developer-journey-map

Start with an interview to fix one audience segment and one end state, then select a published stage model whose terminal stage matches that end state. Gather evidence first through friction logs, support tickets, and exit interviews, then fill one row per stage with an exit event, owner, and signal.

Delete every model whose terminal stage is not your funded end state, then rank survivors by observation cost. The API five-stage arc is cheapest when telemetry exists, the OSS contributor funnel fits contribution goals, and Revell's six-stage model fits organisation-wide standardisation.

No. Circulating per-stage conversion figures are single-company anecdotes with incompatible stage definitions, so no normal discovery-to-production rate exists. The only honest comparison is against your own map's previous version.

Plausible's August 2021 measurement found 58% of a tech audience blocks third-party analytics scripts, with rates varying page by page. Absolute session counts are wrong by a varying factor, so only ratios inside one source survive and early stages should be labelled inferred.

Add a buyer lane in B2B when the developer is not the economic buyer, with its own exit events like security questionnaire cleared and contract signed. Roll individual usage up to accounts first, since forty signups from one company read as forty leads otherwise.

Exactly one. Rank candidate stages by developers recovered per unit of fix effort, then pick the earliest leaking stage with an owner able to change it, because downstream fixes are diluted by losses above them.