HUDDLE - SEARCH FUNCTION
UX Design l UI Design l 2019
Huddle is an online document collaborating SaaS business and each organisation or team and purchasing their high-secured cloud storage. However, due to individuals' different organising behaviour, it is quite reasonable that folders and files are moved or restructured by workspace admin, or difficult to find particular files if new to the workspace.
The baseline question of this project is: How should I help people find things?
After one year of research by survey and data analyst, I joined this Search project to re-design the search function.
Primary goals
-
Improve the search function flow and user experience based on survey and data analyst findings.
-
Feature enhancement from the client’s request to support legal and education use cases.
-
Build hypotheses and define tracking events for the next phase and A/B testing.
-
Reduce the drop rate of searching and increase user retention on the search function.
PROJECT RESULTS
Search function engagement before and after release (from May 2019 to June 2019)
Avg. unique users
+ 20%
Existed user retention rate
90%
New user retention rate
80%
My role
By following findings from previous researchers, I designed an end-to-end task flow, wireframe, and design system and defined the design criteria.
Meanwhile, I compared and contributed to two other related projects: 1. My space 2. Workspace overview, to fulling, understand the user flow on finding things in different scenarios. With the knowledge of the whole user journey, I’m able to contribute to the product vision, hence re-define the product structure by design and improving the information architecture.
FINDING THINGS
Scope of the redesign project
The entire "finding things" problem area includes the below projects:
-
Bookmark
-
Pin files & folders
-
Favourite workspaces
-
Personal dashboard
-
Search
-
and so on
Although the task flow link from feature to feature, based on user scenarios, today, I'll only introduce the search function as an example of using quantitative research findings to support design decision.
SEARCH IMPROVEMENT
Problem statement
Uses have difficulty in finding the correct file by using the search function. The reasons are:
-
Didn't realise that there was a full-page search result, and stopped at the instant search result.
-
Many files have the same or similar name
-
Didn't know or forget the file name
-
The target item is ranking in a shallow position.
-
Get lost when clicking a wrong item and trying to go back to search results
-
Not able to find an item which was created over three years due to the low ranking
"Hold the line!
Even though users always ask for more filters, the actual thing they want is a better suggestion. It's a designer's job to understand the prioritisation, and work closely with the tech team so we can provide better ranking and better logic."
Challenges
There are some tech restrictions which push UX flow from the best practice.
Examples are:
-
Not able to mix result types and rank one list with the best relevancy.
-
Not able to merge the suggested results (combining with bookmarked files and recent files) to the final result list.
User scenarios
-
Finding an old document that I created/uploaded a while ago
-
Searching for relevant new materials, which are created by people I might know
-
Finding documents, I viewed/contributed before but are now being moved
-
Searching for all relevant information for a specific topic, theme or project
-
Finding documents that people review and comment on, which are pertinent to my work, or they mentioned me.
Using data, so the design is more accurate
Data can tell us the most frequent click file type, the data range of a file, and how often users need to search for multiple items in a short time. We then ask the question about the reason that people drop off and the scenario that the user needs to look for an old file. Later I build the priority and needs of tracking particular clicks, so we can verify the hypothesis or answer other unknown behaviour.
Prototype
Design process
Accessibility
-
The goal is to pass AA level
-
Colour and text size ratios were tested during design. The brand colour is too bright so we need to use a secondary colour together.
-
All the keyboard behaviour were fully considered and written down as design criteria.
Learning
I was excited about this project because I can use data and survey results to make a decision. My wish was able to use data journey on the design, but unfortunately, we were not tracking each click on the product, so the next approach is to build hypotheses and track clicks for phase 2 iteration.
An interesting idea: The product and UX team at Huddle agree on fixing visuals and experience across products, step by step. We are not looking for a perfect solution for each deliverable, but aiming for 80/100 first so we could move forward to the next project.