Analytics reporting playbook

This playbook starts with a business question, not a chart. Use it when a client, manager, or content team needs a result they can reproduce next month.

Before you start

• Open Dashboard → Analytics and read the freshness banner first.
• Decide who will use the report and what decision it should support.
• Confirm the workspace timezone and the date when provider data is complete.
• Add campaign tags to posts before reporting if the analysis depends on a launch or initiative.
• Treat unavailable metrics as unavailable, not as zero.

Recipe 1: monthly stakeholder report

1. Select 30 days and compare with the previous period.
2. Start with all intended channels, then switch between Total and By platform.
3. Choose the metrics the reader can act on, such as impressions, engagements, engagement rate, followers, and posts published.
4. Check the channel coverage and data through time before interpreting a change.
5. Save the current view with a durable name such as Monthly brand — rolling 30 days.
6. Choose Create report, name the report, and arrange the narrative: Summary → Metrics over time → Channel performance → Top posts → Insights → Takeaways.
7. Open the report once and verify its date rule, timezone, sections, native post links, and missing-data notes.

A rolling report recalculates its dates when opened. Use a custom fixed range instead when the report must preserve an exact calendar month or campaign window.

Recipe 2: campaign retrospective

1. Select a custom range that covers the campaign, including the final day on which provider metrics are complete.
2. Filter to the campaign tag and relevant channels.
3. Compare with a deliberate baseline: the previous period or another custom range.
4. Open Insights to compare formats, topics, weekdays, publish hours, and approval status across the complete filtered set.
5. Open Posts and inspect the best and weakest posts using native links and the same primary metric.
6. Export Posts CSV for row-level evidence and Insights CSV for grouped findings.
7. Record the next action separately from the observation—for example, “test two short videos” instead of “short video always wins.”

Recipe 3: content experiment

Use one channel, one content kind, and one main metric. Keep the comparison window and timezone stable. In Insights, read the visible sample size before acting on weekday, hour, topic, or frequency patterns. A ranking is a correlation in the selected data, not proof that the format caused the result.

Copy an evidence-linked takeaway into a creation brief only after checking the posts behind it. Give feedback on a takeaway when it is useful or misleading; the feedback does not rewrite provider data.

Choose the right output

Exports use the active range, comparison, filters, metrics, grouping, and timezone. Review those controls before downloading; an export cannot explain an accidental filter.

You are done when

• The question, audience, date rule, timezone, and comparison are explicit.
• The freshness state and data-through time are acceptable for every included channel.
• Report sections tell one ordered story and use only available metrics.
• At least one native post or CSV row supports each important conclusion.
• The next action is framed as a test or decision, not an unsupported causal claim.

If something blocks you

For provider-specific streams and delays, use the channel capability guide. For the complete screen reference, see Analytics.