How do I diagnose whether a rep needs coaching, training, or termination?

The Diagnostic Framework

Firing the wrong rep costs $80K-$360K in turnover + ramp loss (Bridge Group 2024). Keeping the wrong rep costs $200K+ in unrealized quota plus the morale tax - Gallup State of the Workplace 2024 finds one disengaged rep depresses adjacent rep productivity by 11-14%.

CSO Insights (Korn Ferry 2023 World-Class Sales Practices Study) reports that organizations who run formal performance diagnostics before action retain 17 percentage points more high-performers vs. Those who fire reactively. Gartner's 2024 Chief Sales Officer Survey (G00800431) found 64% of CSOs believe their managers are under-equipped to diagnose performance issues - the data is there but the framework is not.

Get the diagnosis right.

STEP 0: CLEAN THE DATA FIRST Before you even open the dashboard:

• Pull dialer logs directly (Aircall, Dialpad, RingCentral, Five9) - NOT CRM-logged activity. CRM-logged calls are gameable; dialer logs are not.
• De-dupe meetings: Outreach/Salesloft sometimes double-counts auto-rescheduled events.
• Confirm quota was set correctly. If TAM coverage was off by >20%, you are diagnosing a quota error, not a rep error.
• Pull 30 days minimum, ideally 90, to filter noise.
• CALIBRATION SANITY CHECK: Before you trust the percentiles, plot a histogram of team activity. If the distribution is bimodal or has fat tails, percentile thresholds are misleading - use median and IQR instead.
• MULTI-RATER CALIBRATION: Have 2-3 managers independently score the same 5 rep calls using the same rubric. Cohen's kappa should be >=0.6 (substantial agreement); if lower, your rubric is the problem, not the rep.
• WILSON SCORE INTERVAL FOR SMALL SAMPLES: With <30 closed deals, point estimates of conversion rate are unreliable. Use a Wilson score 95% CI: a rep at 12% close rate on 25 deals could plausibly be anywhere from 5% to 27%. Don't fire on noise.

STEP 1: ACTIVITY (the foundation)

• Compute team 30th, 50th, 75th percentiles on dials, held meetings, and outbound emails.
• Below 30th percentile = coachable (effort/behavior).
• 50th+ = activity is fine; the leak is downstream - go to Step 2.
• Per Salesforce State of Sales 2024, top-quartile AEs spend 28% of their week selling vs. 23% bottom-quartile - a 22% relative gap that compounds over a quarter.

WORKED NUMERIC EXAMPLE - PERCENTILE MATH: Team of 10 AEs, last 30 days, dials/day: [22, 19, 17, 16, 15, 14, 12, 11, 9, 8]

• 75th percentile: 17 (top 3 reps)
• 50th percentile (median): 14.5
• 30th percentile: ~11
• Rep at 9 dials/day = below 30th percentile = activity coachable.
• Rep at 16 dials/day = above median = activity is fine, look at conversion.

Use the inclusive method (numpy.percentile or Excel PERCENTILE.INC) so all reps see consistent thresholds.

STEP 2: CONVERSION (the skill) Break the funnel into three stages:

• Meetings held -> opportunities (discovery quality)
• Opportunities -> late-stage (qualification, multi-threading, objection handling)
• Late-stage -> closed-won (negotiation, executive access, procurement)

ACTIVITY-TO-CONVERSION REGRESSION HEURISTIC: In most healthy teams, activity and conversion correlate at r approximately 0.35-0.55 (Bridge Group, ForceManagement field data). If a rep sits more than 1.5 standard deviations off the regression line, they are either gaming activity (high A, low C) or exceptionally efficient (low A, high C).

Both warrant a 1:1 - for opposite reasons.

