CommerceLens
Random Variables & Uncertainty

Conversion Forecast Simulator

Model Simulation

A binomial model treats n upcoming sessions as n independent, identical Bernoulli trials, each converting with probability p. Adjust n, p, and a target conversion count k below to see how the resulting probability distribution shifts — this is a forward-looking planning tool, not a re-analysis of historical rows.

Parameters
Defaults to n = 100 sessions at p = 0.2754, the dataset’s observed overall conversion rate — change either to explore “what if” scenarios.
Number of sessions (n)100
Conversion probability (p)0.275
Target conversions (k)28

Live-Computed Probabilities

P(X = 28)
8.81%
P(X ≤ 28)
59.15%
P(X ≥ 28)
49.65%
E[X] = np
27.5
expected conversions
Var(X) = np(1−p)
19.95
SD(X)
4.47
conversions
Binomial formulas
P(X = k) = C(n, k) · p^k · (1 − p)^(n − k) E[X] = n·p Var(X) = n·p·(1 − p) SD(X) = √(n·p·(1 − p))
PMF of X = number of conversions in 100 sessions
P(X = 28) = 8.81% · chart focused on mean ± 4 SD, where essentially all probability mass sits

The Binomial Model’s Assumptions

  1. A fixed number of trials, n. We decide in advance how many sessions we’re forecasting over.
  2. Each trial has exactly two outcomes. A session either converts or it doesn’t — no partial outcomes.
  3. A constant success probability p across all trials. Every session is assumed to convert with the same probability p, trial after trial.
  4. Trials are independent (i.i.d.). One session’s outcome is assumed to give no information about another’s.
Real e-commerce traffic can violate assumption #3 — seasonality, campaigns, and pricing changes all shift the true conversion probability over time, so p is rarely truly constant — and assumption #4, since a single customer’s repeat sessions are correlated with each other, not independent. The binomial model is a simplifying planning assumption. It is different in kind from the observed historical dataset used elsewhere on this site, which doesn’t assume any of this — it just counts what happened.