How do you coach SDRs without becoming the bottleneck?
Most managers sit at one of two extremes: almost no inspection, or constant interruption. 10-2-1 is the loop in between, enough evidence to diagnose, few enough changes to actually stick.
Original SD Lab framework · expanded for SD Lab V2
What this guide covers
- Why too little inspection and too much interruption produce the same result: reps who cannot self-correct.
- The 10-2-1 loop: 10 calls, 2 questions, 1 behavior.
- The operating cycle: EXECUTE → INSPECT → QUESTION → COACH → APPLY → REPEAT.
- What to write down after every session so the next one starts with evidence.
- Why one behavior at a time is the whole point, not a limitation.
Two failure modes, one outcome
The first failure mode is absence. The manager reviews totals, runs a pipeline meeting, and never listens to a full conversation. Coaching becomes whatever the loudest problem of the day happens to be.
The second is interruption. Every call gets a note, every message gets a rewrite, every objection gets a correction in the moment. The rep learns to wait for instruction rather than read the conversation.
Both produce the same rep: someone who executes what they were last told and cannot diagnose their own calls.
If the rep cannot tell you what went wrong before you do, you are not coaching. You are steering.
10-2-1
Ten is a practical operating rule, not a statistically validated sample. The point is that one bad call is an anecdote and a week of totals is a rumour, ten conversations is usually enough to see whether something is a pattern.
- 10: listen to ten calls before forming a diagnosis. Not ten minutes of ten calls; ten conversations end to end where possible.
- 2: bring two specific questions about what you heard, not a verdict.
- 1: leave with one observable behavior the rep will change on the next block of calls.
Three changes at once means you will not know which one worked, and the next session becomes guesswork again. One behavior, observable, checkable next week.
EXECUTE → INSPECT → QUESTION → COACH → APPLY → REPEAT
- EXECUTE: the rep works a defined block with a defined list. Nothing is coached mid-block.
- INSPECT: the manager reviews ten conversations against a known standard, taking timestamps rather than impressions.
- QUESTION: the session opens with two questions: what did you hear there, and what would you do differently?
- COACH: name the single behavior, show the moment it appeared, and model the alternative once.
- APPLY: the rep runs the next block with that one behavior as the focus.
- REPEAT: the following session starts by inspecting whether the behavior changed, before anything new is added.
Questions do more work than corrections
A correction transfers your judgment. A question builds theirs. The two questions should be anchored to a specific moment. A timestamp, not a theme.
Good pairs sound like: 'At 1:40 they said they already had a vendor. What did you hear in that?' and 'If you ran that thirty seconds again, what changes?' The rep almost always finds it. When they do not, that is your diagnosis.
- Anchor to a timestamp, not a general impression.
- Ask what they heard before telling them what you heard.
- If the rep cannot name the moment, the gap is listening, not scripting.
Observable means someone else could score it
'Be more curious' is not a behavior. 'Ask one clarifying question before responding to the first objection' is. If a second person could listen to the next block and tell you whether it happened, it qualifies.
- Name it in the rep's words, not the framework's.
- Write it down where both of you will see it before the next block.
- Check it first in the next session, before adding anything new.
Where the loop sits in the week
10-2-1 is the HOW of coaching a rep once you have inspected their work. It sits inside a wider inspection rhythm, daily execution visibility, weekly call review, monthly conversion review, which decides WHEN and WHAT gets looked at in the first place.
- Weekly per rep is the sustainable default for most teams.
- The session is thirty minutes because the inspection already happened.
- Missing a week is fine. Skipping the inspection and keeping the meeting is not.
Where structured inspection showed up in live engagements
Operational context only. Read each report for what its evidence actually supports, none of these outcomes are attributed to this framework alone.
- Kustomer: coaching and conversation quality inside an existing team.
- ServiceChannel: behavior and process change across an inherited SDR team.
- FusionAuth: execution visibility across a live outbound motion.
Related reading
Stop coaching the loudest problem of the day. Build a coaching system.
The Revenue Rebuild installs the inspection rhythm and coaching loop into a live team, with real calls on the table rather than a framework on a page.