signals-scout-product-analytics

Detects conversion and retention regressions in saved PostHog funnels, retention, and lifecycle insights.

713|118|Updated Aug 11, 2020
One-click install
npx skills add https://github.com/PostHog/posthog-foss --skill signals-scout-product-analytics
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: signals-scout-product-analytics
Source: https://github.com/PostHog/posthog-foss/tree/main/products/signals/skills/signals-scout-product-analytics
Command: npx skills add https://github.com/PostHog/posthog-foss --skill signals-scout-product-analytics

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Product teams rarely notice when a saved funnel's conversion rate or a retention cohort's curve quietly regresses until someone manually opens the chart. This Skill proactively monitors a team's behavioral flows in PostHog and files a report when a derived rate drops while entrant volume holds steady.

Core Features & Use Cases

  • Watchlist-based flow scoring: Maintains a durable scratchpad watchlist of saved funnel, retention, lifecycle, and stickiness insights, re-scoring each due flow against its seasonality-matched trailing baseline.
  • Regression detection with denominator checks: Flags a rate move only when conversion or retention deviates from baseline while entrant volume holds, filtering out capture drops, running experiments, seasonal swings, and partial windows.
  • Report authoring and deduplication: Authors or edits inbox reports via the signals report channel, tracking report pointers per flow and per rate so persistent regressions become edits rather than duplicates.
  • Use Case: A PM asks in a weekly review whether the activation funnel is still converting; this scout has already detected that step-2 conversion fell from 62% to 48% on steady entrants and filed a routed report with baseline evidence.

Quick Start

Ask the agent to run the product-analytics signals scout to check the team's saved funnels and retention insights for any conversion or retention regressions this week.

Frequently Asked Questions about signals-scout-product-analytics

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

FAQPage Schema
How do I monitor PostHog funnels for conversion regressions automatically?

This scout re-scores saved funnel insights each run, comparing the latest complete window's step-to-step conversion against a trailing baseline band. It files a report only when the rate drops while entrant volume holds steady.

What counts as a real product analytics regression versus noise?

A real regression is a derived-rate drop, such as conversion or retention percentage, with a steady entrant denominator. Drops where entrants also collapsed, running experiments, seasonal swings, and partial windows are disqualified as noise.

Does this scout work without saved funnel or retention insights?

It can infer at most one activation funnel from product intent milestones and top events using path analysis, held to a higher emit bar. If the project has no behavioral events at all, it closes out without filing reports.

How does the scout avoid filing duplicate reports for the same regression?

It stores report pointers in scratchpad keyed by flow and affected rate, so a persistent regression becomes an edit with appended evidence rather than a new report. Distinct rates on the same insight each get their own report.

What MCP tools and permissions does the signals scout require?

It needs read-only PostHog analytics tools such as query-funnel, query-retention, query-lifecycle, query-stickiness, query-paths, and execute-sql, plus signal_scout_internal:write for scratchpad and signal_scout_report:write for authoring reports.