Wango® - Dating Industry
Quantitative/Qualitative Research | User Experience |
Visual Design | Prototyping | Testing |
Team Management | Agile Scrum Methodology
The main goal was to find out what is the shortest yet most efficient path to convert matches into dates thus increasing the overall success rate.
UX • UI • Product Management
UX: Quantitative & Qualitative Research | User Personas | User Journeys | User Flows | Low - Mid-Fidelity Wireframes | Prototypes | Usability Testing | AB Testing
UI: UI Design | Brand Guidelines
Product Management: Agile scrum methodology | Development supervision (Backend, Web, iOS & Android) | Stakeholder management
7 people (1 Designer & Product Manager, 4 Developers, 1 Marketing, 1 Legal)
Design: Balsamiq | Overflow | Sketch | Principle | Marvel | InVision | Adobe Photoshop CC | Adobe Illustrator CC
Communication & Management: Slack | Trello
User feedback collection from competitors
Type of Research:
Qualitative • Evaluative (Summative) • Attitudinal
Context of Use:
With the help of our developers we ran a script on AppAnnie to retrieve all the user reviews from our competitors within the past 6 months. The script then outlines the most common keywords per rating (from 1-5 stars) and grouped the comments by rating and keywords in an excel file.
Researching the Match-to-Chat conversion rate
Type of Research:
Quantitative • Evaluative (Summative) • Behavioural
Context of Use:
Several profiles from the Wango team have been created on competitor's apps. A bot was then used to automatically like every user on the app.
Within 24h, the test profiles cumulated over 35,000 matches. The objective of these tests was to find out how many conversations would be initiated by men and how many would be initiated by women out of all these matches.
This metric is called the Match-to-Chat conversion rate.
Conducting user surveys
Type of Research:
Qualitative • Generative (Formative) • Attitudinal
Context of Use:
2 Facebook pages targeting the same audience as Wango had been created. One reached 21,000 fans and the other one 2,000.
The objective for these pages was to redirect traffic once the app is out and to conduct surveys leveraging the fanbase. The latter were posted on the pages to collect customer insights and paired with analytics data from Facebook.
Tapping into existing research
Type of Research:
Quantitative • Generative (Formative) • Attitudinal
Context of Use:
Sometimes, research has already been conducted and UX designers can simply leverage existing research. This is one of those cost and time saving methods that will bring valuable insights for the business.
While it doesn't supplement in-house research because it might mot be as specific to what we are trying to find, it is still a valuable complement.
User Research - Recap
The table below offers an overview of the types of research conducted in order to collect data during the Emphasise stage.
The key to online dating is women. By attracting more women on the platform, men will automatically follow.
For that reason, using the above research (mainly the user surveys), I created three empathy maps with a focus on female users.
While empathy maps can be created for a general understanding, or for specific tasks and situations, more holistic empathy maps help feed into more detailed personas. These maps have been shared with the stakeholders and developers for KYC (Know your customers) purposes.
"The user needs a way to find a significant other with the same expectations while surviving the hostile environment datings apps are because women on there have a catastrophic experience."
As mentioned above, the secret to success in online dating is winning over women. For that reason, all personas developed from the empathy maps were focusing around women.
I created three personas using the survey results and we referred to them throughout the entire product development process.
Personas include the following sections: demographics, quote, background, market size, about, goals/needs, motivations, frustrations, personality type, technology and channels of communication
• Single undergraduates (3rd year and above) and postgraduates, aged between 20 and 24. On their first couple years at university they used apps like Tinder to meet lots of people, and now as the degree becomes more demanding, they look for relationships more.
• Singles aged between 25 and 36. University years are behind them and opportunities to meet new people seem to have diminished as work gets more stressful. At the same time, social pressure increases as more and more of their friends are getting engaged if not married.
• Single parents aged above 45, usually with children and trying to take a second chance on life.
All designers get stuck. Having developed central characters in the story helps separate me (the designer) from the user while embracing empathy over bias.
These personas acted as a source-of-truth that invariably impacted the design by mitigating my personal bias. It kept me connected to the target user. The more detailed the persona and their story - the more user-centered I stay.
Card sorting the pain points
From the competitor app reviews collected and gathered in the excel file, the main pain points have been outlined and card sorted in order to have a better visualisation on the potential action points.
