Why did CAC change?
Separate traffic, cost, conversion, channel mix, attribution, and data-health explanations.
Ask in Codex →ANALYSIS SKILLS
Repeatable analysis for the questions growth teams ask every week.
Each skill defines the scope, evidence operations, metric semantics, caveats, and output structure before the final interpretation.
Are cohorts ready to compare?
Check attribution windows, freshness, maturity, and coverage before drawing a conclusion.
Explore MMP evidence →Where did delivery shift?
Compare channels and campaigns while retaining spend, delivery, and outcome context.
Explore media sources →Which assets may be tiring?
Use available creative delivery evidence and preserve partial attribution caveats.
See planned creative workflows →Is the result trustworthy?
Distinguish unknown, delayed, partial, unauthorized, no-delivery, and true-zero states.
See operating boundaries →Turn analysis into a reusable readout.
Preserve scope, metric definitions, findings, evidence, caveats, and next questions.
Build a report in Codex →The skill selects the work. The evidence supports the answer.
ChatGrowing uses governed operations before language-model interpretation.