Project Career Path
Project Career Path
Project Career Path
My role in this project:
My role in this project:
I am the Product Manager on this project working closely with engineers, designers, data scientists, operations and marketing specialists.
I am the Product Manager on this project working closely with engineers, designers, data scientists, operations and marketing specialists.
My role in this project:
I am the Product Manager on this project working closely with engineers, designers, data scientists, operations and marketing specialists.
This is a personal project that uses kinetic motion to illustrate the song "The History of Everything", which gives the song a different visual narrative in motion.
This is a personal project that uses kinetic motion to illustrate the song "The History of Everything", which gives the song a different visual narrative in motion.
Tagging Photo to Connect Interests
Mobile, B2C, Social, UGC
Background
Nice app is a Chinese dominant photo sharing app, 64M in VC funding, 30M registered users.
The "Adding tags on photo" feature is the core feature that introduced Nice to its target users and helped Nice won the competition in the early days. It's an innovative fun interaction to add clickable tags on a photo, which made the tag a part of the composition. And more importantly, the tags make it easier for both Nice to aggregate content and users to consume content by the same interest(same tag). Here are some of the user scenarios:
Tagging food
Tagging the brands
Tagging a play and location
Tag home page
User Research & Pain Points Analysis
The pain points come from both data analysis and user research. During the user research, I worked closely with operation team to recruit 2 groups of users - new signed-up users and active users to Nice’s office. The users participated were aged from 16-25, who are high-school/college students and young adults interested in new trends. Key findings are as follows:
Pain points for new users:
​
-
Hard to understand how to use the tag feature
-
Don't know what to write on a tag - data shows that many new users write meaningless tags like "。" or "."
-
Get confuses with the 3 types of tag, don't know why do they need to choose a tag type first
-
The black overlay takes over the whole screen which creates an interrupted feeling
Pain points for active users:
​
-
Too many back and forth steps in:
-
adding multiple tags on one photo
-
adding one same tag on multiple photos
-
-
Selecting a tag type first interrupt their thoughts(what to write on a tag) in mind. Regular users often know exactly what to write before using the tag feature.
-
Want to better manage their history tags. They are not able to easily find the tags they use often across different tag types
Select a photo to add tag
Select a tag type from:
Regular/Location/People
Regular tag:
Input box/history
Location tag:
Input box/history
People tag:
Input box
Concept Exploration
Solution and Case Studies
Here are 3 design solutions I’ve developed to fix the pain points that I learned from data and user research:
1. Interaction improvement - Reduce friction by shortening the adding tag steps from 3 to 2 steps, more steps saved when adding multiple tags on one photo or one tag on multiple photos!
Select a photo to add a tag
Tag landing page
2. Information architecture improvement - Integrate content across 3 types of tags in one place, no need to select a tag type first!
​
3. Make relevant and smart recommendations - build the whole recommendation experience by introducing new smart trending tags via visual search tech, nearby location tags, and easy-to-use history tags
Tag landing page
Recommendations
​
-
"Smart Tag" that auto-recommends trending tags on Nice, using image recognition technology. For example, if it's a breakfast dish in the photo, 3 trending tags like "breakfast" "breakfasttime" "healthyfood" would be recommended. This feature turned out to be very helpful to the new users from the results.
-
Nearby location tag recommendation - this feature recommends the nearby locations based on the location info of the photo
Input box - write a new tag
Tag history
Includes all types of tags ranked by recency
Typing - Recommendation
Typing - Switch to location tag
Typing - Switch to people tag
-
Users now don’t need to choose a tag type first, the tag type choices won’t show up until a user write something. Users are able to switch to places or people tag as optional when needed.
-
When typing, the new design enables typeahead recommendation of all tag types(compared to recommend one type of tag at once previously).
-
The new design also comes with an advanced ranking algorithm that gives a the relevant tag type a higher ranking. For example, if it’s a location type of keyword like “The Great Wall”, then the location tag of “The Great Wall” will be ranked higher than that of the regular tag. Same logic to people tags and regular tags.
Testing and Collecting User Feedback
Since this is a big change to Nice's core tagging feature, besides looking at data, I wrote a questionnaire to collect feedback after the initial ramp - 5% new users and 5% regular users.
​
The questionnaire was sent to both new users and regular users asking behavior questions as well as seeking subject feedbacks/suggestions. This feature had been ramped up after I saw positive feedback from both two user groups.
Key Results
Average number of tags per photo and posting conversion rate was lifted greatly.
New user treatment vs control
Regular user treatment vs control
Overall impact
Learnings
Solving the right problem for the right user groups that meets the company’s current business goal is essential especially when resources are limited.
User pain points and needs varies across different user groups, resources are always limited in startup settings, therefore the key is to solve the right problem for the right user group that meets business needs for the company. In this case, there are new user and active user groups, some of their needs are aligned like reducing friction in adding a tag in posting photo flow, while others are not. For example, active users want to better manage their tags as well as adding some customization features, while new users need recommendations to get started and lower the barrier to generate their first tag. Since the company’s business goal at that point is to acquire more new users so I prioritize the needs of new user groups higher than the other group in order to achieve that goal. If it’s increase engagement of the existing user, then the priority would be reversed.
​
Thinking out of the box, leveraging technology in solving product design problems.
Take the Smart Tag feature I designed for an example, the initial thought was to recommend daily trending tags on Nice for users to select and add, that’s definitely better than no recommendation at all compared to the previous design. However, the problem in this solution would be lack of customization which will result in Matthew Effect over time. I pushed myself to think harder by borrowing engineer colleague’s brain which eventually lead me to this Smart Tag idea that using image recognition technology to recommend relevant trending tags based on photo uploaded by user.