Breaking down the current customer journey
By doing so, I have been able to clearly define the major friction points that exist in every dating app and some ideas/directions on how to solve these started forming.
There are 3 major problems with dating apps:
- Dating apps have a bad image and women don’t feel comfortable being associated with a “hook up” app like Tinder.
- Changing branding and image like feminist app Bumble doesn’t solve the second biggest problem: guys. On dating apps, the bad male behaviour is a real scourge for women, making them feel like a piece of meat in a meat locker.
- The industry only contains matchmaking apps not dating apps. The ideal dating app journey is Match – Chat – Meet, yet the industry’s sole focus is matching. They believe that increasing the matching rate will increase the meet up rate. Unfortunately, this approach results in a 0.81% conversion rate from a match to a meet up.
Match-to-Date conversion rate
Really, the most shocking metric that confirmed all the above observations was the match to date ratio. This was was confirmed by an external study involving 200,000 matches.
Fixing the Match-Date conversion rate - 1st attempt
Since the above ratio is the biggest problem, I wanted to test out a few ideas and measure the potential success rate improvements.
One of these ideas was to craft a conversation script and improve it to get as many dates as possible and within the least amount of messages.
Below is the best script tested:
Script tested on competitors' apps:
[1.2]: Hey [Name]! 😊 So what are you doing in London?
: Wanna play 2 facts one lie?
[2.1]: Each of us tells 2 facts and one lie about him/herself and the other has to guess.
[2.2]: Hehe alright so: 1. Statement #1 - 2. Statement #2 - 3. Statement #2
[2.3]: Anyway, it's your turn now ^^
: Shall we do one more round? And how about we add a little bit of a challenge here: this time, if get it right, I can take you out for a drink! Make sure you come up with some tricky facts ;)
: So when would you be available this week? :)
Scripts testing and Findings
The scripts have been tested on several apps and by all the team members' profiles in order to gather enough insights.
• An ice-breaker drastically improves the initial response rate
• 2 truths 1 lie is the most effective and universal ice-breaker
• The gamification touch was positively received
More about 2 truths 1 lie:
What's powerful with this ice-breaker is that it can be used every time and yet it will generate unique conversations each time. In an instant there are 6 directions that can be leveraged to build up a conversation. And since it comes as a mini challenge, a second round would add even more possibilities to find common grounds.
Beta testing phase - Code name "Heyyy"
At this point we had built and MVP app that had all the basic components considered industry standards PLUS 2 new features taking place at the "Chat" stage of the journey.
The MVP app would:
• Suggest to play 2 truths 1 lie to break the ice
• Feature a "meet up request button"
• Display a list of nearby cafes, bars, restaurants, etc. using Yelp's API
I ran a usability test with 30 beta testers in order to verify that the new features were understood, used and useful.
MVP roll out
While the beta testing allowed for some usability and app stability testing, we needed to more users to test the features at a larger scale and obtain validation.
Within 2 months the app got a little over 1,000 organic users. On the backend we had a funnel in place to measure:
Users > Matches > Conversations with more then 20 messages > Ask out button pressed > Dates planned
While the overall Match-Date ratio showed an increase by a factor 2.8, something wasn't right around the use of the "ask out button". We definitely had some good margin for improvement.
We even had some instances where users would skip the ice breaker but use the ask out feature and meet up suggestions later on.
• Icebreakers: ✅
• 2 truths 1 lie: ✅
• Meet up places suggestions: ✅
• Ask out button: ❌
• The whole icebreaker package and how it's introduced ❌
• Yelp API not returning enough search results ❌
Rationale behind those results:
• The (Gentleman/Lady) Badge System impacted the overall branding and added an extra layer of comfort for women
• The Emoji Bio added just the right amount of user information to generate more thoughtful connections rather than matching and vetting after. In some instances it has also been used as a way to break the ice.
• The Progress Bar introduced a layer of gamification which paired with the Smart Assistance created a chat-to-meet-up user journey with very little to no friction. It tackled all the questioning that user undergo by taking care of pace and planning which are the major block points.
• The Date Planning feature removed the hurdle out of coming up with a date plan by offering a package deal solution that can then be modified at will to adapt to availabilities and preferences.