niche-signal-discovery

Compute Laplace-smoothed lift scores from website content and job listings.

Updated May 22, 2026
One-click install
npx skills add https://github.com/georgeportillo/mitratech-sushidata-plugin --skill niche-signal-discovery-georgeportillo
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: niche-signal-discovery
Source: https://github.com/georgeportillo/mitratech-sushidata-plugin/tree/main/skills/niche-signal-discovery
Command: npx skills add https://github.com/georgeportillo/mitratech-sushidata-plugin --skill niche-signal-discovery-georgeportillo

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires sushidata, websearch, webfetch, apify, fullenrich, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill identifies niche first-party signals that differentiate successful from unsuccessful sales accounts, providing insights for account scoring and prospecting.

Core Features & Use Cases

  • ICP Signal Analysis: Analyze closed won and lost customer domain lists to extract differential signals.
  • Differential Analysis: Compute Laplace-smoothed lift scores to identify distinguishing factors.
  • Report Generation: Generate reports with top signals, evidence, and interpretations.
  • Prospect Generation: Find top 10 net-new prospects and provide optional contact discovery.
  • Use Case: Use this Skill to uncover signals that differentiate successful buyers from non-buyers, enabling targeted prospecting and account scoring.

Quick Start

Use the niche-signal-discovery skill to analyze the provided won and lost customer domain lists.

Frequently Asked Questions about niche-signal-discovery

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I identify ICP signals that differentiate won accounts from lost ones?

ICP signal discovery analyzes closed won and lost customer domains using multi-page website content and job listings to compute Laplace-smoothed lift scores for differentiating successful sales accounts.

How does Laplace-smoothed lift scoring work for account scoring?

Laplace-smoothed lift scoring calculates the statistical probability that specific website or job listing signals appear more frequently in won versus lost accounts, providing a mathematical differentiator for account scoring.

Can I use WebSearch and Apify to extract prospecting signals from job listings?

Yes, this analysis requires WebSearch, WebFetch, Apify, and FullEnrich APIs alongside Sushidata swarm to extract multi-page website content and job listings for differential signal analysis.

How do I find net-new prospects using differential account analysis?

Differential account analysis generates top 10 net-new prospects by applying extracted niche signals to new domains, with optional contact discovery provided through FullEnrich integration.

Do I need Sushidata swarm to run ICP analysis on customer domains?

Yes, Sushidata swarm is a required dependency alongside WebSearch, WebFetch, Apify, and FullEnrich APIs to process customer domain lists and execute the ICP signal analysis workflow.

What is the best way to score sales accounts using first-party website signals?

The best way to score sales accounts is extracting niche first-party signals from multi-page website content and job listings, then computing Laplace-smoothed lift scores to identify distinguishing factors between buyers and non-buyers.