Review Marketing Performance

Turn marketing data into a decision-ready view of what changed, what the evidence can support, and what the team should learn or test next.

Do not confuse a chart with an explanation or platform attribution with objective causation.

1. Define the review

Establish:

• business goal;
• marketing goal;
• campaigns;
• channels;
• audience;
• funnel;
• review period;
• comparison period;
• source systems;
• budget context;
• decision the review must support.

Use existing accepted:

• metric definitions;
• funnel stages;
• reporting conventions;
• attribution assumptions

when available.

If the request is broad, begin with the:

• campaign;
• funnel stage;
• channel;
• period

most likely to change a current decision.

Keep this workflow focused on marketing interpretation.

Use a broader operations or client-reporting workflow when the real task is multi-source client progress reporting.

2. Reconcile source data before interpretation

Use approved sources such as:

• analytics;
• campaign platforms;
• CRM;
• lifecycle data;
• product signals;
• content performance;
• previous reports;
• experiments;
• qualitative feedback.

Preserve:

• source;
• source identifier;
• date;
• freshness;
• relevant filters.

Align:

• date ranges;
• time zones;
• currencies;
• metric definitions;
• attribution windows;
• identity rules;
• deduplication logic;
• funnel stages;
• denominators.

Keep platform-reported attribution separate from independently observed outcomes.

When extraction, joining, statistical work, or validation becomes substantial, use the environment's data-analysis capabilities.

Surface data that is:

• missing;
• delayed;
• sampled;
• modeled;
• contradictory.

Do not force false reconciliation.

3. Analyze what changed

Review the evidence across dimensions that can materially affect the decision.

Objective progress

Assess performance against:

• campaign objective;
• marketing objective;
• accepted baseline.

Trend and mix

Review where relevant:

• trend;
• seasonality;
• spend;
• reach;
• response;
• conversion;
• retention;
• revenue;
• mix.

Funnel

Identify where performance changed across:

• acquisition;
• engagement;
• lead;
• conversion;
• activation;
• retention

or the team's accepted funnel.

Audience and segment

Compare, where useful:

• segment;
• geography;
• device;
• audience;
• customer type.

Channel and placement

Review:

• channel;
• placement;
• campaign;
• source;
• medium.

Message and creative

Look for differences in:

• proposition;
• message;
• offer;
• creative;
• format.

External and operational context

Consider:

• launch changes;
• pricing changes;
• website changes;
• product changes;
• seasonality;
• external events;
• tracking changes.

4. Separate observation, explanation, and recommendation

For every important conclusion, distinguish:

Observed change

What the data directly shows.

Plausible explanation

What may explain the pattern.

Recommendation

What the team should do next.

Do not claim causation from:

• simple before-and-after comparisons;
• one platform's attribution model;
• correlation alone.

5. Diagnose the highest-value issues

Identify the few issues most likely to matter across:

• audience;
• proposition;
• channel;
• creative;
• funnel;
• measurement.

Compare hypotheses against:

• quantitative evidence;
• qualitative evidence.

Include alternative explanations where credible.

6. Recommend actions or experiments

Recommend practical next steps.

For each, include:

• rationale;
• owner;
• expected signal;
• review window;
• dependency;
• risk where relevant.

Some results should lead to:

• better measurement;
• smaller experiment;
• additional evidence

rather than immediate optimization.

Do not recommend large changes when evidence is weak and a smaller test could resolve uncertainty.

7. Deliver the review

Provide:

• scope;
• source list;
• metric definitions;
• freshness;
• material limitations;
• most important changes;
• supporting evidence;
• funnel diagnosis;
• audience diagnosis;
• channel diagnosis;
• creative diagnosis;
• hypotheses;
• alternative explanations;
• prioritized recommendations;
• experiments;
• open questions;
• next review point.

When the existing metric framework does not reflect the business goal, recommend a better measurement structure.

Do not change tracking automatically.

8. Keep analysis separate from execution

This workflow does not automatically:

• configure campaigns;
• edit creative;
• launch tests;
• change spend;
• alter tracking;
• publish assets.

Follow active scoped permission for every:

• data source;
• account;
• destination;
• action.

Draft or ask when permission is insufficient.

Stop when:

• identity changes;
• scope changes;
• expected impact changes materially;
• sensitive-data handling changes.

Verify completed external actions when possible.

9. Preserve the accepted review method

After the team accepts:

• definitions;
• comparison logic;
• analysis structure;
• review behavior;

preserve them for future cycles.

A recurring review should still check for:

• definition drift;
• source changes;
• material data gaps;
• attribution changes.

Do not mechanically repeat last period's dashboard when the underlying measurement model has changed.