Sales Coaching at Scale: Why Your Best Closer Can’t Teach What They Do
A SaaS sales organization we workedwith had a star closer. She converted demos to deals at nearly twice the teamaverage, consistently, across quarters. Leadership did the obvious thing: theyhad her run training sessions for the rest of the team. Six months later, theteam average hadn’t moved. Her own numbers were still excellent. The traininghad transferred nothing.
We analyzed her actual call recordingsagainst the team’s. The reason the training failed became obvious. What shetaught in the sessions — her stated methodology — bore little resemblance towhat she actually did on calls. She believed she won deals through productknowledge and thorough demos. The recordings showed she won them through aspecific discovery pattern in the first ten minutes that surfaced thecustomer’s real decision criteria before she ever showed a feature. She wasn’tteaching this because she didn’t know she did it. It was unconsciouscompetence.
This is the central problem with salescoaching: your best performers are usually the worst at explaining their ownsuccess. The technique that drives their results is automatic, invisible tothem, and therefore absent from anything they teach. The conversation data hasit. Their self-report doesn’t.
Traditional sales coaching is built almost entirely on self-reportand observation. Managers ask reps what happened on calls. Reps describe whatthey remember, filtered through what they believe matters. Top performersdescribe a version of their process that’s plausible, coherent, and frequentlywrong about what actually drives their results.
This isn’t dishonesty. It’s how expertise works. When someonebecomes highly skilled at something, the skill becomes automatic and drops outof conscious awareness. Ask an expert driver how they handle a difficult mergeand they’ll give you a simplified, partly inaccurate account, because theactual skill is running below the level they can introspect on.
The consequence for sales coaching is severe. The organization’sbest knowledge about what works is locked inside performers who can’taccurately report it, and the coaching system is designed to extract thatknowledge through exactly the channel — self-report — that can’t access it.
Conversation analytics changes salescoaching by replacing self-report with observation of what actually happened.
The patterns that separate top from average performers becomevisible and specific. Discovery depth, talk ratio, question types, objectionresponses, the timing of when they introduce price, the language they use atthe close. These are measurable across hundreds of calls, and the differencesbetween performance tiers are usually concentrated in two or three specificbehaviors rather than spread across a vague notion of “skill.”
The coaching artifact becomes real. Instead of “be moreconsultative” — advice that means nothing actionable — the coach can show therep three of their own calls where a shallow discovery question led to a lostdeal, alongside three of a top performer’s calls where a deeper questionchanged the trajectory. The specificity makes the coaching land.
The teaching library becomes evidence-based. The behaviors thatactually correlate with conversion in your specific sales motion, with yourspecific customers, become the curriculum. Not a generic methodology from asales book. Your own winning patterns, extracted from your own data.
Sales coaching has always faced an impossible math problem. Amanager with eight reps can’t listen to enough of each rep’s calls to coachthem well. Even one call per rep per week is eight hours of listening beforeany coaching happens, and one call per week is far too small a sample to seereal patterns.
The result is that most sales coaching is based on a handful ofcherry-picked or random calls, which is statistically inadequate to identifywhat’s actually driving each rep’s results. Managers coach on impressionsrather than patterns, and the impressions are biased toward recent andmemorable calls rather than representative ones.
Analytics across 100% of calls removes the sampling constraint. Themanager can see each rep’s actual patterns across all their calls, identify thespecific behaviors limiting their conversion, and coach on evidence rather thanimpression. The eight hours of listening becomes eight minutes of reviewingflagged patterns.
1. Compare your top and bottomperformer on one specific behavior. Pick talk ratioor discovery depth. Measure it across 10 calls each. The gap will likely belarger and more specific than you expected.
2. Ask your best closer to explaintheir process, then check it against their calls. Thegap between what they say and what they do is the unconscious competence youneed to extract.
3. Identify the two behaviors thatcorrelate most with conversion. Across your team,which measurable call behaviors predict deals? Those are your coachingpriorities.
4. Build one coaching session aroundreal call clips. Replace methodology slides withactual recordings — winning and losing — from your own team. Watch howdifferently reps engage.
5. Stop coaching on self-reported callnotes. They’re a record of what the rep believeshappened, which is systematically different from what did. Coach on therecording.
Your best closer is a goldmine of salesknowledge that even they can’t access. The technique that drives their resultsis automatic, invisible, and absent from everything they’d teach you. The onlyplace it exists in retrievable form is the conversation data — and that’s theone place most sales coaching never looks.
The AI terminology chaos is real. Your "divide and conquer" framework is the clarity we needed.
Finally, a clear way to cut through the AI hype. It's not about the name, but the problem it solves.