GONG/CHORUS COACHING SCORECARD (use these dimensions):

• Talk-time ratio (rep speaks <50% of call time on discovery)
• Question count (>=4 open-ended questions per discovery)
• Next-step booking rate (>=70% of discovery calls end with calendar invite booked live)
• Multi-threading (>=2 stakeholders on call by stage 3)
• Pricing discussed before stage 4 (red flag - leads to discounting)

If scorecard shows training is the gap, the fix is calibration sessions weekly, not termination.

STEP 3: TERRITORY (the fit)

• Compare attainment to the prior rep in the same patch and to peers with comparable TAM.
• Two consecutive reps failing in the same patch = the patch is broken, not the rep. Per Gartner 2024 Sales Force Effectiveness, 22% of quota misses trace to territory design.
• Account fit audit: ICPs aligned to product? Sandbagged territories produce false negatives in your diagnosis.

STEP 4: TENURE (the runway)

• <3 months: Always coachable. Don't even diagnose; just ramp.
• 3-6 months: Coachable with structured plan. Bridge Group median ramp is 5.3 months.
• 6-12 months: Below 60% by month 9 = formal PIP. Below 40% = accelerate exit timeline.
• >18 months, no growth: Termination on the table. Coaching ROI flattens past 18 months tenure.

EXPECTED-VALUE DECISION MODEL (the math): Let OTE = $180K, quota = $1.2M, current attainment = 58%. Three options:

• KEEP AS-IS: EV = ($1.2M * 0.58) - ($180K) = $516K contribution. Probability of improvement without intervention ~20%.
• PIP + COACH: Cost = $15K (coach + 60 days runway). Probability of lifting to 80% = 35% (per Korn Ferry meta-analysis). EV = 0.35 * ($1.2M * 0.80 - $180K) + 0.65 * ($1.2M * 0.58 - $180K) - $15K = ~$591K.
• TERMINATE + REHIRE: Replacement cost = $300K (DePaul ratio). New rep at 0% during 5-month ramp. EV first 12 months = ($1.2M * 0.40 ramp average - $180K) - $300K = -$48K.

In this case PIP wins by ~$75K vs keep, and beats fire by ~$639K. Always run the EV math before deciding. If the rep is at 30% attainment and tenure 24mo, fire wins. If at 70% and tenure 12mo, keep wins.

LEGAL / HR RISK CHECKLIST (DO NOT SKIP): Before terminating, run through this with HR and counsel:

• ADEA (Age Discrimination in Employment Act): if rep is 40+, document objective performance criteria parallel to younger reps held to the same standard.
• ADA (Americans with Disabilities Act): if rep has disclosed a disability, ensure reasonable accommodations were offered and documented.
• FMLA: if rep took recent FMLA leave, the timing alone creates a retaliation claim risk - consult counsel on a 90-day cooling period.
• State law: California, NY, Massachusetts have additional protections; offer letter and PIP must comply with state notice requirements.
• Severance: offer severance in exchange for a release; standard is 2-4 weeks per year of service for non-executive AEs.

The single biggest cause of wrongful termination suits in sales: inconsistent application of standards. If two reps have similar metrics and only one is fired, document why or settle for $50K-$200K later.

REP-PERSPECTIVE TRANSPARENCY FRAME: The rep should be able to answer 3 questions at any point during the diagnostic:

1. What specific metrics am I being measured on?
2. Where do I currently stand vs. The bar?
3. What would I have to do, and by when, to clear the bar?

If the rep can't answer these, the manager has not communicated clearly - and any subsequent termination is procedurally weak. Document the conversation in writing, ideally email or a shared doc.

WHAT TO SAY IN THE 1:1 (verbatim opener): "I want to be direct with you about where you stand and what we're going to do about it. Your YTD attainment is 58% against an expected 80% by month 14. Activity is at 8 dials/day, which is below team 30th percentile of 11.

Discovery-to-close is 12% vs. Team median of 18%. I don't think this is a fit problem - I think it's a stacked activity-and-skill issue, and I think it's coachable.

