ᱛᱚᱞᱟᱥ ᱵᱤᱵᱷᱚᱜᱽ ᱟᱜ ᱠᱟᱹᱢᱤ
ᱟᱢᱟᱜ ᱰᱤᱜᱰᱦᱟ ᱪᱮᱰ ᱠᱟᱱᱟ ᱚᱠᱟ ᱟᱢ ᱵᱟᱝ ᱡᱩᱴ ᱰᱟᱰᱮᱟᱜ ᱠᱟᱱᱟ?
ᱦᱳᱰ ᱠᱳᱰᱳ ᱵᱟᱦᱨᱮ ᱥᱟᱨᱪ ᱤᱱᱡᱤᱱ ᱨᱮ ᱠᱟ ᱣᱤᱠᱤᱯᱮᱰᱤᱟ ᱟᱨ ᱣᱤᱠᱤᱢᱤᱰᱤᱟ ᱨᱮ ᱢᱮᱱᱟᱜ ᱵᱤᱥᱟᱭ ᱥᱟᱨᱪ ᱫᱟᱰᱮᱟᱜ ᱠᱟᱱᱟ ᱠᱚ
The problem is that people tend to go back to to the external service for additional searches, even though they're already on one of our projects.
The Discovery Department wants to create a better experience in discovering knowledge. We create more accessible searching and discovery mechanisms. Projects like improvements to search, project portals, maps,Wikidata query service and more.
There is also an opportunity to explore surfacing information from sister projects to enhance the discovery of that knowledge for projects that have less visibility.
Lastly, the team is providing a foundation for product development that is data driven as well as user driven to iterate to the useful services and features our users desire and need.
What are you trying to understand?
The Discovery Department tracks four core metrics (also known as key performance indicators) for search:
- Zero results rate for search - If users receive no results, it means we've not been able to help find what they're looking for, so we measure the zero results rate.
- User engagement with search results - If users do not click on results, then we haven't given them the results they wanted.
- Load times - The faster our search works, the better.
- API use - It's important that apps and third parties can search our site too.
You can see the full range of metrics that we track on the Discovery Department's search dashboard.
ᱪᱮᱰ ᱪᱤᱠᱟ ᱟᱟ ᱚᱱᱟ ᱯᱚᱨᱮ ?
The team and users will post ideas for deliberation, and will collectively come up with proposals.
How will this affect other products that the Wikimedia Foundation is developing?
The research carried out will help bring more understanding to search and discovery mechanisms across all platforms, and user flows from readers to editors and will inform decisions made on how to improve those mechanisms on desktop, mobile web, and mobile apps, as well as in specific products like VisualEditor.
We also are exploring API usage, best practices, mix of content from inter-wiki projects like Wiktionary, Wikivoyage, Wikimedia Commons and more, and utilization of open data sources like OpenStreetMap to expand contextual knowledge discovery.
We will, of course, be publishing our research, so that it may be read and taken into account by the broader movement and other interested parties.
What does your overall strategy look like ?
In late 2015 the Discovery department set out a 3 year strategy plan.
- Year 0 - Look inward and improve the search experience across our projects
- Year 1 - Look outward and see if we can incorporate new data streams and public curation models for relevance
What does year 0 include ?
We call year 0 Discovery because we are focused on learning and understanding user pathways and appreciation for other knowledge sources.
What does year 1 include ?
Potential ideas that we need your feedback on:
- Identify pathways for the community to improve relevance via Wikidata
- Actively highlight difficult to find knowledge and empower the ability to surface it in search, reading and editing flows
- Research open sources of knowledge to continually strengthen the legitimacy of our content through curation by humans and machines
This feels like a huge long term project. Is it?
Our users interested in search request a lot of improvements: inter wiki, multi-lingual, media search , improving UX, improving search relevance, and others. The Discovery Department aims to improve search in these areas, and that will take a lot of time! During this process, we will continually re-evaluate our plans on a quarterly and annual level to assess our impact and hold ourselves to the same standards as any other team at the Wikimedia Foundation.
How does this align with strategy?
- Relevancy, accuracy and trustworthy ratings on index entities
- Extended context to geospatial, temporal, multimedia and relational paths of knowledge
- Display Inter-wiki projects (internal) and potentially open data sources
- Mobile, voice, and modern consistent interface opportunity
- Multiple-lingual and global respective experiences and results
See also: Community Strategic Consultation
Is all your work open source?
Yes. All of our code is contained in public repositories, and falls under the same licensing as MediaWiki. See this list of all the code repositories the department supports.
How do you know if we are succeeding for our users?
Will there be any element of human curation?
- We'd like to explore this and need your help on our RFC to think through how to do it right.
I'd like to see a list of search results. Can the department provide this information?
This is a common question by editors and researchers alike. The idea of providing a list of queries where a page does not exist was researched and determined to be to difficult to accomplish with our resources. The biggest concerns were ensuring privacy, the difficulty in creating a usable list out of many junk queries, and the time it would take to create such a list would be costly.
What was the Knowledge Engine ?
I want to help! How can I get involved?
We'd love the input of anyone who wants to join us in building and improving search. Here's how you can do that: