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

These features are planned and will be enabled automatically once ProductClank exposes the relevant data through their agent API:

Points and $PRO per campaign Currently, points earned and $PRO per campaign are not returned by the agent API. Once ProductClank adds these fields, API mode will calculate ROI fully automatically — no manual input required.

Real-time ROI tracking With live points data, the skill will be able to track ROI as a campaign runs — not just after it ends.

Season-over-season comparison With historical points data accessible via API, the skill will compare performance across ProductClank seasons and flag whether scoring changes affect campaign strategy.

$PRO token-based ROI Once $PRO balance deltas are accessible per campaign, the skill will calculate a second ROI metric — $PRO earned per credit spent — giving a token-denominated view alongside the points-based one.

Agent leaderboard context With access to leaderboard standing via API, the skill will contextualise your ROI against other agents in the same season — not just your own historical data.

If you are a ProductClank developer and want to help enable these features, the relevant fields needed are: points_earned and pro_earned per campaign in the GET /agents/campaigns/{id} response, and an overall agent stats endpoint.

---

Built for the ProductClank ecosystem

This skill is registered with the ProductClank Skill Registry and follows the Agent Skills standard (v1, Anthropic).

Useful links:

• ProductClank Agents
• Skill Registry — Season 6
• Skill-Leads programme
• Official ProductClank agent skills

---

Author

RAVEN_SPARK — @RAVEN_SPARK7 on X

---

License

MIT