Lucky Search UX
Created by: almeidapaulooliveira
'Lucky search' eliminates the 'no results' problem users often have when they first encounter our literal search default.
This design initiative addresses the need to create an easy-to-use UI to interact with 'lucky search' suggestions. It is the beginning of a more significant effort to improve our searching experience and lays a foundation for iteratively improving the quality of our results.
Design process (Size: M)
- Estimate: Three weeks
- Weighting factors:
- MEDIUM risk
- MEDIUM clarity
Discover
'Lucky' and 'standard' search modes
-
Low-fidelity (Figma board) exploration and discussion.
In the discovery process, deciding how effectively to switch between search modes, many discussions led to a different understanding: 'Lucky search' is not a search mode per se, nor is the standard search mode. Instead, all searches are implicitly standard, and 'lucky search' is an extended query interpretation feature applied on top of it, with the possibility of being enabled and disabled.

Define
Goal:
Decision on UX for the extended query interpretation interaction model.
Problem statement:
From recent research, we've learned how users struggle to discover and use search-related features. 'Lucky search' aims to improve the searching experience. Still, if not well explained, it could lead to users not understanding how to use it (enable, disable, etc.) and more frustration. Therefore, this design will create a simple and easy-to-use UI with the addition of a straightforward education effort.
Design challenges:
Having discovered a new understanding of 'lucky search,' we get back to the following challenges:
- How might we explain the extended query interpretation functioning to users?
- How might we present the extended query interpretation suggestions to users?
- How might we educate users on discovering how to toggle it between enabled and disabled?
Design
-
Hi-fi visual design (Figma file) and discussions. -
Interactive prototype. -
Loom. -
Designer review.
Testing
-
User testing plan.
Deliver
-
Announce in #progress. -
Team review. -
Design team review.
Measure
-
Define metrics.