One of the Discovery team's KPIs, and a crucial goal for this quarter, is measuring user satisfaction with search. We've started thinking about implicit measures for doing this and how we might validate them. We would really appreciate it if Halfak, with his background in session reconstruction work and understanding our users, could take a look at our proposed hypothesis around where we might go about measuring user satisfaction with search. The documentation is at https://meta.wikimedia.org/wiki/Research:Measuring_User_Search_Satisfaction
In particular we're looking for:
- Does this seem viable?
- If not (or even if so!) what can you suggest to improve it?
- What should we be thinking about that we're not?
This shouldn't take more than a few hours - although obviously I'd also love for this to expand into an active collaboration on the data analysis and hypothesis testing, if Halfak concludes there's stuff in this project that makes sense for research to be thinking about and benefits the systems we're building.