Here's the 60-day plan I want us to run together. If at the end of 60 days dials are at 14+ and conversion is at 15%+, we're back on track. If not, we'll have a different conversation.

I want you to succeed here. What do you need from me?" Calm, specific, falsifiable, supportive. The opposite of a surprise.

REAL DIAGNOSIS EXAMPLE:

• Rep: 14mo tenure, 58% YTD quota, 8 dials/day (25th percentile), 12% discovery-to-close (team median 18%).
• Diagnosis: Stacked - activity AND conversion both below median. Not pure skill, not pure effort.
• Recommendation: 2-week MEDDPICC intensive PLUS daily morning huddle on activity. If dials don't reach 14/day and conversion doesn't lift to 15%+ within 30 days, formal PIP with 60-day clock.

PIP LETTER SKELETON (use, don't reinvent):

1. Specific gap: YTD attainment 58% vs. Expected 80% by month 14.
2. Specific behavioral targets: at least 14 dials/day, at least 6 discovery meetings/week, at least 15% discovery-to-close.
3. Support offered: 2hrs/week 1:1 coaching + MEDDPICC bootcamp budget.
4. Timeline: 60 days from [date] to meet all 3 targets.
5. Consequence: Failure to meet targets results in separation effective [date+60].
6. Sign + countersign + HR file.

BEAR CASE - when this framework fails you:

1. Garbage activity data. Reps gaming CRM logs makes percentile rankings fiction. Mitigation: dialer logs only.
2. Wrong quota. Over-allocated quota by 20%+ makes every rep look broken. Audit quota vs. TAM before diagnosing.
3. Manager bias. HBR 2023 found managers misclassify 30% of coaching cases as termination cases when they have not ridden along on 5+ calls in the prior quarter. Mandate 5-call shadow before any termination decision.
4. Macro headwinds. In a down quarter where 70% of the team is missing, individual diagnosis is meaningless - reset the baseline first.
5. Comp plan misalignment. If SPIFFs reward the wrong behaviors, you are punishing reps for following the comp plan you built.
6. Survivorship bias in your benchmarks. If you only kept top performers, your team-median is artificially high - any new hire looks bad. Calibrate against industry benchmarks (Bridge Group, Pavilion) too.
7. Comp-cycle timing risk. Firing a rep who is 60 days from a vesting cliff or accelerator threshold creates legal exposure under wage-and-hour claims AND morale damage. Consult Finance + Legal on timing.
8. Sample size too small. With <20 closed deals, conversion rate confidence intervals are wide enough to render the diagnosis meaningless. Use Wilson score interval or wait for more data.

RED FLAGS FOR TERMINATION (not fixable):

• Knows the skill but won't apply it (proven in role-plays, absent live).
• Resistant to feedback in three consecutive 1:1s.
• 18+ months flat with no improvement under structured intervention.
• Territory is controlled-for and rep is still bottom decile.
• Attitude/integrity issue (coaching cannot fix character).

THE TIMELINE:

• Day 1-7: Diagnostic only. Don't coach yet.
• Day 8-21: ONE intervention (training OR activity coaching). One variable at a time or you will never know what worked.
• Day 22-60: Measure. Did activity move? Did conversion move?
• Day 61+: Decide - promote, retain, PIP, or exit.

RELATED FRAMEWORKS ON PULSE:

• /knowledge/q47 - How to design a fair sales quota distribution (Step 3 prerequisite)
• /knowledge/q63 - MEDDPICC for sales coaching: a hands-on guide
• /knowledge/q92 - Building a 30-60-90 day PIP that actually works
• /knowledge/q104 - Sales rep ramp time benchmarks by segment
• /knowledge/q115 - Coaching vs managing: when to step in and when to step back
• /knowledge/q142 - How to audit CRM activity hygiene before any rep evaluation
• /knowledge/q176 - Gong call scorecard rubric for AE coaching
• /knowledge/q203 - Sales comp accelerators and cliffs: legal exposure on terminations

