Definition
The percent agreement between supply chain digital twin simulations (vendor network behavior, multi-echelon model output, resilience scenarios) and actual realized lead times, vendor responses, and network outcomes.
The Measurement Gap
Your supply chain digital twin tells you that if your top-three steel vendors all face capacity constraints simultaneously, the network can absorb the shock by shifting to next-tier vendors with a maximum 9-day delay on critical fasteners. Your resilience team signs off on the consolidation plan. Eighteen months later, two of the three face simultaneous constraints. The predicted recovery path does not materialize: next-tier vendors have lost qualified status with your quality team, the supposed dual-source has consolidated to the same parent company, and the actual delay is 47 days. The twin was wrong, and its wrongness was invisible until the event. Every decision that depended on it (sourcing strategy, vendor consolidation, buffer placement, dual-source policy) was committed during the period the twin was silently drifting from reality. Vendor on-time delivery measures realized performance against stated lead time, not against simulated performance. Stockout rate measures consequences after the fact. Working capital measures the financial expression of resilience decisions but says nothing about whether those decisions were grounded in a trustworthy simulation. A twin that is accurate in steady state and inaccurate during disruptions is not a planning instrument: it is a confidence-generating illusion.
Decision Supported:
Whether resilience plans, scenario tests, and what-if analyses produced by the digital twin can be trusted to inform sourcing strategy, dual-source decisions, and buffer placement, and on which segments trust should be withdrawn.
Formula
Plain language: Run the twin on a recent real event, compare its predicted behavior to what actually happened. The closer the match, the more trustworthy the twin. Compute separately for steady-state operations and disruption-response scenarios.
Notation:
SNTF = 1 − mean(|sim_metric_i − actual_metric_i| / actual_metric_i)
Variables:
| Variable | Definition | Unit |
|---|---|---|
| sim_metric_i | Twin-predicted value for event i (lead time, fill rate, recovery time) | Varies |
| actual_metric_i | Realized value for event i | Varies |
| i | Index across vendor segments, lead times, and disruption responses | Index |
Confidence Grading (required alongside every SNTF value):
| Grade | Basis | Interpretation |
|---|---|---|
| A: Empirical | 5+ observed events in measurement window provide direct tail observation | Measurement grounded in observed outcomes; highest confidence |
| B: Hybrid | 1-4 observed events plus Monte Carlo simulation with Bayesian updating | Blends observation and simulation; moderate confidence |
| C: Simulated | Zero observed events; derived entirely from Monte Carlo simulation | Modeled, not observed; treat as indicative, not validated |
A SNTF of any value at Grade C is a simulation that agrees with itself. It tells you the model is internally consistent but not that it matches reality. Grade C scores should trigger proactive stress-testing against historical near-miss events and cross-asset analogues rather than acceptance as calibrated.
Proposed formula, pending validation pilots.