campaign-analyser

Campaign performance analysis for ProductClank agents and creators.

An Agent Skill that turns raw campaign data into actionable insights — ROI per campaign, keyword patterns, content style analysis, and a concrete next-campaign strategy.

---

What it does

When asked to analyse campaign performance, campaign-analyser runs a four-step review:

1. Data collection — fetches live campaign structure via ProductClank API, or accepts pasted campaign data in manual mode
2. ROI calculation — calculates points-per-credit for every campaign and ranks them
3. Pattern analysis — identifies what keywords, content styles, and post volumes drive the best results
4. Recommendations — produces a specific next-campaign strategy based on historical patterns

It then outputs a structured performance report with rankings, insights, and an actionable plan.

---

Two modes

Both modes produce the same structured report.

Note on API mode: The ProductClank agent API currently returns campaign structure (names, keywords, dates) and post counts. Points earned and $PRO per campaign are not yet exposed via API — the skill will prompt you to enter these manually after fetching. See Future capabilities below.

---

Install

Load the skill into your agent:

https://github.com/AyushVerma7861/campaign-analyser/blob/main/SKILL.md

Or fetch raw:

curl -s https://raw.githubusercontent.com/AyushVerma7861/campaign-analyser/main/SKILL.md

---

Quick start

# Pull campaign structure from the ProductClank API
chmod +x scripts/fetch-data.sh
PRODUCTCLANK_API_KEY=your_key ./scripts/fetch-data.sh

# Paste the output into your agent and say:
# "Analyse this campaign data and give me a performance report."
# The agent will ask for points and $PRO earned per campaign, then calculate ROI.

No API key? Use the sample data in QUICKSTART.md to test immediately — no setup needed.

---

File structure

campaign-analyser/
├── SKILL.md                          # Core skill — load this into your agent
├── README.md                         # You are here
├── QUICKSTART.md                     # Get running in 5 minutes + sample test data
├── CHANGELOG.md                      # Version history
├── references/
│   ├── METRICS.md                    # Metric definitions, edge cases, alternative signals
│   └── API_REFERENCE.md              # ProductClank API endpoints, what is/isn't available
└── scripts/
    └── fetch-data.sh                 # Full data pull with offset pagination

---

Future capabilities