Path analysis

Path analysis looks at the actual sequences of events a user performs, in order, rather than a single conversion metric. It answers questions a funnel can't: what do users do before and after a given step, which sequences most often lead to a purchase or churn, and where do unexpected detours appear.

Analysts typically visualize paths as a tree or Sankey diagram branching out from a starting event, showing the most common next steps at each stage, and can filter by segment to see whether paths differ between, say, new and returning users. It differs from funnel analysis, which tracks conversion through a small number of predefined, ordered steps; path analysis is exploratory and can surface routes nobody anticipated, but that also makes it noisier and harder to summarize in a single number. Both rely on a consistent event taxonomy to produce comparable sequences.

Path analysis matters for finding unplanned but effective routes to value, diagnosing why users abandon a flow at a point that isn't the "official" funnel step, and spotting confusing navigation. The main pitfalls are combinatorial explosion — the number of distinct paths grows quickly, making patterns hard to read without careful grouping — and treating a popular path as a good one, when it may simply reflect users struggling to find what they need.