Sales Process vs. Customer Reality

FAQ

What does the Process-Reality Gap (PRG) index measure and at what value is a stage "fiction"? PRG is computed per stage as the max of three components: PRG_time (gap between documented and actual median time-in-stage), PRG_conv (gap between documented and actual conversion rate), and PRG_value (gap in average deal value at stage).

Below 0.15 the process matches reality and needs only a quarterly audit; 0.15-0.30 is drift requiring adjusted definitions and retraining; above 0.30 is fiction, and you must halt forecasting against that stage until it is rebuilt.

Can you walk through the worked PRG example in the article? The playbook says Discovery to Scoping takes 14 days, but the CRM median is 31 days, giving PRG_time of 0.55. Documented conversion is 70% while actual is 41%, giving PRG_conv of 0.29. The stage PRG is the max, 0.55, which is fiction, meaning either reps are parking deals to look busy or buyers are not ready on that timeline, and you cannot forecast against it.

What minimum sample sizes does the audit require for pipeline and for win/loss interviews? The stage-reality audit needs at least 80 closed deals (N>=80) from 6-12 months of pipeline, because below that per-stage variance dominates the signal. The buyer-signal interviews require 5+ wins, 5+ losses, and at least 3 no-decisions; skipping the no-decision bucket loses 50% of the signal.

The Primary Intelligence opener is "Walk me through the real timeline, not the one in our CRM."

What healthy stage conversion benchmarks does the article cite? Per OpenView and Korn Ferry / CSO Insights, healthy reverse-funnel benchmarks are 60-80% for Qualified to Proposal, 50-70% for Proposal to Negotiation, and 65-85% for Negotiation to Close. Forrester's 2025 buyer study adds that 68% of B2B journeys are non-linear, so the process must support parallel tracks like evaluation, procurement, and security review running concurrently.

What CRM-hygiene threshold signals the time-in-stage data is unusable? If fewer than 70% of stage transitions carry a same-day timestamp, reps are backfilling stages at quarter-end and the time-in-stage data is fictitious. Another contamination check is segmenting by month: any 30-day window whose median diverges more than 2 sigma from the rolling median indicates the sample includes an event like a pricing change, product launch, layoff, or competitor exit.

The fix is to run two separate audits, pre and post event, and never blend them.

Bottom Line Up Front

If the median actual time-in-stage in your CRM differs from your documented playbook stage by more than 30%, your process is fiction and your forecast is built on it. The instrument is a quarterly four-lens audit producing a Process-Reality Gap (PRG) index per stage. Anything PRG > 0.30 must be rebuilt before next forecast call.

Gartner's B2B buying journey research finds the median enterprise buyer touches 6.8 stakeholders and spends only 17% of total decision time with sellers — a process assuming linear, seller-led progression is structurally mis-calibrated before the first deal lands.

SUBAGENT_VERIFIED.

The PRG Index (Math)

For each stage transition compute three components, then take the max:

• PRG_time = | (documented median time-in-stage − actual median) / actual |
• PRG_conv = | documented conversion rate − actual conversion rate |
• PRG_value = | documented average deal-value-at-stage − actual | / actual (revenue-at-risk lens)

PRG_stage = max(PRG_time, PRG_conv, PRG_value)

Interpretation:

• < 0.15 → process matches reality. Quarterly audit.
• 0.15–0.30 → drift. Adjust stage definitions and retrain.
• > 0.30 → fiction. Halt forecasting against this stage until rebuilt.

*Worked example*: Playbook says 14-day Discovery → Scoping. CRM median is 31 days. Conversion is documented as 70%, actual is 41%. PRG_time = 0.55, PRG_conv = 0.29, PRG_stage = 0.55 → fiction. Either reps are parking deals to look busy, or buyers aren't ready on that timeline. You cannot forecast against this.

The Operator's View

McKinsey's 2024 B2B Pulse finds top-quartile sellers re-baseline stage definitions every 90 days; the rest re-baseline only after a missed quarter. Bain's commercial excellence diagnostic recommends a buyer-mirror review every two quarters.

Salesforce's State of Sales 2024 reports only 28% of reps say their process reflects how customers actually buy. Deloitte's commercial transformation index puts the cost of process-reality misalignment at 6–11% of annual revenue.

1. Audit Stage Reality

Pull 6–12 months of pipeline with N≥80 closed deals (below this, per-stage variance dominates signal — see /knowledge/q05 on minimum sample sizes). Compute median, p90, conversion rate, per-rep variance. Flag stages where p90/median > 4× — undefined stage where reps improvise.

