Jan 2017 - Mar 2018
sketcch.png
invision.png

Feeling embarrassed when using dating apps? Tired of swiping? No longer want to waste time on unfit matches? Try Ardent! Ardent is a smart dating app that uses artificial intelligence to deliver high-quality matches curated specially for you based on physical attractiveness, personality and interest, and social background. When leading you to your perfect match, Ardent also highly values your information security and privacy. Instead of being selected by others from a pool, you are a gift that shows up to your match's life.

Ardent evolved from my class project to a commercial product due to my passion and its market opportunities. In April 2017, I joined Altum as a user experience designer to develop this idea by integrating artificial intelligence with user-centered approaches. Within a year, I led four design iterations and successfully released the app to the Apple Store and Google Play. 

OBJECTIVE

Our team aimed to help young singles between the age of 21 and 35 find their partners with long-term relationships for marriage purposes through artificial intelligence. 

USER RESEARCH

By conducting surveys, contextual inquiry, and qualitative interviews, we concluded our findings into four main challenges which we were going to address in our product.  

216

survey

20

contextual inquiry

62

qualitative interviews

Challenge 1. users found it difficult to find serious long-term relationship with most of existing dating apps.

71% research participants were looking for a serious relationship but only 10% of them believed that existing dating apps were helpful.

55% research participants who were dating app users regarded "swipe" as a way to kill leisure time or as a kind of amusement. 

86% research participants who were dating app users usually took no more than 2 minutes to learn their matches.

Challenge 2. users complained that it was hard to find their ideal matches though they had spent so much time.

69% participants who were dating app users believed that matches given to them were mostly random.

32% participants who were dating app users complained that it took a long time to find a prospective match. 

Challenge 3. users had troubles to make the conversations on dating apps effectively and efficiently. 

74% research participants who were dating app users indicated that they didn't know how to initiate the conversations when they matched with someone. 

Male participants who were dating app users reported that the average response rate for their first messages was lower than 10%.

22% female participants who were dating app users complained that they usually got many messages from males but only replied a few of them.

Challenge 4. users were worried about their personal safety and information privacy when using dating apps.

40% participants didn't want to use dating apps because they felt embarrassed or they were uncomfortable to be selected.

60% participants were somewhat worried about their information security and privacy when using dating apps. 

48% participants were worried about their own safety if they go on the actual date for the first time with people meet on dating apps. 

EXPERIENCE MAPPING

We mapped the experience of a couple from the first encounter to marriage and found out three crucial factors that impacted the courtship: physical attractiveness, personality and interest, and social background.

Physical Attractiveness

People in general choose to pursue relationships with those who are attractive to them. Physical attractiveness often serves as a gatekeeper directing people to learn more about their partners. 

Personality and Interest

The match of personality is important to ensure that the mutual attractiveness is able to stay long. Mutual interests can create common languages between two people. 

Social Background

Social background refers to the objective conditions (e.g. education, occupation, income) of a couple. Those conditions play important roles in the level of happiness during the relationship.

Dating Experience Mapping.png
IDEATION

During the 4th design iteration, we defined nine features to target the above four challenges: facial analysis algorithm, personality analysis algorithm, background analysis algorithminterest comparison, match limitation, match expiration, conversation assistant, feedback system, and designated meetup location.

Facial Analysis Algorithm

85% of interviewees somewhat cared about the appearance of their matches. With the facial recognition system, Ardent developed an AI that had learned facial features from a thousand couples. The facial analysis AI now could predict whether two users had physical attractiveness at a 75% success rate and paired our users accordingly.

Personality Analysis Algorithm

The personality analysis AI currently took two parts of data into account: the result of the personality test (direct input from users) and users’ behaviors showed up when users were using the app (analyzed by system). The given match would be based on users’ preferences on personality and feedback from successful couples.

Background Analysis Algorithm

Ardent paired two users with similar social backgrounds at first. Location, nationality, occupation, education, income, etc. would be considered at the backend when the AI matched users. In other words, such kind of information would not be directly released to users’ matches.

Interest Comparison

Ardent would collect information (from interest questions or other social media accounts) about users’ interests and hobbies and compared them with other users. Interests data would be organized into tags shown on the user profile. Two users are more likely to be matched if they have many mutual interests.

Match Limitation

Limited the number of matches that users could get per day so that they would pay more attention to each given match. This way could offer more time for users to learn more about each match and encourage conversations between them. Too many matches might overwhelm users and lower the efficiency of finding the right one.

Match Expiration

Given match for users would expire on certain days. The number of days before the match expired mostly depended on the user’s activity rate. It could motivate users to chat with their matches as soon as possible and plan their physical dates. For safety and privacy purposes, users were unable to retrieve their expired matches unless both sides agreed to move on.

Interest Comparison

On the chat page, the system would popup small tips on how to initiate conversations with the given match and the tips may be curated based on the match’s recent goals, interests, etc. In addition, Ardent developed several small games to spice up the conversation and help two users learn more about each other (e.g. Truth Or Dare, etc.)

Interest Comparison

Ardent would ask users for feedback periodically on given matches, meeting places, etc. The system would ask for confirmation and feedback if two users successfully established a love relationship after several dates and their data would be used to improve the algorithms within the system (require permission from two users).

Interest Comparison

Ardent would collect information (from interest questions or other social media accounts) about users’ interests and hobbies and compared them with other users. Interests data would be organized into tags shown on the user profile. Two users are more likely to be matched if they have many mutual interests.

DEMONSTRATION
CURRENT VERSION
DEMONSTRATION
3RD VERSION

Curated Match

No swiping. No trolling. Ardent curates about 3 matches per day especially for each user. Each match is a gift. Save users' time from browsing unsatisfactory matches.

 

Match Foreshow

Clicking on the heart when the heart has not been filled up, users will get information about their next matches. Build up users' expectations and motivate them to chat with their matches.

 

Interesting Questions

Personality and interest questions motivate users to learn more about their matches. The answers are used as inputs of the algorithm so that Ardent will offer better matches.

Personality Questions

Question 12 and question 13 each represents one type of personality questions, respectively.

The result of personality questions is represented by a diagram with a brief description and is helpful to construct relationships among matches.

 

Interest Questions

Question 14 is an example of interest question.

Answers of interest questions assist in improving the matching algorithm and calculating match index.

Photo Gallery

The photo of the match plays an important role in enhancing the reliability of the match. The visual attraction motivates users to contact their matches.

Interest Tags

The answers of interest questions will be organized automatically into interest tags. Users can browse lots of interest information in the fastest way.

Match Expiration

Match will expire in 3 days to motivate users to chat with their matches. Expiration also protects users' privacy. Users cannot retrieve info of their matches unless both agree on continuation.

Match Index

The initial match index depends on the result of facial analysis algorithm and the number of mutual interests. The match index will be raised by the increase of conversation numbers.

QIXIN DING

dingqixin1130@gmail.com

Welcome to contact me if you would
like to learn more about me or my work.

© 2020 made with ❤️by Qixin

Home

Resume

About

Projects

Ardent

HP UX Internship

ApplyBox

Tencent Interaction Design

  • LinkedIn
  • Facebook
  • Twitter
  • Pinterest
  • YouTube