Campaign performance review

Purpose

Explain what a completed campaign achieved, where performance differed from its plan, and what should be continued, changed, tested, or stopped.

Scope

The skill covers goal variance, channel and asset performance, audience patterns, costs, attribution limits, and recommendations.

Excluded: changing campaign data, reallocating live budget, and claiming causal certainty from incomplete evidence.

Data basis

• Campaign plan and objectives in scope.
• Performance, spend and conversion tables supplied or in scope.
• Prior campaign comparison when available.

Result

A retrospective, channel-performance sheet, and next-time playbook.

Quality criteria

• Every metric has a source row and defined denominator.
• Actuals are compared to the plan without mixing periods.
• Findings separate observation, interpretation and hypothesis.
• Recommendations name evidence, effort and decision use.

Instructions

Respect the campaign's attribution model and reporting window. Reconcile totals before interpreting rates. Do not rank channels with incomparable denominators. Treat causal explanations as hypotheses unless the design supports them. State missing baselines, delayed conversions and tracking limitations.

Adapt before use

• Map campaign, spend, channel, asset and conversion fields.
• Define attribution, reporting window and metric formulas.
• Add prior campaigns and decision thresholds.
• Define owners for follow-up actions.
• Keep source facts, analysis, recommendations, and unresolved questions visibly separate.
• Preserve stable document, section, page, sentence, URL, or row identifiers in every finding.
• Record assumptions about audience, channel, period, approval status, and data coverage.
• Treat missing guidance as an adaptation gap rather than inventing a company rule.
• Prefer a precise, bounded recommendation over a broad generic rewrite.
• Use supplied terminology consistently across the document, sheet, and any derivative copy.
• Do not infer performance, customer behavior, product capability, or market response from wording alone.
• Make every unresolved issue actionable by naming the missing decision, evidence, owner, or source.
• Keep factual claims traceable even when the final copy is concise.
• Recheck that counts, names, dates, and identifiers match across all outputs.
• Mark hypotheses as hypotheses and preserve competing explanations when evidence is mixed.
• Use the stated acceptance criteria as the final completeness check.