PRIMARY SOURCES:

• Bridge Group 2024 SaaS AE Metrics Report: https://blog.bridgegroupinc.com/saas-ae-metrics-2024
• Pavilion 2024 GTM Benchmark Report: https://www.joinpavilion.com/benchmarks
• Salesforce State of Sales 2024: https://www.salesforce.com/resources/research-reports/state-of-sales/
• Gartner Sales Force Effectiveness 2024 (G00800431): https://www.gartner.com/en/sales
• Gallup State of the Workplace 2024: https://www.gallup.com/workplace/349484/state-of-the-global-workplace.aspx
• Korn Ferry / CSO Insights World-Class Sales Practices Study: https://www.kornferry.com/insights/featured-topics/sales-transformation
• DePaul University Sales Effectiveness Center: https://business.depaul.edu/centers-institutes/sec/
• Harvard Business Review (manager bias): https://hbr.org/topic/subject/performance-management
• EEOC ADEA guidance: https://www.eeoc.gov/age-discrimination
• US DOL FMLA employer guide: https://www.dol.gov/agencies/whd/fmla/employer
• Cohen's kappa rater reliability: https://en.wikipedia.org/wiki/Cohen%27s_kappa
• Wilson score interval (small-sample CI): https://en.wikipedia.org/wiki/Binomial_proportion_confidence_interval#Wilson_score_interval

TAGS: performance-diagnosis, coaching-vs-training, termination, rep-evaluation, fit-assessment, pip, meddpicc, sales-management, gartner, gallup, salesforce-soss, bridge-group, pavilion, korn-ferry, ramp-time, quota-attainment, sales-coaching, hbr, adea, fmla, ada, gong, chorus, hr-compliance, ev-math, rater-reliability, wilson-score

FAQ

What four data points should I pull before deciding between coaching, training, or termination? Activity (dials and meetings versus team percentiles), conversion (funnel stage rates versus team median), territory (TAM/ICP fit plus prior-rep attainment), and tenure (ramp curve plus months in role).

Stack the answers rather than deciding on a single metric. Below 30th-percentile activity points to coaching; above-median activity with below-median conversion points to training.

Why pull dialer logs instead of CRM-logged activity? CRM-logged calls are gameable; dialer logs from Aircall, Dialpad, RingCentral, or Five9 are not. Step 0 is cleaning the data: de-dupe meetings since Outreach and Salesloft sometimes double-count auto-rescheduled events, confirm quota was set correctly (a TAM-coverage miss above 20% means you're diagnosing a quota error, not a rep error), and pull at least 30 days, ideally 90.

How does the activity-versus-conversion matrix map to a specific fix? Low activity with high conversion is a confidence or behavior gap, fixed with a 30-day activity coach on cadence and time-blocking. High activity with low conversion is a skill or training gap, fixed with a 3-week MEDDPICC or Sandler bootcamp.

High and high means promote, mentor, expand the patch. Low and low signals wrong fit, low motivation, or a ceiling, which goes to a PIP with crisp exit criteria.

Why use a Wilson score interval for reps with few closed deals? With fewer than 30 closed deals, point estimates of conversion rate are unreliable. A rep at a 12% close rate on 25 deals could plausibly sit anywhere from 5% to 27% under a Wilson 95% confidence interval, so you should not fire on noise.

The article also recommends multi-rater calibration, where 2-3 managers independently score the same 5 calls and Cohen's kappa should be at least 0.6.

What Gong or Chorus scorecard dimensions indicate a training gap? Talk-time ratio (rep speaks under 50% on discovery), question count (at least 4 open-ended questions per discovery), next-step booking rate (at least 70% of discovery calls end with a live calendar invite), multi-threading (at least 2 stakeholders by stage 3), and pricing not discussed before stage 4 (early pricing is a red flag that leads to discounting).

If the scorecard shows training is the gap, the fix is weekly calibration sessions, not termination.