

The practical shape of the week changes more than the technology suggests. Instead of blocking two hours to sit in on calls, a manager opens a queue of moments already sorted by what they show: the objection that went unanswered three times this week, the two reps whose discovery calls never surface a second stakeholder, the deal where the next step has slipped twice without anyone saying so out loud.

Each of those is a five-minute conversation with a recording attached, and the recording ends the argument about what was said. Most coaching disputes are not disagreements about technique. They are disagreements about what happened, and those disappear when both people are listening to the same thirty seconds.

The second-order effect is that coaching stops being an event. A rep who gets three short, specific notes a week improves faster than one who gets a thorough review each quarter, for the same reason that frequent small corrections beat infrequent large ones in any feedback system.

The metric that tells you it is working

Watch the variance, not the average. A team where the best rep converts at four times the worst rep has a coaching problem, not a talent problem, and the useful signal is whether that spread narrows. Averages improve when your strongest performer has a good month, which tells you nothing about whether anybody learned anything.

Expect the spread to widen slightly before it narrows. Making behaviour visible surfaces problems that were previously invisible, and the first month of honest measurement usually looks worse than the last month of comfortable ignorance.

Start with one behaviour, not a scorecard

The failure mode when teams adopt this is measuring everything at once. A twelve-point scorecard produces twelve mediocre conversations and no behaviour change, because a rep cannot hold twelve intentions in a live call. Pick the single behaviour with the clearest link to advancement, usually setting an explicit next step with a date, and coach only that until it stops being a coaching topic.

Then move to the next one. This is slower than it sounds and faster than the alternative, because a behaviour that has genuinely landed keeps paying out while you work on the following one, whereas twelve simultaneous priorities decay together the moment the quarter gets busy.

The evidence makes this sequencing possible in a way it never was before. You can see whether the behaviour actually moved across two hundred calls rather than guessing from the handful you happened to hear, which means you find out you were wrong about a coaching priority in a fortnight instead of at the end of a quarter.

• Playbooks
• Coaching
• Analytics