- Home
- Skills
- Sales & Lead Gen
- lead source and attribution tracker
Works with the AI Agents you already use
lead source and attribution tracker
Marketing teams struggle to justify spend when lead sources are inconsistent or windows are poorly defined.
$2.99
lead source and attribution tracker
Example session with this skill installed
Analyze attribution for our Q3 leads. Use a 30-day default window and first-touch model. Check our recorded sources against the touchpoint log. Here are the files for leads, sources, and touches.
- Read your context and instructions
- Compiled the lead source
I have completed the attribution analysis for Q3 using the first-touch model. All calculations were performed via the deterministic script.
Lead Source and Attribution — Global Marketing, 2023-10-25
1. Scope, sources and assumptions
- Analysis Date: 2023-10-25
- Model: first_touch
- Scope: all_opted_in
- Access Scope: Demand Gen Team
- Assumptions: 30-day default window applied to E3, E4 (default).
2. Windows and funnel
| Source | Channel | Approved | Window (days) | Window origin |
|---|---|---|---|---|
| E1 | Paid Search | True | 30 | user_supplied |
| E2 | Webinar | True | 45 | user_supplied |
Lead Funnel:
- Supplied: 1,200
- Excluded (Consent/Scope): 150
- Analysed: 1,050
- Attributed: 980
- Unattributed: 70
3. Credit by source and channel
| Source | Leads credited | Credit | Share of attributed |
|---|---|---|---|
| E1 | 540 | 540.00 | 0.5510 |
| E2 | 320 | 320.00 | 0.3265 |
| E3 | suppressed | suppressed | suppressed |
4. Model comparison
| Model | E1 | E2 | E3 |
|---|---|---|---|
| first_touch | 540.00 | 320.00 | 120.00 |
| last_touch | 410.00 | 450.00 | 120.00 |
5. Recorded source check and proposed actions
- Discrepancies found: 42
- Actions:
- ACT-001: Change Lead L-882 from 'Organic' to 'Paid Search' (Draft, requires human approval).
- ACT-002: Change Lead L-901 from 'Unknown' to 'Webinar' (Draft, requires human approval).
6. Warnings and exclusions
- W-CON-001: 45 leads excluded due to missing consent flags.
- W-SUP-002: Source E3 results suppressed in channel view due to small cell size.
7. Interpretation and items for human review
- Paid Search (E1) shows higher efficiency under first-touch than last-touch (computed result).
- 7.14% of opted-in leads remain unattributed within the defined windows (computed result).
- Management decision required: Review 42 recorded-source discrepancies before CRM sync.
This analysis supports, and does not replace, professional judgment and management's decision.
Next steps
- Review the
proposed_actionslist to verify source changes for the 42 flagged leads. - Adjust the
window_daysin your source catalog if the 70 unattributed leads fall just outside the 30-day limit. - Run a linear model comparison to see how credit distributes across middle-funnel touchpoints.
Connects securely to your tools. The creator never sees your data.
About this skill
Marketing teams struggle to justify spend when lead sources are inconsistent or windows are poorly defined. This skill connects leads to touchpoints using verifiable attribution models while enforcing data governance. It calculates credit distribution across first-touch, last-touch, linear, and position-based rules without modifying your live data.
What it does
- Multi-model attribution credits sources using first-touch, last-touch, linear, or position-based logic.
- Window enforcement filters touchpoints against documented attribution windows per source.
- Governance filtering automatically excludes leads based on missing consent, out-of-scope access, or provenance gaps.
- Integrity checking compares recorded lead sources in the CRM against actual touchpoint evidence.
- Draft generation produces proposed CRM updates as drafts for human review rather than executing live writes.
How it works
- Cataloging maps your acquisition sources to specific channels, status levels, and attribution windows.
- Schema mapping organizes leads and touchpoints into the required JSON format using pseudonymous IDs.
- Execution triggers the deterministic Python script to calculate credits and apply suppression rules.
- Validation identifies warnings, small-cell suppressions, and discrepancies in recorded sources.
Frameworks & tools
The skill relies on a self-contained Python execution environment. It interfaces with CRM data via JSON schemas for leads, touchpoints, and source catalogs.
Why this beats prompting it yourself
Manual prompting often fails at complex fraction math across thousands of touchpoints and ignores small-cell suppression for privacy. This skill uses a deterministic script to ensure calculations remain consistent, auditable, and compliant with your defined attribution windows.
Use cases
- Compare how different attribution models shift credit across paid and organic channels.
- Audit CRM data to find leads where the recorded source contradicts touchpoint logs.
- Calculate marketing impact while strictly excluding non-consented lead data.
- Generate batch update drafts for lead source fields based on evidence.
Known limitations
Does not perform lead scoring, pipeline velocity forecasting, or duplicate merging. It cannot provide legal opinions on consent validity or calculate financial ROI/CPA.
Requires Python 3 environment for script execution. No live CRM writes; all changes are drafts. Leads must be pseudonymous. Small-cell suppression applies to results under a count of 5.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 2 days ago