See /knowledge/q44 for stage-gate exit criteria.

*Salesforce SQL sketch*: ``sql SELECT StageName, PERCENTILE_CONT(0.5) WITHIN GROUP (ORDER BY DaysInStage) AS median_days, PERCENTILE_CONT(0.9) WITHIN GROUP (ORDER BY DaysInStage) AS p90_days, COUNT(*) AS n FROM OpportunityHistory WHERE CloseDate >= DATEADD(month, -12, GETDATE()) GROUP BY StageName; ``

2. Map Buyer Signals

Conduct win/loss/no-decision interviews (5+ wins, 5+ losses, 3+ no-decisions). Skipping the third bucket loses 50% of the signal. Use the Primary Intelligence opener: "Walk me through the real timeline, not the one in our CRM." Sense-check against HBR's new B2B sales playbook.

See /knowledge/q88 for the interview script and /knowledge/q400 on no-decision program design.

3. Track Conversion Math

Reverse-funnel from closed-won. Healthy benchmarks (per OpenView's SaaS benchmarks and Korn Ferry / CSO Insights):

• Qualified → Proposal: 60–80%
• Proposal → Negotiation: 50–70%
• Negotiation → Close: 65–85%

See /knowledge/q12 on funnel math and /knowledge/q210 on no-decision rate as truest fit signal. Forecast accuracy methodology in /knowledge/q301.

4. Measure Predictability

Run Sandler-style reversals in weekly 1:1s. Track time-in-stage by deal size, source, industry. Use MEDDPICC or Challenger to audit rep questions.

Forrester's 2025 buyer study confirms 68% of B2B journeys are non-linear — your process must support parallel tracks (eval + procurement + security review running concurrently).

5. Check Behavior Alignment

Survey reps anonymously: "Rate process adherence 1–5." If <3.5 avg, process is theater. See /knowledge/q150 on CRM adoption diagnostics — low stage updates are a credibility issue, not data-entry laziness.

Bear Case: When This Audit Fails (and How to Recover)

Six failure modes operators repeatedly hit, with detection thresholds and concrete remediation:

1. Sample contamination. Window includes pricing change, product launch, layoff, or competitor exit. *Detect:* segment by month; any 30-day window with median diverging >2σ from rolling indicates contamination. *Mitigate:* run two audits (pre/post-event); never blend.
2. CRM hygiene rot. Reps backfill stages at quarter-end so time-in-stage is fictitious. *Detect:* if <70% of transitions have a same-day timestamp, data is unusable. *Mitigate:* in HubSpot, lock backdating via Pipeline Settings → Stage History permissions; in Salesforce, enable Stage History audit and require forward-only logging via validation rules.
3. Survivorship bias in win/loss. You interview only closed deals — but 40–60% go dark. The no-decision cohort holds the real signal. *Mitigate:* mandatory no-decision interview track; pay an external firm to reach ghosted contacts who won't take your call.
4. Champion capture. Win interviews dominated by your champion confirming the narrative. *Mitigate:* require one non-champion stakeholder per won deal (procurement, technical evaluator, end user).
5. Stage-name drift across segments. Enterprise and SMB reps both call stage 3 "Proposal" but mean different things. *Detect:* compute PRG by segment; if segment variance > overall variance, you have two processes labeled the same. *Mitigate:* split the playbook by segment.
6. Tool-induced false positives. Some CRMs auto-advance stages on activity (email open, demo booked) creating phantom velocity. *Detect:* compare manual vs. Automated stage moves; if >30% are automated, velocity numbers are inflated. *Mitigate:* disable auto-advance on revenue-critical stages; require manual confirmation with exit-criteria checklist.

Typical Mismatches

30/60/90 Instrumentation Plan

• Days 1–30: Export closed-won + lost + no-decision deals (N≥80). Calculate PRG per stage. Schedule 13 interviews (5 win, 5 loss, 3 no-decision).
• Days 31–60: Run interviews. Map actual buyer journey on a whiteboard. Side-by-side compare to playbook. Identify the 1–3 stages with PRG > 0.30.
• Days 61–90: Rewrite the broken stages to match buyer reality. Retrain reps with new exit criteria. Lock CRM stage logic. Set 90-day re-audit cadence on calendar.

Key insight: If your CRM doesn't reflect your playbook, one of them is fiction. Fix the playbook, not the CRM. SUBAGENT_VERIFIED.