Monte Carlo Simulation
Stress-test your supply plan across thousands of randomized scenarios to quantify shortage risk and determine optimal safety stock levels.
Monte Carlo Simulation is a premium feature that runs thousands of randomized scenarios to quantify supply risk and determine optimal safety stock levels.
What Is Monte Carlo Simulation?
Your deterministic forecast shows one scenario: a single best-guess demand line based on configured enrollment rates, visit schedules, and production lead times. But reality rarely follows the plan exactly.
Monte Carlo simulation runs thousands of "what-if" scenarios -- each one slightly different. In every scenario, the system randomly varies enrollment speed, dropout rates, visit timing, manufacturing lead times, and other parameters within realistic ranges. After thousands of runs, you get a full picture: not just one demand number, but a range of possible outcomes and the probability of each.
Instead of planning around a single forecast, you can see the probability of a supply shortage, how large it might be, and how much safety stock you need to reach your target confidence level.
Why It Matters
- Risk quantification: Know the actual probability of a kit shortage, not just a hunch
- Buffer sizing: Determine exactly how much safety stock achieves 95% or 99% confidence
- Decision-making: Compare options -- increase safety stock, negotiate faster manufacturing, or accept the risk
- Stakeholder communication: Present clear data like "30% chance of shortage with current plan" instead of vague risk statements
Deterministic vs. Monte Carlo
| Aspect | Deterministic Forecast | Monte Carlo Simulation |
|---|---|---|
| Approach | One best-guess scenario | 1,000 to 10,000 possible scenarios |
| Assumption | Everything goes as planned | Accounts for real-world variability |
| Output | Single demand/supply line | Range of outcomes with probabilities |
| Risk visibility | Limited -- no shortage probability | Comprehensive -- probability, magnitude, timing |
| Use case | Baseline planning | Risk assessment, safety stock decisions, stakeholder reporting |
Accessing Monte Carlo Simulation
Premium Feature: Monte Carlo Simulation is available on paid plans only. Free tier accounts are redirected to the Analytics overview.
Location: From any trial, navigate to the Simulation page in the trial outputs section.
URL pattern: /home/[account]/trials/[trial-id]/simulation
Prerequisites:
- Trial setup must be complete (all 8 wizard steps finished)
- Account must be on a paid plan
Configuring a Simulation
Quick Presets
Three presets to get started fast:
| Preset | Service Level | Precision | Variability | Skew |
|---|---|---|---|---|
| Conservative | 99% | High (10,000 runs) | Maximum | Higher (worst-case bias) |
| Balanced | 95% | Standard (5,000 runs) | Default | As Projected |
| Lean | 90% | Quick (1,000 runs) | Minimum | Lower (best-case bias) |
Pick Conservative for critical supply decisions. Lean for a fast directional check. Balanced is a good default.
Service Level Target
Choose the confidence level: 90%, 95%, 97%, or 99%.
A 95% service level means the simulation recommends enough supply to cover 95% of all simulated scenarios. Higher targets require more safety stock but reduce shortage risk.
Variability Settings
Control how much real-world uncertainty the simulation introduces:
| Setting | Range | What It Controls |
|---|---|---|
| Visit Window | 0-7 days | How much patient visit timing can drift from schedule |
| Screening Period | 0-28 days | How much screening delays can compress enrollment |
| Lead Times | 0-14 days | How much production and shipping times can extend |
Higher values mean more uncertainty and wider outcome ranges. Lower these if you have strong confidence in operational timelines. Increase them for historically variable logistics.
Rate Adjustments
- Screen Failure Skew: Shift screen failure rates Higher, Lower, or As Projected
- Dropout Skew: Shift patient dropout rates Higher, Lower, or As Projected
Each control shows the currently configured rate from your trial setup for context.
Simulation Precision
| Level | Runs | Estimated Time | Best For |
|---|---|---|---|
| Quick | 1,000 | ~1 minute | Fast directional check |
| Standard | 5,000 | ~3-5 minutes | Routine planning |
| High Precision | 10,000 | ~10 minutes | Critical decisions, stakeholder reports |
Start with Quick to validate configuration, then re-run at Standard or High Precision for final results.
Scenario Label
If previous simulation runs exist, a label field lets you name the new run (e.g., "Baseline Q2", "Increased safety stock") for easy comparison later.
Running the Simulation
Click Run Monte Carlo Simulation to start. Only one simulation can run per trial at a time. The simulation processes as a background job -- you can leave the page and return later.
What the Simulation Varies
Each simulated scenario randomly perturbs these parameters:
Enrollment Variables (per country):
- Site activation delay (up to 30 days late)
- Screening delay (up to your configured screening period)
- Enrollment duration (varies by +/-10%)
- Monthly enrollment rate (Poisson-Gamma model with +/-30% coefficient of variation)
- Screen failure rate (varies +/-5 percentage points, with optional skew)
Treatment Variables (per arm):
- Dropout rate (varies +/-5 percentage points, with optional skew)
- Visit timing (shifts around scheduled offsets based on your visit window setting)
- Arm allocation (slight per-country variation)
Supply Variables:
- Production lead time (extends up to your configured lead time days)
- Shipping lead time (extends up to your configured lead time days)
Result Tabs
After the simulation completes, seven tabs present different lenses on your trial's risk profile.
