How to Run Call-Shadowing So New Reps Learn Something Instead of Just Listening
Most shadowing programs are new reps taking notes on autopilot. A structured observation rubric turns the same hours into measurable skill transfer.
Most shadowing programmes consist of a new rep with a notepad, sitting through eleven calls in a row, nodding along. Ask what they learned afterwards and you get "seemed like a good call" or a paraphrase of whatever the closer said in the last two minutes. The hours logged are real. The learning mostly isn't — because watching an expert do something fluently is not the same skill as doing it yourself, and nobody has told the new rep what, specifically, to watch for.
Why passive shadowing doesn't transfer
A novice doesn't yet know what "good" looks like in component parts, so an expert call flows past them as a blur of general competence rather than a sequence of discrete decisions. Without something to score against, new reps pattern-match on the surface features they can actually perceive — tone, energy, a good joke at the right moment — instead of the substance underneath: how the qualifying question was framed, how the objection was reframed rather than argued with, whether the next step was actually secured or just implied. The shadowing hours accumulate. The skill doesn't.
The fix: score before the call, score after it
The mechanism that actually moves a new rep from spectator to learner is prediction and comparison, not observation on its own.
Before the call. Give the new rep the account context five minutes ahead of time. Have them predict, against a short rubric, what the experienced rep will do at three or four checkpoints — the opening, the depth of discovery, the first objection response, the close or next-step ask — and score their own confidence in each prediction, one to five. This step is the one almost every shadowing programme skips, and it's the one doing most of the work: a prediction that turns out wrong is a far stronger teaching moment than an observation that simply happened.
During the call. The new rep takes notes against the same rubric categories, not free-form notes. Structure forces attention onto the substance rather than the personality of whoever's on the call.
After the call. Score what actually happened against the same rubric, then compare it directly to the prediction. The debrief question isn't "how was that call" — it's "where was your prediction wrong, and what does that tell you." A rep who predicted a price objection would be met with a discount and watched it get met with social proof instead has learned something real about reframing versus conceding, in a way that "take good notes" never produces.
The debrief. Ten minutes, focused only on the single biggest delta between prediction and reality — not a general retrospective on the whole call. A GROW Model Coaching Conversation Script gives that ten minutes a shape rather than letting it drift into "so what did you think." Anchor the rubric itself to something durable rather than reinventing it per manager — a Call Shadowing Checklist for New Reps covers the general structure, and a Discovery Call Scorecard is worth pulling in specifically for the discovery checkpoint, since that's the stage most rubrics score too loosely.
The rubric, compressed
| Checkpoint | What to score | Question the new rep asks themselves |
|---|---|---|
| Opening | Did they earn attention or just start talking | Would I have known where to take this in the first 30 seconds? |
| Discovery | Depth, not just breadth, of questioning | Did that question get a real answer or a polite one? |
| Objection handling | Reframe vs. concede vs. argue | Did they change the buyer's frame, or just push back on it? |
| Close / next step | Specific commitment secured, not implied | Is there an actual date and action, or a vague "sounds good"? |
How many calls before it stops paying off
Six to eight, provided every one of them is scored rather than merely watched. That's the point at which a new rep has extracted most of what a single type of call has to teach them — the checkpoints stop surprising them, the predictions start landing correctly, and the marginal call adds less than the one before it. Running fifteen shadowed calls of the same call type past that point is close to wasted time; you're better off running six to eight structured ones spread across genuinely different scenarios — discovery-heavy, negotiation-heavy, a renewal, a multi-stakeholder deal — than doubling the count within one scenario type.
Past roughly call eight, flip the model rather than extending it. The new rep should be the one on the call, with the mentor scoring them against the same rubric — reverse shadowing. Passive observation has hit its ceiling by then; the remaining skill transfer needs the rep in the seat, making the actual decisions the rubric has been training them to notice. A Sales Buddy/Mentor Program Playbook-style structure works well for this handoff, because it's built around a named mentor relationship rather than a rotating cast of whoever's free to be shadowed that week.
The hours were never the scarce resource here. Attention with a target was. Give a new rep something specific to be wrong about before the call, and eleven hours of half-remembered listening turns into six or eight calls' worth of skill they can actually name.