Interactive tool
GTM Failure Mode Simulator
See how one broken stage propagates through a demand system.
Inject a constraint into a healthy funnel and trace what breaks downstream, what stays healthy, and where to intervene.
Scenario controls
One constraint at a time. Each scenario reduces the local rate of exactly one stage and leaves every other rate at baseline.
Severity
The affected stage retains 55% of its healthy local rate.
Clear the constraint and return every stage to baseline.
The baseline is an illustrative systems model, not benchmark data. It exists to make propagation legible, not to predict your numbers.
Routing breakdown, Material severity. Primary constraint: Lead routing. Final pipeline output 6.6 opportunity equivalents, −45.0% against the healthy baseline.
Primary diagnosis
Lead routing
Routing breakdown · Material · 95% → 52.3%
System effect
Submitted demand remains healthy, but substantially less of it reaches ownership. Every downstream stage receives less volume even though downstream local conversion performance is unchanged.
Pipeline impact
- This scenario
- 6.6
- Healthy baseline
- 12.0
- Throughput change
- −45.0%
Opportunity equivalents from 200,000 starting audience.
Different constraints can produce the same final number. That is why the last metric in the chain cannot tell you which stage failed.
Do not optimize yet
Paid media, landing-page conversion, or form completion. All three are performing at baseline.
Inspect next
- Assignment rules
- Ownership queues
- Routing eligibility
- CRM synchronization
- SLA coverage
Decision
Restore routing reliability before increasing acquisition volume.
System trace
Nine stages in sequence. Each one converts a share of what the stage above it passed down, so a single reduced local rate changes every volume beneath it.
- Primary constraintThe stage whose local rate was reduced.
- Downstream impactLocal rate unchanged. Receives and passes less volume.
- HealthyLocal rate and volume both at baseline.
- 01
Audience Coverage
Healthy- Local rate
- 80% · unchanged
- Output
- 160,000 reachable people
- 02
Ad Engagement
Healthy- Local rate
- 2.5% · unchanged
- Output
- 4,000 engaged visits
- 03
Landing Progression
Healthy- Local rate
- 25% · unchanged
- Output
- 1,000 visitors reaching the form
- 04
Form Completion
Healthy- Local rate
- 40% · unchanged
- Output
- 400 submissions
- 05
Lead Routing
Primary constraint- Local rate
- 95% reduced to 52.3%
- Output
- 209 leads reaching ownership
−45.0% vs baseline (380)
The injected constraint. This is the only local rate the scenario changed.
- 06
BDR Acceptance
Downstream impact- Local rate
- 70% · unchanged
- Output
- 146.3 accepted leads
−45.0% vs baseline (266)
From here down, every stage converts at its healthy rate on a smaller input. These stages are not broken — they are receiving less.
- 07
Trial Start
Downstream impact- Local rate
- 45% · unchanged
- Output
- 65.8 trial starts
−45.0% vs baseline (119.7)
- 08
Activation
Downstream impact- Local rate
- 40% · unchanged
- Output
- 26.3 activated users
−45.0% vs baseline (47.9)
- 09
Pipeline Conversion
Downstream impact- Local rate
- 25% · unchanged
- Output
- 6.6 opportunity equivalents
−45.0% vs baseline (12.0)
The method behind it
Four moves, in order. Most reporting stops at the metric that moved, which is usually the last stage in the chain rather than the one that failed.
- 01
Locate
Find the stage whose local rate deteriorated. Not the stage whose output looks worst — the stage whose own conversion changed.
- 02
Trace
Follow the lost throughput downstream. Every stage below the constraint inherits a smaller input, so its volume falls without its performance changing.
- 03
Separate
Distinguish the failed stage from stages that are merely receiving less. This is the distinction that decides where work goes.
- 04
Intervene
Fix the constraint before optimizing its symptoms. Improving a downstream stage that is already converting at baseline buys very little.
What this model is not
The baseline is an illustrative systems model, not benchmark data. It exists to make propagation legible, not to predict your numbers. No forecasting, no statistical inference, no causal proof. Nothing is stored, and nothing you do here leaves the browser.
The same reasoning runs through how I build demand systems: name the constraint, then decide, rather than optimizing whichever number happens to be visible.