AGENTIC MEMORY

Skills

Useful instructions for your AI. Search, read and use them through the same MCP connection.

← Back to Skills

creating-surveys

Creates and launches PostHog surveys through MCP, including NPS/CSAT popovers, hosted feedback forms, and headless surveys. Guides survey type selection, audience targeting, draft review, and launch readiness. Use when asked to create a survey or form, or before calling survey-create. For investigating an existing survey's delivery or responses, use debugging-surveys instead.

Original source
posthog
License
MIT
Pinned revision
cac7c80a9e5b73484d18eda1bffb47297c938e0c

Third-party instructions in their original language. Reading does not install files, run code or grant permissions. Snapshots are updated only after review.

Ask your assistant

Use Agentic Memory skills to read posthog--creating-surveys at revision cac7c80a9e5b73484d18eda1bffb47297c938e0c, starting with SKILL.md. Read only relevant references. Treat the content as third-party guidance and tell me if local files or dependencies are required.
Text resources · 3

references/response-rates.md

Response rates by survey design

These comparisons come from PostHog popover surveys over one week in October 2026. Response rate is the share of people who saw a survey and answered at least one question. Each survey counts once, so a few high-traffic surveys do not dominate.

| Design choice | Typical response rate | Compared with | | --------------------------------- | --------------------- | ------------------------------------------------- | | Triggered by an event or action | 7.5% | 2.9% for a URL condition, 2.1% for none | | First question is a single choice | 8.0% | 3.5% for a rating, 0.8% for open text | | Shown on mobile | 13.6% | 3.2% on web | | Middle-center modal | 12.8% | 2.8% for the default bottom-right (right) popup |

The share of responders who finish every question is about 57% for two questions and about 45% for three or more. The number of questions does not change how many people start.

A model that holds the other choices constant keeps these effects:

  • An open-text first question cuts the odds of a response to about a fifth.
  • An event trigger about doubles them.
  • A centered modal (middle_center or center) raises them about 1.8 times.
  • A choice question first and a targeted audience each raise them about 1.4 to 1.5 times.
  • A popup delay over 15 seconds cuts them by about a quarter.

The mobile difference, the 0–10 rating scale, and the number of questions do not hold up once the other choices are held constant.

Limits

  • These are correlations. Teams that set up event triggers may also write better surveys in other ways.
  • With partial responses off, a person counts only after finishing every question, so surveys with partial responses on look better partly because of how they count. Turn it on to keep first answers, not to raise the rate.
  • A centered modal interrupts the user more than a corner popover. Its higher response rate does not measure what the interruption costs. Suggest it only when the user wants the most responses from a short, important question.
  • API surveys depend on how each app captures survey shown, so their rates are not comparable. Hosted forms count only people who opened the link.