Performance Report

Performance Report

Generate structured weekly and monthly performance reports that turn raw metrics into actionable insights. The report answers three questions: what happened, why it happened, and what we should do differently.

Purpose

A performance report that nobody reads is worse than no report at all. This skill produces reports designed for one specific reader — the founder — and answers exactly what they need to know in the order they need to know it.

When to Use

• Weekly: Monday morning review of previous week
• Monthly: First week of month — trends across 4 weeks
• Campaign post-mortem: within 48 hours of campaign end
• Founder asks "how's social performing?"

Inputs Required

• Metrics data for the period (from metrics-track)
• A/B test results (from ab-test-analyze)
• Content calendar (planned vs actual)
• Any external events that might have influenced performance

Quick Reference

| Report Type | Frequency | Key Sections | Audience | |------------|-----------|-------------|---------| | Weekly Snapshot | Every Monday | Top/bottom posts, metrics summary, 1 insight, 1 action | Founder + team | | Monthly Deep-Dive | First week of month | Trends, format analysis, growth trajectory, forecast | Founder + strategy | | Campaign Post-Mortem | After campaign | Goals vs actuals, lessons, format winners | Founder + team |

Procedure

1. Gather data: metrics, A/B test results, competitor benchmarks, content calendar.

1. Build metrics summary:

• Total posts per platform
• Total impressions
• Average ER per platform
• Follower growth (net + growth rate)
• Link clicks / conversions
• Compare to previous period

1. Identify top 3 and bottom 3 posts:

• What made them work/fail (hook, format, timing, topic)?
• Include the defining metric

1. Analyze by content type:

• Threads vs short posts vs carousels vs videos — which format wins?
• Has the winning format changed?

1. Analyze by content pillar:

• Which pillars drive engagement vs conversions?
• Is the pillar mix matching strategy?

1. Track conversion funnel: Impressions → engagement → profile visits → link clicks → signups.

1. Extract 3-5 key learnings:

• What worked (do more)
• What didn't (stop doing)
• What to test next
• Update always-on learnings doc

1. Provide 2-3 actionable recommendations:

• Specific and testable (not "post more")
• "Increase thread cadence from 2/week to 3/week targeting Tue/Thu mornings"

Output Format

# [Weekly / Monthly] Report: [Period]

## Executive Summary
[3-4 sentences — founder reads this in 30 seconds]

## Metrics at a Glance
| Metric | This Period | Last Period | % Change |
|--------|-----------|------------|----------|
| Total posts | [N] | [N] | [+/-N]% |
| Total impressions | [N] | [N] | [+/-N]% |
| Avg ER (X) | [N]% | [N]% | [+/-N]% |
| Avg ER (LinkedIn) | [N]% | [N]% | [+/-N]% |
| Follower growth | [+N] | [+N] | [+/-N]% |
| Link clicks | [N] | [N] | [+/-N]% |

## Top 3 Posts
### 1. [Title]
**Platform:** [X / LinkedIn / IG / TikTok] | **Format:** [Type]
**Impressions:** [N] | **ER:** [N]% | **Why it worked:** [Analysis]
### 2-3...

## Bottom 3 Posts
### 1. [Title]
**Impressions:** [N] | **ER:** [N]% | **Why it underperformed:** [Diagnosis]

## Format Analysis
| Format | Posts | Avg ER% | vs Overall |
|--------|-------|---------|-----------|
| Thread | [N] | [N]% | [+/-N]% |
...

## Pillar Performance
| Pillar | Posts | Avg ER% | Top Post |
|--------|-------|---------|----------|
...

## Key Learnings
1. [Specific learning with supporting data]
2-3...

## Recommendations
1. [Specific, actionable recommendation]
2-3...

Done Criteria

The skill is complete when:

1. Executive summary is written (founder can read in 30 seconds)
2. Top and bottom 3 posts are identified with analysis
3. Format and pillar analysis are included
4. 3-5 key learnings are extracted
5. 2-3 actionable recommendations are provided
6. Always-on learnings doc is updated

Pitfalls

• Data dump without insight (metrics without interpretation = noise)
• Comparing platforms naively (X ER != LinkedIn ER)
• Highlighting problems without solutions
• Cherry-picking favorable comparisons
• Report so long nobody reads it

Verification

Can the founder read the executive summary and know whether it was a good week? Are there at least 2 actionable next steps? Does this report make the next content calendar smarter than the last one?

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: gokulb20
• Source: gokulb20/Crewm8-Social-Media-Manager-Skill-Graph
• License: MIT
• Homepage: https://crewm8.ai

Install and usage instructions live in the source repository linked above.