top of page
  • Writer's pictureTiffany

Foundational: How Specificity of Item Needs Impacts Shopper Behavior


As the Senior UX Researcher on Mercari and a UXR lead for Shopper Experience, I noticed our team's needs related to understanding how shoppers who are in various buying needs/specificity differ in behaviors, perceptions, and experience. This project showcases an example of data triangulation and stakeholder management across multiple product teams and departments, and how foundational insights may

Note that, due to NDA, I've purposefully made some details of the project/insights vague.

Research Objective:

Users have varying needs of search specificity when shopping; for example, one may know exactly what they want to search for (e.g. Nike AirMax 90 Size 7.5) , while another may be open for exploration and discovery (e.g. I'm like BTS fandom goods). This study helps gain foundational understanding about shoppers' needs, behaviors, and expectations when in browsing/searching/exploring mode, to help shape designs that will best support shoppers in their mode.

Research Approach:

  • External paper review previously published paper on searching/browsing behavior on eComm sites

  • Hypothesis workshop with stakeholders + UX team to identify hypothesized events related to each mode

  • Collaborate with Data Science team to identify behavioral patterns with three buyer modes (previously identified through research)

  • 45-min 1:1 user interviews to understand intent, needs, and expectations for each mode

    • Each session to include walkthrough + inquiry of past searches + item purchases

  • In-app feedback survey from users on search recommendations


Before jumping into recruitment for primary research, I needed to identify specific behavioral criteria of each buyer "mode". I conducted a hypothesis mapping workshop with product + UX team to map out relevant actions/behaviors that we hypothesize about each buyer mode.

Then worked with Data Science to identify behavioral patterns of various buying modes:

  • Investigate and validate each modes and behavioral patterns through double_log.

  • Define metrics and thresholds that represent each “modes”.

  • Sample extreme patterns of users who are extreme in their modes.

Based on these behavioral patterns we hypothesized, Data Scientist Victor pulled sample users based on extreme representations of each user mode for 1 week time span, and ensured that there was no overlap in sample across groups.

1:1 Interview:

Once the criteria was identified, we recruited the 15 participants for the 1:1 interview, with n=5 in each buyer mode. Each session was split into: background, contextual inquiry (with scenarios on specificity of item needs), and walkthrough of most recent purchases, searches, and discovery.


Tl;dr... This research pointed us to the varying ways our shoppers were using Search based on specificity of item needs.

Research Impact:


While there were immediate takeaways and recommendations for product team, there were also opportunities for discussion and ideation. For example, "how might we allow for easier context switching between interests?"

I brought the stakeholders together, and hosted a workshop that kicked off with a Highlight Reel Watch Party that summarizes interview findings from participants, where teammates and stakeholders were encouraged to jot notes down. Then I led the Crazy 6 activity where we ideated on search-related opportunities which were major insights from the research.

📚My Learnings:

  • Value of partnering with the data team and connecting quantitative + qualitative data

  • The criteria matched from data perspective was not accurate, so the results weren’t so clear cut across mindsets.

  • Loved the different ways I tested out to share out research findings for fast, easy, and accurate digestion of information. Highlight reels should be short and direct!

  • More opportunities for research...



bottom of page