Skip to content

Post-survey module

Extra question sets appended after a quiz, to read the current mood.

Decision: ⚪ idea

Context

A post-survey is a short set of questions shown once a quiz is finished - the para-poll. The quiz measures political values, which move slowly; the post-survey asks about current affairs, which move every week.

  • Technically a quiz - authored in the editor with the same schema, only served differently.
  • Attached by the front-end - the questionnaire ends, and the front-end appends the post-survey as extra questions before the result screen closes the session.
  • Optional - it is offered, never forced. The user has already finished what they came for.
  • Two payoffs for the taker - a bit more insight about themselves, and access to how everyone else answered.
  • Rotating - which set is attached changes over time, so we can always be asking about what is current.

The answers are collected for myPolitics data - live analysis, internal analysis and reports.

Opportunity

  • Reach we already have - a question added today rides on existing quiz traffic and collects thousands of answers with no separate promotion, no panel, no budget.
  • Fresh data on top of stable data - the quiz says what someone believes, the post-survey says what they think about this month. Joined, that is a dataset nobody else in Poland has.
  • Cheap - no new product surface. A set is written in the editor and attached.
  • Something back for the taker - seeing how the country answered is a reason to take the survey, and a reason to come back for the next one.
  • Feeds the credibility layer - reports and analysis need current data, not just an annual quiz.

Risk

  • Cost at the finish line - extra questions sit exactly where the result and the share are, which is the worst possible place to add friction.
  • Our audience is not the population - a fast, cheap sample is still a skewed one, whatever demographics we attach to it.
  • Wording is the whole game - current-affairs questions are where neutrality is hardest to hold and easiest to be accused of losing.
  • Consent boundary - these answers are collected in a different context than the quiz, so they have to sit inside the same anonymisation rules - see exports.
  • Ragged time series - rotation means questions come and go; comparing periods only works if the wording stays stable when it is reused.