Monte Carlo Journey Simulation

Description

Simulates thousands of synthetic user journeys from observed step success probabilities to forecast end-to-end completion as a distribution with intervals and to test what-if improvements.

Monte Carlo simulation estimates outcomes by repeated random sampling. Applied to UX funnels, each virtual user passes or fails each step according to empirical step probabilities, and the share completing across many runs gives a completion distribution rather than a single number. In Vivek Gopinath's playbook, step probabilities of 81%, 75%, 85% and 95% across 10,000 virtual users give a mean completion of 49% [46%, 52%]; improving OTP to 92% gives 54% [51%, 57%], showing small step tweaks compound into larger overall lifts. Assumptions are stated explicitly: stationarity, approximate independence of steps, like-for-like traffic and no leakage. Sensitivity checks vary the improved step by ±10–20%, split by segment, and scale by expected weekly starts. CIs can be converted to SD with SD ≈ (upper − lower)/(2 × 1.96). The same technique forecasts ΔU under redesign scenarios in the Nash protocol.

Attributes

Loss type
Latent
Stage
Design, Strategy
Effort
Days

Steps

  1. Collect empirical step-level success probabilities from experiments or analytics.
  2. For each of 10,000 virtual users, sample success or failure at each step using those probabilities.
  3. Aggregate the share of virtual users finishing and repeat to obtain a completion distribution.
  4. Report the mean and 95% interval.
  5. Run what-if scenarios by changing the probability of a target step.
  6. Run sensitivity checks (best/mid/worst ±10–20%, segment splits, volume scaling) and state assumptions.
  7. Compare the forecast with the observed lift after launch.

Sources

  • Vivek Gopinath (2025), Forward-Thinking UX Research Playbook: A Mixed-Methods, Decision-Ready Approach from Prediction to Validation
  • Vivek Gopinath (2025), Equilibrium and the Design Object: How Nash Equilibrium Illuminates Why Designs Stabilise, Plateau, and Change

Also known as

  • Monte Carlo funnel forecasting
  • Synthetic journey simulation

Categories

  • Analytics & Behaviour
  • Statistics & Modelling