Journey generator

Journeys to simulate: 4000

Removal-effect test

Channel to test for removal

Model stats

Journeys simulated—

Empirical conv. rate—

Model conv. probability—

Channel under test—

Conv. prob. w/o channel—

Removal effect (drop)—

Normalized attribution—

Attribution % by channel

Empirical transition probabilities are fit from the simulated journeys. Total conversion probability is solved exactly from Start via the absorbing-Markov-chain fundamental matrix N=(I−Q)⁻¹. Each channel's removal effect is the drop in that probability when its incoming transitions are redirected to the Null state.