Summary
Risk overview cards with high-level metrics (shortage probability, supply coverage status) and a per-kit, per-country summary table. Start here to quickly identify the most at-risk kit-country combinations.
Enrollment
A cone chart showing possible enrollment outcomes over time. Shaded bands represent percentile ranges (P5 to P95/P99), with the median (P50) as the center line. Available in monthly, cumulative, and per-country views. The wider the cone, the more uncertain enrollment is.
Monthly Demand
Per-kit, per-country demand with confidence intervals for each month. Shows how many kits you likely need (P50), the range of possible demand (P5 to P95), and which months have the widest uncertainty.
Production
Production requirements with confidence bands, plus overage recommendations:
| Column | Meaning |
|---|---|
| Kit | Kit name |
| Country | Destination country |
| Current % | Your configured overage percentage |
| Recommended % | Overage needed to meet your service level target |
| Status | Optimal (within 2%), Increase (under-buffered), or Reduce (over-buffered) |
| Kit Delta | How many more or fewer kits the recommendation implies |
| Cost Impact | Estimated cost difference |
The overage recommendation is one of the most actionable outputs. If it suggests increasing overage, your current safety stock is insufficient for your chosen service level.
Inventory
Projected inventory balance over time with shortfall risk overlay. The inventory cone chart shows running balance with percentile bands -- when lower bands dip below zero, there is stockout risk. Supply coverage card shows Full, Partial, or Critical status.
Cost Analysis
Two visualizations:
- Cost Distribution Summary -- Percentile breakdown (P5, P25, P50, P75, P95, P99) of total trial cost
- Sensitivity Tornado Chart -- Shows which input parameters have the biggest impact on cost. The wider the bar, the more influence.
If enrollment variability dominates the tornado chart, focus on tightening enrollment assumptions. If lead time variability dominates, focus on manufacturing reliability.
Risk Analysis
The most direct view of shortage risk:
- Shortfall probability -- For each kit, country, and month, what percentage of scenarios resulted in a shortage
- Expected shortfall -- If a shortage occurs, how many kits short on average
- P50 and P90 shortfall -- Median and worst-case shortage magnitudes
Color coding for quick scanning:
| Color | Meaning |
|---|---|
| Green | Low risk (shortage in less than 10% of scenarios) |
| Yellow | Moderate risk (shortage in 10-30% of scenarios) |
| Red | High risk (shortage in more than 30% of scenarios) |
Comparing Scenarios
Run multiple simulations with different configurations and compare them using the Scenario Switcher at the top of the results dashboard.
Example workflow:
- Run a Balanced simulation -- see 25% shortage risk
- Increase overage by 10% in your trial configuration
- Run a new simulation labeled "Increased overage" -- see 8% shortage risk
- Compare both scenarios side by side to justify the overage increase
When viewing a historical scenario, a banner shows which one you are viewing with a Back to Latest button.
Exporting Results
Click the Export dropdown to download CSV files:
- Enrollment Percentiles
- Kit Demand
- Cost Distribution
- Shortfall Risk
- Production Requirements
- Inventory Projections
- All Data Combined
Use exports for stakeholder presentations, further analysis in Excel, or archival.
Decision-Making Workflow
Step 1: Check Overall Risk
Start with the Summary tab. Is the probability of shortage acceptable?
| Shortage Probability | Risk Level | Suggested Action |
|---|---|---|
| Below 10% | Low | Current plan is likely sufficient |
| 10-30% | Moderate | Consider increasing safety stock |
| Above 30% | High | Adjust supply plan before proceeding |
Step 2: Identify At-Risk Periods
Switch to Risk Analysis. Which months and kit-country combinations show the highest shortage probability?
Step 3: Consider Your Options
- Accept the risk -- If shortage probability is within tolerance and mitigation cost outweighs the benefit
- Increase safety stock -- Use the overage recommendations from the Production tab
- Improve manufacturing -- If lead time variability is a major driver (check the tornado chart)
- Adjust enrollment -- If enrollment uncertainty drives risk, add countries or sites to reduce variability
Step 4: Run What-If Scenarios
Adjust your trial configuration, re-run the simulation, compare to baseline, and iterate until risk is acceptable.
Tips
- Start with your deterministic forecast. Use the deterministic analytics tabs to establish your baseline plan, then run Monte Carlo to stress-test it.
- Validate assumptions. Review variability settings with your team before sharing results. Are the configured ranges realistic for your trial?
- Use scenarios for stakeholder conversations. Run a "current plan" and "improved plan" scenario. Present both to justify safety stock or manufacturing investments with data.
- Iterate, don't just run once. The most value comes from the cycle: run simulation, identify risks, adjust plan, re-run, confirm improvement.
- Match precision to the decision. Use Quick for exploratory checks while adjusting configurations. Switch to Standard or High Precision for the final run that informs actual supply decisions.