As a reader on Wikipedia for intrinsic learning, I want to satisfy my curiosity by navigating to recommended articles.
Hypothesis: displaying links to articles in the empty state of search on wikipedia.org and article pages will increase internal referrals on both desktop and mobile Wikipedia.
Requirements:
* Use the Codex [[ https://doc.wikimedia.org/codex/main/components/demos/typeahead-search.html | TypeaheadSearch component ]] in both Vector and Minerva
* Allow one or more recommendation APIs to contribute to the recommendations list at the same time.
* As soon as a user starts typing, remove recommendations and display results for their query
Experiment parameters:
TBD
Qualitative testing
Potential metrics:
- internal referrals
- time to query abandonment
- clickthrough rate
- session depth (time x number of articles, i.e. does this promote rabbit hole reading?)