Marketing performance is easy to summarize and hard to interpret. Different platforms can count the same journey differently, definitions drift, and a visible change does not automatically explain why it happened.

Strawberry can work across logged-in campaign tools, analytics, CRM or lifecycle data, product signals, prior reports, and qualitative feedback. Your companion can preserve the definitions and source limits behind the analysis, then carry the accepted learning into the next campaign or content decision.

Define the review and the decision

Clarify the business and marketing goals, campaign or channel, audience, funnel, period, comparison, source systems, budget context, and decision the review should support. Begin with the slice most likely to change current work.

Want to try it?

Ask your Strawberry companion: “Review our marketing performance, make the attribution and data limits clear, and recommend the decisions or experiments the evidence supports.”

Skill

Review marketing performance

Review the performance evidence behind a current marketing decision and make the limits explicit.

Reconcile before comparing

CheckWhy it changes the conclusion
Date range, time zone, and currencyPeriods that look aligned may contain different activity or budget.
Metric and funnel definitionsA lead, conversion, activation, or qualified opportunity may not mean the same thing across tools.
Attribution window and modelPlatforms may claim the same outcome or omit effects they cannot observe.
Identity, deduplication, and denominatorRates and audience comparisons fail when the counted populations differ.
Freshness and modeled dataDelayed, sampled, or estimated results should not carry the same confidence as observed records.

Use shared Data capabilities when extraction, joins, statistics, or validation becomes substantial. Marketing still owns the audience, channel, creative, funnel, and campaign interpretation.

Separate change, explanation, and action

Start with the few changes that matter across the objective, baseline, funnel, audience, geography, channel, placement, message, offer, creative, spend, and external context. Then test plausible explanations against both quantitative and qualitative evidence.

  1. State what changed and the evidence behind it.
  2. List plausible explanations and credible alternatives.
  3. Explain what the current evidence cannot distinguish.
  4. Recommend the smallest decision, test, or measurement improvement that could reduce uncertainty.

Do not claim causation from a before-and-after chart or give one platform's attribution model more certainty than it deserves.

Leave with a learning plan

Recommend a small set of actions or experiments with the rationale, owner, expected signal, review window, and major dependency. Some results should lead to better measurement or a smaller test rather than immediate optimization.

The review can define a better metric framework. It does not change tracking, configure campaigns, edit creative, launch tests, or change spend. Keep each external action separately permitted and verifiable.

Carry the learning back into marketing

Preserve accepted definitions, comparisons, source rules, analysis shape, and review behavior for the next cycle. Still check for definition drift and material data gaps rather than mechanically repeating the dashboard.

Skill

Plan a marketing campaign

Use accepted performance learning to improve the next campaign decision.

Skill

Build a content plan

Use audience, channel, and creative learning to improve the next content cycle.

Official Strawberry skill
Copy skill file Download skill file
name
review-marketing-performance
description
Review marketing performance across campaigns and channels by reconciling definitions, source quality, baselines, trends, attribution, funnel behavior, audiences, and creative. Use when a team needs a decision-ready performance view and recommendations; not for client reporting, deep data engineering, platform configuration, or live spend changes.

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.

1. Define the review

Clarify the business and marketing goals, campaigns or channels, audience, funnel, period, comparison, source systems, budget context, and the decision the review should support. Use the team's accepted definitions and reporting habits when they exist.

If the request is broad, begin with the campaign, funnel stage, or period most likely to change a current decision. For client-facing, multi-source progress reporting, use strawberry/operations/build-client-progress-report-project-tools; keep this skill focused on marketing interpretation.

2. Reconcile the evidence

Use approved analytics, campaign platforms, CRM or lifecycle data, product signals, content data, prior reports, experiments, and relevant qualitative feedback. Preserve source links or record identifiers and freshness.

Align date ranges, time zones, currencies, metric definitions, attribution windows, identity and deduplication rules, funnel stages, and denominator choices before comparing results. Keep platform-reported attribution separate from independently observed outcomes.

Use available shared Data capabilities when extraction, joins, statistical work, or validation becomes substantial. Surface missing, delayed, sampled, modeled, or conflicting data instead of forcing false reconciliation.

3. Analyze what changed

Review the evidence across the dimensions that can change the decision:

  • progress against the stated objective and useful baseline;
  • trend, seasonality, mix, spend, reach, response, conversion, retention, or revenue;
  • funnel movement and the point where performance changed;
  • audience, segment, geography, channel, placement, message, offer, and creative differences;
  • operational changes or external events that could explain the pattern; and
  • uncertainty, concentration, and missing evidence.

Separate observed change, plausible explanation, and recommendation. Do not claim causation from a before-and-after chart or give one platform's attribution model more certainty than it deserves.

4. Diagnose and recommend

Identify the few audience, proposition, channel, creative, funnel, or measurement issues most likely to matter. Compare hypotheses against both quantitative and qualitative evidence.

Recommend practical actions or experiments with their rationale, owner, expected signal, review window, and major dependency. Some results should lead to more measurement or a smaller test rather than immediate optimization.

5. Deliver and keep actions separate

Provide:

  • scope, sources, definitions, freshness, and material limitations;
  • the few changes that matter, with their evidence;
  • funnel, audience, channel, and creative diagnosis where supported;
  • hypotheses and alternative explanations;
  • prioritized recommendations or experiments; and
  • open questions and the next review point.

The review may define a better metric framework when current measures do not match the goal. It does not change tracking, configure campaigns, edit creative, launch tests, or change spend.

Follow Strawberry's active scoped permission for every source, account, destination, and action. Draft or ask when permission is insufficient. Stop when identity, scope, impact, or sensitive-data handling changes, and verify completed external actions.

After the team accepts the definitions, comparisons, analysis shape, and review behavior, preserve them for the next cycle. A recurring review should still surface definition drift and material data gaps rather than mechanically repeating last period's dashboard.