signals-scout-surveys

Monitors PostHog surveys for score regressions, abandonment spikes, and recurring open-text themes.

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

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Teams running PostHog surveys rarely notice when an NPS score drops, a survey's response rate craters, or five users independently complain about the same feature until weeks later. This Skill continuously watches active surveys and surfaces validated anomalies and themes as actionable inbox reports.

Core Features & Use Cases

  • Anomaly detection on survey metrics: Detects NPS/CSAT/rating score regressions, response-volume drops, dismissal-rate spikes, and targeting drift using tiered statistical bars matched to each survey's baseline volume.
  • Open-text theme aggregation: Clusters free-text responses into recurring themes (e.g., five users complaining about the same checkout step) and files reports with verbatim quotes as evidence.
  • Report lifecycle management: Authors new reports or edits existing ones via the signals report channel, deduplicates against prior runs using a persistent scratchpad, and routes findings to the right reviewer.
  • Use Case: A product team's NPS survey drops from 32 to 24 over two weeks while open-text responses converge on dashboard slowness; the scout validates the regression against the trailing baseline, cross-references a recent deploy, and files a report routed to the survey owner.

Quick Start

Ask the agent to run the surveys scout to check all active PostHog surveys for score regressions, response drops, and recurring user feedback themes.

Frequently Asked Questions about signals-scout-surveys

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

FAQPage Schema
How do I detect NPS or CSAT score regressions in PostHog surveys?

Query survey sent events over the last 30 days, coalesce the id-based and index-based response keys, dedupe by $survey_submission_id, and compare the recent 7-day average against the trailing 21-day baseline. High-volume surveys need a 10% scale drop on 30+ responses; low-volume surveys need a 15% drop against a 60-day baseline.

How do I aggregate open-text survey responses into themes?

Pull recent non-empty open-text responses with the dedupe SQL, then read them for convergence on a noun phrase or feature name. A theme is report-worthy when 5+ distinct respondents converge within 14 days, or 3+ converge and match a recent deploy or flag change.

Why does my survey response SQL return no responses?

PostHog writes answers under two key schemes: id-based $survey_response_<question_id> and legacy index-based $survey_response or $survey_response_<n>. Querying only the id-based key reads as empty on legacy surveys, so always coalesce both keys.

Does a high survey dismiss rate always mean a problem?

No. Interpretation depends on the survey type: popover dismiss spikes are genuine fatigue signals, but widget surveys naturally run 50-70% dismiss rates since shown only fires after a user clicks the trigger. Check the survey type via survey-get before flagging.

When should a survey finding not be filed as a report?

Skip filing when the survey is at the end of its scheduled window, the sample is under 30 responses for score drift, themes are evenly split between praise and complaints, a scratchpad entry already marks the theme as addressed, or responses are internal test submissions.