Skip to content

Internationalisation

Every module goes international. The quiz goes first.

Decision: ⚪ idea

Context

The endgame is the largest international platform for political and psychometric self-awareness. The method behind it is not Polish - only the interface and the question sets are. Internationalisation is the work of separating those two things.

Three layers, and only the first is translation:

  • Interface - the product's own strings, handled by translation tooling and shipped per language.
  • Content - questions, orientations, archetypes, descriptions and the copy pools behind checkpoints and the loader. These are written per language, not translated - a joke about a third reading and a question about the Round Table mean nothing outside Poland.
  • Context - the parties, candidates and axes a country actually argues about. Nothing here transfers, and nothing here should.

Quiz goes first, because it is the only module that does not need local staff. A creator in another country can author a quiz for their own politics on our infrastructure the day the interface speaks their language - see community quizzes. Data follows, since datasets and reports are per market. Polls and Media come last: both depend on local sources, local aggregation and people who read the local press.

What has to change underneath:

  • Language becomes a property of a quiz, not only of the interface, so a quiz is authored, browsed and scored in one language.
  • Discovery has to be per language - a Polish quiz surfacing to a Spanish taker is noise, so quality score and the catalogue need language as a filter.
  • Moderation has to be possible in that language - the approval gate assumes we can read what we are approving.
  • The legal frame is per market - EEA rules carry across Europe, everything beyond it needs its own review, and age thresholds differ - see privacy and legal.
  • Domains and search - we already run on two domains, and each language needs to be findable on its own terms.

Opportunity

  • The infrastructure is the thing worth sharing - the editor, the result modules and the scoring took eight years to build, and a country that wants a political compass currently starts from nothing.
  • Creators bring the content we cannot staff - the same argument as community quizzes, applied to a whole country. We supply the platform, they supply the politics.
  • It smooths the election cycle - traffic today rises and falls with Polish elections, and every additional market is another cycle on a different calendar.
  • The non-political side travels furthest - psychometrics and myQuizzes carry across borders far more easily than any political question set, and that is also where the paid tier lives.
  • Open source makes it cheap to adopt - a country can run the stack rather than commission one.
  • One product, many markets - nothing per-country gets built: language is configuration, not a fork.

Risk

  • Moderation in a language we cannot read - the credibility gate is the whole community model, and it does not survive being applied to text nobody on the team understands.
  • Neutrality is culturally specific - a question that reads as balanced in Poland can read as partisan elsewhere, and we would not know until it does.
  • Content debt multiplies - every copy pool, description and result text is written again per language, by someone who knows both the language and the politics.
  • A translated question is not the same question - answers collected in two languages are not automatically comparable, which limits what pooled analysis can honestly say.
  • It splits a team that is already thin - a second market before the first product is finished is how both get half-built.
  • Legal exposure grows with the map - special-category data outside the EEA is a different problem in every jurisdiction.
  • The brand fragments - several domains, several languages and community-authored content in each is a lot of surface for a name whose only asset is trust.