| name | cold-outbound-performance-analysis |
| description | Analyze cold outbound performance across any outbound sending platform and available CRM or outcome data to identify top-performing campaigns, attribution patterns, copy and offer learnings, audience characteristics, opportunities, deals won, and recommended new campaign sequences. |
Cold Outbound Performance Analysis
Use this skill when the user wants to understand which cold outbound campaigns are working and turn that evidence into campaign strategy, targeting, copy, and optional draft campaign setup.
This skill is CRM-agnostic and platform-agnostic. It can work with HubSpot, Salesforce, Pipedrive, Attio, Close, a warehouse, product events, CSV/XLSX exports, or any other outcome dataset. It can work with Instantly, Smartlead, Lemlist, HeyReach, Salesloft, Outreach, Outbound, Woodpecker, La Growth Machine, Drip, Dripp, Waalaxy, or any other outbound sending platform.
References
Load only what is needed:
references/setup-questions.md for the intake questions and branching logic.references/data-sources.md for CRM/outcome and outbound-platform extraction guidance.references/analysis-workflow.md for attribution repair, matching, ranking, and insight extraction.references/output-formats.md for the expected report formats and campaign brief structure.references/campaign-creation.md for optional draft/paused campaign creation rules.
Core Process
- Start with setup questions before querying data: CRM/outcome source, outbound tool, connector/API/export availability, date range, success events, and whether CRM attribution is attached to contacts, accounts, opportunities, or deals.
- Use the user's answer about CRM attribution to choose the analysis path:
- attributed CRM outcomes available: rank by repaired CRM outcomes first.
- partial attribution: join platform contacts to CRM records and label confidence.
- no attribution: build directional analysis from platform data plus any matched outcomes, and state limits clearly.
- Pull or inspect CRM/outcome data and outbound-platform data using the available connector, MCP, API, or export.
- Normalize the data into campaign, contact/account, opportunity/deal, copy, reply, and rollup views.
- Repair attribution before ranking. Keep raw source fields and repaired source campaign fields separate.
- Exclude or separate muddy attribution buckets before ranking, especially follow-up, nurture, signup, lifecycle, reactivation, meeting follow-up, and other non-source campaign buckets that overwrite original outbound source.
- Rank campaigns by downstream outcomes first: opportunities/deals created, open pipeline when available, deals won, and won revenue. Use platform engagement metrics only as supporting context.
- Extract copy, offer, reply, audience, and CRM/outcome patterns from winners and weak performers with explicit quantitative evidence.
- Produce a clear report with one recommended campaign, targeting/list criteria, draft copy, and notes behind the recommendation.
- Ask whether the user wants to create the recommended campaign if a safe outbound-platform creation tool is connected. Create only draft/paused campaigns after explicit approval.
Output Requirements
Always return:
- data access status and source coverage
- attribution status and caveats
- campaign leaderboard ranked by opportunities, deals won, and pipeline/revenue
- copy, offer, reply, audience, and outcome patterns backed by explicit counts, rates, pipeline, or revenue
- one recommended campaign with targeting filters/list criteria
- draft outbound sequence copy
- next steps, including whether campaign creation is possible with connected tools
Never launch, activate, unpause, schedule sending, or add leads to a live campaign without explicit user approval. Never present internal CRM outcomes as external customer proof.
