---
name: Microsurvey
slug: microsurvey
category: pattern
status: published
created: 2026-08-21T00:00:00.000Z
modified: 2026-08-21T00:00:00.000Z
definition: A one or two question survey placed inside the product at the moment
  being asked about, answered in place rather than on a separate page.
aliases:
  - name: in-app survey
    source: community
  - name: one-question survey
    source: community
  - name: NPS prompt
    source: community
  - name: feedback mechanism
    source: cloudscape
  - name: was this helpful
    source: community
tags:
  - onboarding
relations:
  contrastWith:
    - rating-prompt
  variantOf: []
  partOf: []
  seeAlso:
    - status-message
    - progressive-profiling
implementations: []
sources:
  - title: "Cloudscape: Feedback mechanisms"
    url: https://cloudscape.design/patterns/
demo: inline
exhibit: false
useWhen: a one-question survey inside the thing it asks about
---

A microsurvey trades length for context. A full survey can ask thirty questions but has
to be answered somewhere else, days later, by whoever still cares enough to click a link
in an email. A microsurvey asks one question, at the foot of the page or the end of the
task it is about, while the experience being described is still in front of the person
describing it. The answer is worth more because it is anchored to something specific, and
it costs a click instead of five minutes.

The canonical instance is the "Was this page helpful?" pair under a documentation page,
and Cloudscape files the whole family under feedback mechanisms. The variants are all the
same shape: three faces after a task completes, a thumbs pair under a generated answer, a
zero to ten scale that a product calls an NPS prompt because that is the number it feeds.
Whatever the scale, the rule is that the question is asked where the thing happened, and
answered there too.

That last part is easy to get wrong. Clicking an answer must not open a page, a modal, or
a new tab: the panel replaces its own question with a thank-you in the box it already
occupied, so nothing under the reader's eye moves and the survey is over. If a second
question is genuinely needed, it goes in the same box after the first, and there are two
questions at most. The moment the reader is taken somewhere to finish answering, the
response rate collapses and the pattern has stopped being micro.

The two neighbours worth keeping straight are both about who the question serves. A
[rating prompt](/rating-prompt) asks the reader to rate the product in public, usually in
a store listing, and is often gated so that only the happy answers get there. A
microsurvey asks about the thing that just happened and the answer stays private, which
is why it can be honest about failures. And like any in-product ask it needs a frequency
cap and a memory: one question per visit at most, never on every page, and never again
once it has been answered or refused, or it turns into [nagging](/nagging) wearing a
survey's clothes.