E24 · Explain recorded paths

Synthetic practice, not company performance.

One decision

Where does the last recorded step occur?

Run the exact query

WITH b AS (SELECT * FROM v_task WHERE ('all'='all' OR persona='all') AND ('all'='all' OR workflow='all') AND ('all'='all' OR substr(signup_at,1,7)='all') AND date(created_at)>='2026-05-01' AND date(created_at)<'2026-10-01') , checkpoints AS(SELECT task_id,'created' stage,created_at at,1 priority FROM b UNION ALL SELECT task_id,'ready',ready_at,2 FROM b WHERE ready_at IS NOT NULL UNION ALL SELECT task_id,'preview',preview_at,3 FROM b WHERE preview_at IS NOT NULL UNION ALL SELECT task_id,'export',exported_at,4 FROM b WHERE exported_at IS NOT NULL UNION ALL SELECT task_id,'review',reviewed_at,5 FROM b WHERE reviewed_at IS NOT NULL UNION ALL SELECT task_id,'distributed',distributed_at,6 FROM b WHERE distributed_at IS NOT NULL UNION ALL SELECT task_id,'use',confirmed_use_at,7 FROM b WHERE confirmed_use_at IS NOT NULL), edges AS(SELECT task_id,stage source,LEAD(stage) OVER(PARTITION BY task_id ORDER BY at,priority) target FROM checkpoints) SELECT source,COALESCE(target,'last_recorded') target,COUNT(*) tasks FROM edges GROUP BY 1,2;

Download SQL · Result CSV · Metabase saved question

Chart and reading

Use timestamp-ordered edges and conservation; review/export can swap order.

Challenge your recommendation

Are skipped events instrumentation gaps or true behavior?

Guardrails: Ordered by actual timestamps; review/export can occur in either order. Last recorded is not proven abandonment.

Next evidence: Observed task traces and missing event tests.

PRD sections: 6, 10, 14. Open proposed PRD

Your spoken decision

State one takeaway, one caveat and one action in 30 seconds. Then stop.

Connect this finding

What this chart helps decide · Follow the connected decision · Canonical metric definitions

All exercises